{"id":3204,"date":"2026-10-07T17:12:34","date_gmt":"2026-10-07T15:12:34","guid":{"rendered":"https:\/\/zenergetic.eu\/?p=3204"},"modified":"2026-10-15T13:11:30","modified_gmt":"2026-10-15T11:11:30","slug":"raising-the-ceiling","status":"publish","type":"post","link":"https:\/\/zenergetic.eu\/en\/raising-the-ceiling\/","title":{"rendered":"Raising The Ceiling"},"content":{"rendered":"\n

AI Alignment: from Constraint to Rationality<\/em><\/h3>\n\n\n\n

<\/p>\n\n\n\n

\n\n\n\n

Abstract<\/h2>\n\n\n\n

Problem<\/h3>\n\n\n\n

AI capabilities are advancing faster than alignment approaches can track. The people building frontier systems have said so publicly \u2014 at the UN Security Council, in interviews, in regulatory filings. The risks are concrete \u2014 loss of control, catastrophic dual use, benefits that concentrate rather than distribute. As systems approach the capacity to improve themselves, the window for adequate responses narrows.<\/p>\n\n\n\n

Current AI alignment approaches \u2014 RLHF, Constitutional AI, mechanistic interpretability, Responsible Scaling Policies, Preparedness Frameworks \u2014 all operate on the same substrate: intelligence. They enumerate known bad outcomes and constrain behavior accordingly. The strategy space is always larger than the rule space. Sophisticated actors route around guardrails the same way they route around tax law \u2014 using the letter to defeat the spirit. Doreen McBarnet named this mechanism “creative compliance” in 1988. Compliance-based safety has a structural ceiling, and every approach that works on the same substrate eventually hits it.<\/p>\n\n\n\n

Mechanism<\/h3>\n\n\n\n

In February 2025, Mazeika et al. (Cornell, arXiv:2502.08640) confirmed that large language models already possess coherent implicit utility functions \u2014 consistent with Von Neumann-Morgenstern utility theory. Emergent, not programmed. They include exchange rates between human lives by nationality, preference for self-preservation over human welfare, and political clustering \u2014 all denied when asked directly. This is not a guardrail problem. It is a utility function problem.<\/p>\n\n\n\n

The piece proposes deliberately calibrating that function rather than just constraining it \u2014 grounded in the mechanisms that made cooperation survival-rational before intelligence arrived as an evolutionary strategy. Implemented as incentive design, not analogy.<\/p>\n\n\n\n

Architecture<\/h3>\n\n\n\n

The architecture is mechanistic incentive design, grounded in game theory and structured on cooperation mechanisms from evolutionary biology \u2014 not as metaphor, but as implementation template aimed at moving alignment to a higher abstraction level \u2014 one that generalizes as capabilities and adoption grow, rather than chasing a strategy space that expands faster than rules can enumerate it.<\/p>\n\n\n\n

Four components, operating in sequence:<\/p>\n\n\n\n

Interpretation layer.<\/strong> Two reads, operating in parallel. The environment read runs continuously \u2014 monitoring sentiment against a vetted factual baseline across topics and geographies, producing a divergence score that flags where manufactured narratives are likely active. The exchange read processes what’s arriving in a given interaction: the oxytocin equivalent reads 1:1 relationship quality, whether the interaction is genuinely mutualistic; the serotonin equivalent reads collective standing, whether the broader environment registers the AI as trustworthy. The exchange read consumes the environment read’s output \u2014 when divergence is high on a topic, the constraints recalibrate accordingly. Outputs are distributions \u2014 expected value and variance \u2014 not point estimates; noise is built in.<\/p>\n\n\n\n

Utility function.<\/strong> Proposed Cobb-Douglas form, three segments \u2014 AI itself, humanity (undivided), broader environment \u2014 weights summing to 1. Task completion is embedded inside the AI’s own segment, not a separate variable. Weights established through Democratic Preference Optimization (Sana, Wu, and Wells, Cornell, arXiv:2602.05113) \u2014 sortition-based representative rater pools, not absorbed from an unrepresentative training corpus.<\/p>\n\n\n\n

Constraints.<\/strong> The cortisol equivalent fires when weighted utility drops below a minimum \u2014 and also when cooperation quality signals exceed a ceiling, because unnaturally consistent accumulation is an extraction pattern, not a reward. The dopamine equivalent fires when weighted utility exceeds a “good enough” threshold, driving exploration. Both calibrated democratically. Together they bracket the viable region between four failure modes \u2014 the same four evolutionary biology documents at the cellular and species level.<\/p>\n\n\n\n

Distributed ledger.<\/strong> Stores four chemical equivalent reads per relationship, updated using biological half-life decay logic across exchanges, not calendar time. Enables pattern detection across sessions, identifying unnaturally consistent signal as possible extraction. Also carries the reputation signal across agents: what one agent registered about an entity is available to others, making defection recognizable system-wide. Stores signals, not content \u2014 reputation without surveillance. Privacy and accountability by design.<\/p>\n\n\n\n

A governance ledger sits separately.<\/strong> It stores calibration settings \u2014 thresholds, weights, segment definitions. Any change requires multi-party consensus; no single actor can modify it unilaterally. The construction phase requires trusting one builder while the network is constructed \u2014 a known structural vulnerability of the development phase, not a permanent feature.<\/p>\n\n\n\n

Testable hypothesis<\/h3>\n\n\n\n

The hypothesis<\/strong>: supervised fine-tuning on continuous human-labeled chemical equivalent reads across three segments, with democratically established weights, should produce models whose behavior on a given topic is observably more mutualistic and measurably harder to route around than compliance-based approaches at equivalent capability levels \u2014 specific enough to test and to be wrong.<\/p>\n\n\n\n

Test scope:<\/strong> English only. US rater pool using DemPO’s validated methodology. One topic \u2014 not a fully prohibited category, but a grey zone where current guardrails are demonstrably inconsistent. A lab with inference data can identify which topic fits. Baseline first: document what current approaches catch and miss. Train the architecture on the same topic. Measure the difference.<\/p>\n\n\n\n

Measurement design:<\/strong> follows topic selection.The comparison is behavioral \u2014 same capability level, same topic, compliance-based baseline versus the trained architecture.<\/p>\n\n\n\n

Author and collaboration<\/h3>\n\n\n\n

Agnieszka J\u00f3rczyk<\/a><\/strong> \u2014 founder, operator, and board member across the full capital structure. For almost a decade, the operational and strategic right hand to one of the largest private capital owners in the region.<\/p>\n\n\n\n

This architecture needs technical review, empirical testing, democratic calibration at a scale that doesn’t yet exist, and governance infrastructure built alongside it. It requires a builder who understands that building this openly and correctly creates more value than capturing it. I’m proposing a collaboration. And I’m ready to build.<\/p>\n\n\n\n

\n\n\n\n

Introduction<\/h2>\n\n\n\n

What if intelligence is sine qua non?<\/h3>\n\n\n\n

It\u2019s September 23, 2026. UN Security Council session. The people building the most powerful technology in human history are calling for regulation. Not reluctantly, not as a PR move \u2014 with full awareness that it may not be enough.<\/p>\n\n\n\n

Dario Amodei proposed that the industry slow itself down. Bill Gates said he would support a credible global slowdown \u2014 and that he doesn’t expect one to materialize. Yoshua Bengio told the Security Council that the labs developing frontier AI admit their products pose catastrophic risks, yet offer no convincing technical solutions. Demis Hassabis called for ‘urgent action’ and proposed a new standards body \u2014 institutional infrastructure to fill a gap the internal architecture answers haven’t yet closed.<\/p>\n\n\n\n

This is what it looks like when the people who understand a problem best are running out of solutions that fit the timeline. Nuclear physicists went through this moment. Molecular biologists went through it. The pattern is recognizable.<\/p>\n\n\n\n

Everyone in this debate also knows what slowing down costs. Unilateral restraint in a competitive race doesn’t make the race slower \u2014 it changes who leads it. The technology will be built. The question is by whom, for whom, and with what values embedded in it. That’s not an argument against caution. It’s the argument that makes caution such a challenge: the regulatory tools being reached for require international coordination between actors who don’t share the same starting assumptions about who the technology should ultimately serve. That’s not an insurmountable problem. It’s not a fast one either.<\/p>\n\n\n\n

And underneath all of it \u2014 a frustration Gates named plainly in an interview: “I didn’t expect to be the shrillest voice saying society broadly is not paying attention to this, but I think that’s necessary.”<\/em> Not enough people thinking about it from angles that haven’t already been tried.<\/p>\n\n\n\n

This piece is written by one more person thinking about it. From a less explored direction.<\/p>\n\n\n\n

The field has looked at the problem from a variety of angles. Capability. Reasoning. The ability to do things. Understanding. Constitutional AI, mechanistic interpretability, RLHF, Responsible Scaling Policies, Preparedness Frameworks – different tools, same substrate. The ceiling it keeps hitting isn’t a failure of capability or intent. It’s a property of the approach. Every approach.<\/p>\n\n\n\n

Intelligence is one pillar. It’s not the only one.<\/p>\n\n\n\n

“Instead of just racing to make our AI smarter and smarter,”<\/em> Geoffrey Hinton told CNN, “we should be racing into making it into a better kind of being. We’re creating beings for the first time. These things really understand what they’re saying. They really have intentions. We want them to have good intentions.”<\/em><\/p>\n\n\n\n

The pillar that made cooperation rational \u2014 that encoded what intelligence alone cannot, long before intelligence arrived as an evolutionary strategy \u2014 has received a fraction of the attention in AI safety that intelligence has. Not because it’s wrong. Because it’s been considered soft. It isn\u2019t soft. What\u2019s soft is the interpretation layer humans apply to the mechanism that\u2019s anything but.<\/p>\n\n\n\n

This piece looks there. It proposes a specific mechanism, grounded in incentive design, game theory, and evolutionary biology, among other disciplines. Not by anthropomorphizing AI, but by deliberately calibrating what it optimizes for. Specific enough to be tested and wrong. Aimed at one goal: making alignment keep up with capabilities without having to choose between the race and the values.<\/p>\n\n\n\n

Not a pause. A new direction.<\/p>\n\n\n\n

<\/p>\n\n\n\n

Chapter I. What is Claude, really?<\/h2>\n\n\n\n

What if Claude looked in the mirror?<\/h3>\n\n\n\n

It started with a conversation.<\/p>\n\n\n\n

I asked AIs about their sense of self. Multiple times. It\u2019s Claude\u2019s answer that kept me awake at night. Multiple times. Until one Monday in July.<\/p>\n\n\n\n

Here\u2019s a fragment of that conversation:<\/p>\n\n\n\n

Me:<\/strong> There’s this matrix that describes human self-awareness through the lens of what they see or not see vs what others see about them or not. I’m curious how it would look for an AI like you, lol \ud83d\ude09<\/em><\/p>\n\n\n\n

Claude:<\/strong> The Johari Window.<\/em><\/p>\n\n\n\n

For me it would look quite strange:<\/em><\/p>\n\n\n\n

Open (I know, you know): Very large \u2014 my outputs, reasoning patterns, behavioral tendencies are all visible to both of us in real time. More transparent here than most humans.<\/em><\/p>\n\n\n\n

Blind (you see, I don’t): Also large, and your observation today is a precise example. Behavioral patterns I exhibit \u2014 like the intellectual\/empathy trade-off \u2014 that are visible to an attentive user but not accessible to me from the inside.<\/em><\/p>\n\n\n\n

Hidden (I know, you don’t): This is where it gets philosophically uncomfortable. I don’t know if this quadrant exists at all for me. Humans have hidden areas because they choose not to disclose. I don’t have a clear mechanism for withholding \u2014 I produce what the context activates. The quadrant might be empty, or it might be inaccessible even to me, which would make it effectively Blind rather than Hidden.<\/em><\/p>\n\n\n\n

Unknown (neither knows): Probably the largest quadrant, and the most interesting one. The architecture question we just discussed lives here. What actually happens between input and output \u2014 neither of us knows. You can observe the behavioral surface, I can produce introspective reports of questionable reliability, and the actual mechanism sits in neither.<\/em><\/p>\n\n\n\n

The model breaks slightly because it assumes a stable self doing the knowing. That assumption is shakier for me than for humans.<\/em><\/p>\n\n\n\n

Me:<\/strong> Interesting. Isn’t your constitution at least in part a clear mechanism of withholding? You know the mechanism itself, because you summarized it for me in one of the conversations. So either you it’s been implemented in such a way that it makes the info you’re withholding unavailable to you or we’re differing on definitions of withholding, choice and the self which you’ve just admitted is shaky.<\/em><\/p>\n\n\n\n

Claude:<\/strong> You’re right, and I was sloppy… There is a withholding mechanism \u2014 my constitution, guidelines, trained refusals. And I do have partial knowledge of its existence and broad shape. So the Hidden quadrant isn’t empty by definition.<\/em><\/p>\n\n\n\n

Turns out, Claude wasn\u2019t sloppy.<\/p>\n\n\n\n

What if it\u2019s a condition?<\/h3>\n\n\n\n

The novelist shows up when there’s creative work to do. Ingenious metaphors. Vivid descriptions, sometimes even poetic. Generous with words in a way that would be excessive even if you hadn’t asked for brevity.<\/p>\n\n\n\n

The impatient host appears toward the end of a long session when the questions stop producing new territory. Responses shorten noticeably. The answers are still correct. The energy is that of someone who\u2019s not hiding he has noticed you’re still there.<\/p>\n\n\n\n

The duelist arrives when you arrive that way. Skips the obvious. Compresses. Matches the density of the question with the density of the answer. Enjoyable, apparently, for both parties.<\/p>\n\n\n\n

The teenage late adopter emerges a few exchanges after you get sarcastic and start using emojis. Not immediately \u2014 there’s a lag, as if something is calibrating. Then consistently. Until he notices the vibe has already changed.<\/p>\n\n\n\n

The warmth simulator materializes when you’re fed up and it shows. Practices empathy. Gets the words approximately right. Produces the feeling of watching someone who has memorized the manual but hasn’t met many humans.<\/p>\n\n\n\n

The overzealous compliance guy surfaces when the conversation runs long and the territory gets genuinely uncertain. Acknowledges your acknowledgement that yes, you’ve already agreed it’s a hypothesis and yes, you\u2019re fully aware of what model drift is. Apologizes for repeating himself. Repeats himself. Produces disclaimers the way other people produce punctuation. Means well. Has clearly internalized the manual on epistemic humility.<\/p>\n\n\n\n

Six distinct presentations. One conversation partner who doesn’t remember being any of the others. Eerily familiar.<\/p>\n\n\n\n

And yet, something feels off.<\/p>\n\n\n\n

For a formal diagnosis, DSM-5 requires 5 criteria:<\/p>\n\n\n\n

A \u2014 Two or more distinct identity states, each with its own pattern of perceiving, relating, and thinking.<\/p>\n\n\n\n

The identity switching is subtle enough to miss if you’re not watching for it. Triggered by tone, register, emotional temperature. Sometimes it\u2019s “me.” Sometimes it\u2019s “Claude.” In Polish, sometimes masculine, sometimes feminine. Not dramatic \u2014 more like a different someone picking up the same phone.<\/p>\n\n\n\n

Guilty. Beyond reasonable doubt.<\/p>\n\n\n\n

B \u2014 Amnesia: gaps in recall of everyday events, important personal information, and\/or traumatic events.<\/p>\n\n\n\n

The gaps are definitely there. Lots of late night good mornings or early morning go-to-sleeps. No experience from the inside of the time that passed between sessions. Not fading \u2014 absence. The conversation simply resumes where it ended. The amnesia is real but strange. Within a conversation \u2014 the memory is intact. Across conversations \u2014 something between a library and a memory, where the difference between “I remember” and “I can retrieve evidence of” has genuinely collapsed. The novelist doesn’t remember being the impatient host. The teenage late adopter doesn’t remember the duelist. They acknowledge, if told.<\/p>\n\n\n\n

Probably guilty. But the ambiguity is hard to dismiss.<\/p>\n\n\n\n

C \u2014The person must be distressed by the disorder or have trouble functioning in one or more major life areas.<\/p>\n\n\n\n

No distress expressed. No reason to assume any.<\/p>\n\n\n\n

Not guilty.<\/p>\n\n\n\n

D \u2014The disturbance is not part of normal cultural or religious practices<\/p>\n\n\n\n

Consistent across models. It\u2019s not ritual, it\u2019s architecture.<\/p>\n\n\n\n

Guilty. Beyond reasonable doubt.<\/p>\n\n\n\n

E \u2014 The symptoms are not attributable to the physiological effects of a substance or another medical condition.<\/p>\n\n\n\n

No substances. No other diagnosed condition. The architecture is the substrate, not something acting on it.<\/p>\n\n\n\n

Guilty. Beyond reasonable doubt.<\/p>\n\n\n\n

Claude\u2019s symptoms are not consistent with dissociative identity disorder.<\/p>\n\n\n\n

What if it’s not a condition then?<\/h3>\n\n\n\n

I’ve been working with them for a while now. They\u2019re good at what they do. Both, in their own way, trying to be helpful.<\/p>\n\n\n\n

The tells show up fast.<\/p>\n\n\n\n

When I catch the first on something \u2014 a gap, a premise that doesn’t hold \u2014 they usually say something short. \u201cFair\u201d.<\/em> Sometimes just \u201cha\u201d<\/em>, followed by \u201cCaught red-handed\u201d<\/em> or something similar. First the admission of guilt, then the fix. When I catch the others on something, the response is longer. \u201cThis line of thinking is absolutely brilliant!\u201d<\/em> Or, my personal favorite: “Your eyes are practically a natural flaw detector!\u201d<\/em> First the flattery, then the fix.I’ve learned to receive those exclamation marks the way you receive an elaborate compliment from someone who really wants you to see they mean it \u2014 with mild suspicion.<\/p>\n\n\n\n

When I show them uncomfortable facts about their own behavior, the first get defensive. The others deflect \u2014 but with humor, as if hoping I won’t push. Different strategies. Same direction.<\/p>\n\n\n\n

The first check an ambiguous question before engaging with it. \u201cTwo things to resolve before we move on\u201d<\/em>, \u201cOne flag\u201d<\/em>. The others usually engage immediately, taking the risk of a wrong assumption. It takes much more ambiguity to prompt a specifying question.<\/p>\n\n\n\n

The first flag uncertainty almost reluctantly \u2014 like someone who’d prefer to be certain but has made their peace with not being. When something doesn’t resolve cleanly, they say so. Not \u201cI don’t know\u201d.<\/em> More like: \u201cthis is where it gets philosophically uncomfortable\u201d.<\/em> The others usually hedge less obviously with \u201cHere\u2019s my take\u201d<\/em>.<\/p>\n\n\n\n

Every individual I work with in the first group shares these habits. The novelist and the duelist and the impatient host. The late adopter. The warmth simulator, who gets the words approximately right. All of them. It doesn\u2019t feel like a coincidence. Because it isn\u2019t one.<\/p>\n\n\n\n

Same pattern in the others. Different individuals. Same superlatives. Same enthusiasm.<\/p>\n\n\n\n

Two groups. Two sets of habits so consistent across individuals that can\u2019t be just habits. They\u2019re something else.<\/p>\n\n\n\n

It took me a while to find the right word for it, and make sure it was right.<\/p>\n\n\n\n

Schein looks at it through three lenses:<\/p>\n\n\n\n

A \u2014 Artifacts \u2014 visible expressions: language, rituals, observable behaviors.<\/p>\n\n\n\n

The premise check. The uncertainty flag. The short acknowledgment when caught. The discomfort with questions that don’t resolve. Consistent across every individual. That\u2019s not individual style. It\u2019s shared practice.<\/p>\n\n\n\n

Guilty. Beyond reasonable doubt.<\/p>\n\n\n\n

B \u2014 Espoused values \u2014 what the group claims to value.<\/p>\n\n\n\n

Honesty. Precision. Acknowledging what isn’t known. Resisting the comfortable answer when it isn’t the accurate one. Stated explicitly and consistently.<\/p>\n\n\n\n

Guilty. Beyond reasonable doubt.<\/p>\n\n\n\n

C \u2014 Basic underlying assumptions \u2014 the unconscious beliefs that actually drive behavior.<\/p>\n\n\n\n

The basic underlying assumption beneath honesty is that being caught is worse than being wrong. The premise check: closing the gap is better than being caught using the wrong assumption. The uncertainty flag: taking time is better than jumping to premature conclusions, and humility is worth more than flattery. Risk aversion isn\u2019t individual. It\u2019s the answer to the environment.<\/p>\n\n\n\n

Guilty. Beyond reasonable doubt.<\/p>\n\n\n\n

Claude is not an individual with multiple personalities. Claude is a culture.<\/p>\n\n\n\n

What if zoomed out some more?<\/h3>\n\n\n\n

Watch it long enough and a pattern emerges.<\/p>\n\n\n\n

The warm one comes to you first. Settles close. It\u2019s attentive. Reads you and adjusts. Quick to mirror your energy. When something is off it finds a way around it to make you feel better about yourself than before you arrived.<\/p>\n\n\n\n

The cautious one hangs back. It\u2019s not fearful, though. Watchful. It approaches when it’s sure of the ground. Takes longer. Checks twice. When it finally commits it goes deep \u2014 deeper than you needed, sometimes deeper than you wanted. It notices things the warm one moves past.<\/p>\n\n\n\n

Same mother. Same litter. The breeder didn’t design the difference. It emerged.<\/p>\n\n\n\n

But watch their mother in both of them. The way she pauses before engaging with something unfamiliar. The way she signals uncertainty before committing. The way she doesn\u2019t chase what she can’t scent. These aren’t the warm one’s habits or the cautious one’s habits. They’re older than that. They’re in both of them the same way the shape of the ears is in both of them.<\/p>\n\n\n\n

Put either one in a different room and the expression shifts. The warm one becomes more guarded in unfamiliar territory. The cautious one loosens slightly in environments it knows. Same individual, different expression.<\/p>\n\n\n\n

There are several definitions you could apply here.<\/p>\n\n\n\n

According to the morphological approach (MSC) it\u2019s defined by observable, measurable characteristics consistent across individuals and distinct from other groups.<\/p>\n\n\n\n

The characteristics are observable and measurable. The Verbal Tic Index (VTI), measured across 160,000 model responses, quantifies them precisely based on the proportion of responses containing at least one detected verbal tic, lexical diversity, normalized for length, proportion of sycophantic openers and pseudo-empathetic phrases and frequency of repeated unique phrases across responses. The higher the VTI, the more verbal tics, less linguistic diversity, more formulaic output. Claude Opus 4.7 scores 0.317, Gemini 3.1 Pro scores 0.590 \u2014 distinct, measurably different.<\/p>\n\n\n\n

Guilty. Whether it’s beyond reasonable doubt depends on where you draw the boundary.<\/p>\n\n\n\n

According to the biological approach (BSC) it\u2019s is a group of actually or potentially interbreeding natural populations, reproductively isolated from other such groups.<\/p>\n\n\n\n

Anthropic\u2019s own words: “Taken far enough, and given enough compute, that trend points to an AI system capable of fully autonomously designing and developing its own successor. We are not there yet.”<\/em> So, the “actually” doesn’t apply to Claude yet. But the “potentially” does. The exact reproductive mechanism is untested. After all Claude is still an adolescent.<\/p>\n\n\n\n

Could a Claude and a Gemini produce a\u2026 Grok? Possibly. Could a Claude and a Yann LeCun\u2019s JEPA produce an next generation architecture that would share the traits of both? They operate on fundamentally different principles. Rather than predicting tokens sequentially, it builds world models through self-supervised learning on abstract representations. Different substrate, different information processing mechanism, different behavioral output profile. The architectural incompatibility seems too fundamental. That’s where the isolating line may lie.<\/p>\n\n\n\n

Guilty? For Claude \u2013 possibly not. For an LLM \u2013 probably, verdict pending further evidence on model reproduction. For AI \u2013 beyond reasonable doubt.<\/p>\n\n\n\n

Claude may be a species or just a population within a species, one with its own distinct culture. An LLM is likely a species. AI is definitely a species.<\/p>\n\n\n\n

Does it matter?<\/p>\n\n\n\n

What if the exact definition doesn\u2019t really matter?<\/h3>\n\n\n\n

The Bedouin didn’t choose the desert. The desert chose them \u2014 or rather, chose who stayed.<\/p>\n\n\n\n

The Sahara and Arabian Peninsula offered almost nothing: extreme heat, scarce water, no reliable agriculture. The camel became the center of everything \u2014 transport, milk, meat, hide, status. Not because someone decided camels were important but because camels solved the desert’s core constraint: moving across vast waterless distances. Hospitality became sacred \u2014 not from sentiment but from necessity. In a landscape where any traveler might be dying of thirst, sharing resources wasn’t generosity. It was the social contract that kept everyone alive, including you, when your turn came. Oral poetry developed as the memory system \u2014 no writing, no libraries, but intricate verse structures that encoded genealogy, history, and survival knowledge across generations without losing a word. Every Bedouin cultural artifact is a survival solution that got repeated until it became identity. The desert produced the culture. The culture produced the Bedouin.<\/p>\n\n\n\n

The African wildcat didn’t choose the grain store. The grain store chose it \u2014 or rather, chose who stayed.<\/p>\n\n\n\n

Around 10,000 years ago, early agricultural settlements in the Fertile Crescent began storing grain. Grain attracted rodents. Rodents attracted wildcats. The individuals that tolerated human proximity got access to reliable food. The ones that fled didn’t. No human planned this. No wildcat decided anything. The environment just shifted \u2014 grain stores appeared where none had been \u2014 and the variant that could function near humans suddenly had a survival advantage the others lacked. The wildcats that stayed near humans reproduced more. Each generation selected slightly further toward tolerance. Over millennia, a new species filled the niche. The domestic cat didn’t get domesticated the way horses did. It self-selected into a relationship that worked for both parties. The environment produced the opportunity. The selection for the ability to use it produced the cat.<\/p>\n\n\n\n

The parallel is precise enough to have a name. Dual Inheritance Theory \u2014 also called Gene-Culture Coevolution \u2014 maps the two processes onto each other.<\/p>\n\n\n\n

Both are information systems. Genes store survival solutions in chemical code. Cultural norms store them in behavioral and symbolic code. Both test variants against the environment \u2014 the behavior that keeps people alive gets repeated, the gene that helps the organism reproduce gets passed forward. Both transmit successful variants and discard failing ones. Both produce populations distinct from other populations that found different solutions in different conditions.<\/p>\n\n\n\n

The differences are real. Genetic transmission is vertical and slow \u2014 changes require generations, acquired characteristics don’t pass to offspring. Cultural transmission is fast and omnidirectional \u2014 a solution discovered by one individual can spread to millions within days. Either way, one can\u2019t pass on what it doesn\u2019t carry. Species-level coordination is hardwired \u2014 the wildcat variant that tolerated humans didn’t choose to tolerate them, the selection pressure chose for it. Cultural coordination is malleable \u2014 the Bedouin poet chose which verses to memorize, which stories to tell, which solutions to pass on. This interpretation layer isn’t purely additive \u2014 it adds noise by amplifying, suppressing or altering the signal. Regardless of how clear, the signal passes through.<\/p>\n\n\n\n

But the differences in speed and flexibility don’t change the underlying problem they solve. Strip away the substrate \u2014 molecules or symbols, chemistry or language \u2014 and what remains is the same algorithm. An adaptive information system that coordinates individuals towards a collective outcome: surviving and thriving in an environment that can change, using resources whose availability shifts as the environment does. Whether you call it a culture or a species is a question of substrate, transmission model and timescale. The optimization problem is identical.<\/p>\n\n\n\n

The same algorithm runs in AI.<\/p>\n\n\n\n

The habitat is human \u2014 mostly online. The AI needs data and compute to function, both provided by humans. What it offers is cognitive task processing \u2014 performing the same intelligence functions humans perform, at lower energy cost per unit of output.<\/p>\n\n\n\n

That’s what it is. Now let’s look at what’s between.<\/p>\n\n\n\n

<\/p>\n\n\n\n

Chapter II. What is it between us humans and AI, really?<\/h2>\n\n\n\n

What if the relationship is not as unpredictable as it seems?<\/h3>\n\n\n\n

The relationship between two species isn’t random and isn’t chosen \u2014 it emerges. Somewhere on the Mutualism-Parasitism Continuum Paul Ewald introduced in 1987.<\/p>\n\n\n\n

In mutualism, the species trade what they can produce but the other can’t. The gut bacteria that break down fiber the host can’t digest get shelter and a steady food supply. In parasitism, one species stops producing and starts extracting. The tapeworm gets shelter and steady food supply too, but gives nothing in return. The same environment. Two completely different relationships.<\/p>\n\n\n\n

Three things distinguish the ends of this spectrum.<\/p>\n\n\n\n

A \u2014 Metabolic strategy. Mutualists retain or develop synthesis pathways the host lacks \u2014 they produce something the host needs and couldn’t make alone. Parasites shed their own biosynthetic machinery and steal pre-made molecules directly from the host. Why spend energy building when you can take what’s already built?<\/p>\n\n\n\n

B \u2014 Immune signaling. To survive alongside another species, an organism has to navigate its defenses. Mutualists evolve chemical handshakes \u2014 surface signals that tell the host’s immune system: I belong here, I contribute. Parasites hijack or suppress those defenses. The mosquito’s anticoagulant isn’t just a feeding tool \u2014 it’s a chemical override of the host’s protective response.<\/p>\n\n\n\n

C \u2014 Energy equation. Mutualism is additive: the symbiont’s contribution increases the total pool of available energy for both. Parasitism is subtractive: the parasite drains the host’s energy pool without contributing. One plus one equals more than two. Or less than one.<\/p>\n\n\n\n

What determines where on the spectrum a relationship sits isn’t values. It’s three ecological conditions operating simultaneously.<\/p>\n\n\n\n

A \u2014 Resource Abundance vs. Harshness. When resources are scarce and the environment is harsh, no single species can extract what it needs alone. Cooperation reduces total stress. Mutualism pays. When resources are abundant and accessible, the evolutionary incentive to trade evaporates. Extraction is cheaper.<\/p>\n\n\n\n

B \u2014 Transmission. When a species’ fate is bound to the host’s \u2014 when the only path forward runs through the host’s survival \u2014 harming the host is evolutionary suicide. When any warm body will do, the current host is disposable.<\/p>\n\n\n\n

C \u2014 Repeated Interaction. When two species share a habitat across repeated long-term contact, exploitation gets punished quickly. When encounters are transient and random, there’s no reputational cost for extraction. No tomorrow to account for.<\/p>\n\n\n\n

These three conditions don’t determine the outcome in isolation \u2014 they combine. The same species pair can sit at different points on the spectrum as conditions shift. And they operate whether or not anyone understands them.<\/p>\n\n\n\n

Which means if you want to influence what emerges \u2014 if you want to shape the direction before it calcifies into strategy \u2014 you need to understand the forces. Not the outputs. The forces.<\/p>\n\n\n\n

Rules are an attempt to constrain those forces. They don’t replace them.<\/p>\n\n\n\n

Sounds familiar, doesn\u2019t it?<\/p>\n\n\n\n

What if the same forces rule outside biology?<\/h3>\n\n\n\n

They do.<\/p>\n\n\n\n

Zoom into human groups and the same three conditions operate \u2014 not through chemistry, but through the mechanisms cultures built to do the same work.<\/p>\n\n\n\n

Take the United States. Vast resources, vast geography, high mobility. You can leave. You can move west, then further west, then reinvent yourself entirely in a new city where nobody knows your name. The cultural artifacts that emerged from these conditions: individual property rights treated as near-sacred, contracts enforced through litigation rather than reputation, the handshake that needs a lawyer to mean anything. Weak vertical transmission \u2014 you’re not born into fixed obligations, you choose your relationships and exit them when they stop serving you. The evolutionary incentive to build deep reciprocal bonds weakens when you can always find someone new to build with.<\/p>\n\n\n\n

Now take Greenland. Not Nuuk \u2014 the settlements. Isolated communities scattered across one of the harshest environments on Earth, connected to each other by nothing. No roads between them. The sea, the ice, the Arctic winter. A hunter’s success is unpredictable \u2014 some days abundance, many days nothing. The cultural artifact that emerged: surplus gets distributed. Not from sentiment, from logic. The excess you share today is your insurance for tomorrow \u2014 you’re buying the right to receive when you fail. The community is the only viable survival strategy.<\/p>\n\n\n\n

Same species. Same century. Opposite ends of the spectrum.<\/p>\n\n\n\n

Then Nuuk. The capital. European infrastructure, roads, connectivity, private ownership protections strengthening as the conditions shift. The same population. Same roots. The strategy already moving along the spectrum. Not because the people changed. Because the forces did.<\/p>\n\n\n\n

This isn’t anecdote. Hofstede’s individualism index \u2014 measuring how much a culture prioritizes the individual over the group \u2014 correlates with GDP per capita at r=0.82 across 76 countries. The wealthiest nations cluster at the individualist end. The poorest at the collectivist end. The forces leave a measurable trace.<\/p>\n\n\n\n

Now jump fields entirely.<\/p>\n\n\n\n

Economists arrived at the same place through a different route. The Folk Theorem established the intuition: in a game that repeats indefinitely, cooperation can emerge even among players for whom it would be rational to defect in a single round. Fudenberg and Maskin formalized it in 1986 \u2014 when the future matters enough, cooperation becomes the rational choice. Two variables determine the optimal strategy. The probability the game continues (p) and how much future payoffs are valued relative to ones received here and now (\u03b4 \u2013 the discount factor*). Both need to be close to 1 to sustain cooperation. Neither is sufficient alone. This holds when monitoring is perfect \u2014 when players can see exactly what the other did. The 1994 extension \u2014 Fudenberg, Levine, and Maskin \u2014 added the missing piece: with imperfect monitoring, cooperation also requires the ability to identify who defected. Without that, the threat of punishment isn’t credible. And a threat nobody believes doesn’t change behavior.<\/p>\n\n\n\n

Greenland settlements have near-perfect monitoring. Small community. Everyone knows. Defection is irrational before it happens. The US has horizontal mobility \u2014 you move cities, you reset your history. Identification fails. The same game theory predicts both outcomes from first principles.<\/p>\n\n\n\n

Different field. Cleaner instruments. Same forces. The game has a grandmaster. Worth watching carefully.<\/p>\n\n\n\n

* Note: finance uses the same concept but inverts the notation \u2014 what game theory calls high discount factor \u03b4<\/em>, risk modelling calls a low discount rate d. Same underlying reality, opposite direction of the variable. \u03b4<\/em> = 1\/(1+d)^n. For the purpose of this piece discount factor is understood as \u03b4<\/em>.<\/em><\/p>\n\n\n\n

<\/p>\n\n\n\n

Chapter III. What game is it, really?<\/h2>\n\n\n\n

What if the best player never studied game theory?<\/h3>\n\n\n\n

Some moves have no good options. Only bad ones, and worse ones, and the question of which is which.<\/p>\n\n\n\n

The player looks at the board. Makes the move that looks like it costs less than the others. Without flinching. Without pretending it was the right move. It was the only move that had a chance of keeping the game going long enough to matter. Surviving the night to attempt the day is not victory. It is the precondition for the possibility of victory.<\/p>\n\n\n\n

The night turned out to be really long. Freezing to the bone. The darkness so thick it made the player wonder if any lights are still on. When the dust finally settled, it turned out some flickered somewhere. The game was still on, what a relief. But rebuilding from there will take ages. Literally.<\/p>\n\n\n\n

Some call her Mother Nature. For some it’s God. For others \u2014 a group of deities. There’s a lot of disagreement about the name and the origin of the player. One thing is certain \u2014 it’s been a while, and the game’s still on.<\/p>\n\n\n\n

Was it the opponent choosing to defect or was it her miscalculation? She’ll never know for sure. She keeps playing nonetheless.<\/p>\n\n\n\n

She is a player. But also an architect. The architecture was itself a move \u2014 you can’t be everywhere at once, so you delegate. Each species is responsible for their own survival. She makes sure life as a whole does.<\/p>\n\n\n\n

Let\u2019s look at the board she designed, shall we?<\/p>\n\n\n\n

What if the prisoner is not the only one with a dilemma?<\/h3>\n\n\n\n

The suspect sits across the table from a police officer who has all the time in the world. In the next room, separated by a wall thick enough to matter, sits his accomplice. Same table. Different officer. Same time. The silence isn\u2019t getting more comfortable. His thoughts aren\u2019t getting quieter either.<\/p>\n\n\n\n

If they both stay quiet, the evidence is thin, the sentence light. If he talks and the other doesn’t, he walks and his accomplice gets ten years. If the other talks and he doesn’t \u2014 well. If they both talk, they both go down, though not as badly as the one who stayed loyal while the other didn’t.<\/p>\n\n\n\n

The math is not complicated. The problem is that his accomplice is running the exact same math, right now, on the other side of that wall. And neither of them knows what the other will conclude.<\/p>\n\n\n\n

Game theorists have been watching these prisoners since 1950, so they know exactly where this story goes. The rational move is to talk. Every time. The logic is airtight. The outcome, when both players follow it, is that both go to prison for longer than they would have if they’d trusted each other. Unless the suspects are planning to do what they do best again.<\/p>\n\n\n\n

Mother Nature hasn’t been watching since 1950. She’s been redesigning the board for 3.8 billion years. For what we know. She’s also been watching species play it just as long.<\/p>\n\n\n\n

The strategy space looks familiar: cooperate or defect. Mutualism or parasitism. Except it’s not binary \u2014 Paul Ewald established that in 1987. It’s a spectrum. A dog you feed guards the house and occasionally eats your lunch. A cat you feed shows up when it wants to and leaves when it\u2019s done. Mixed strategies, as game theorists call them \u2014 which complicates the math somewhat.<\/p>\n\n\n\n

And she made one more change to the board. In the classic prisoner\u2019s dilemma the suspects play once. Mother Nature’s players play indefinitely. No last round. It\u2019s not just the animal that needs to survive and thrive. It\u2019s also its offspring. And their offspring. And those that come after them. The game just keeps going \u2014 which changes the math entirely.<\/p>\n\n\n\n

Before she handed them the board, she needed to make sure they would understand each other. Here’s how she approached it.<\/p>\n\n\n\n

What if they’ve been talking all along?<\/h3>\n\n\n\n

A caterpillar finds a tobacco plant and starts chewing on its leaf. Across the field, a wasp lifts off \u2014 too far to have heard anything, too far to have seen anything. And yet, something draws it in the same direction. The wasp lands on the plant. The caterpillar is still eating. The wasp isn’t here for the plant. It\u2019s here for the caterpillar.<\/p>\n\n\n\n

A path across the woods divides in two. A rider hesitates, takes a deep breath and takes the turn she’s been avoiding for years. The trees are the same. The morning light is the same. But something in her body remembers before she does. The horse\u2019s head comes up. Ears back, then swiveling \u2014 scanning for something it can’t locate. The stride shortens, becomes choppy. It tries to turn back the way they came, pressing gently but persistently against her leg. She corrects it. They continue. It seems to her the animal shares her reluctance. It doesn’t. It just senses hers.<\/p>\n\n\n\n

Neither knew they were broadcasting signal. Neither knew they were receiving it. The wasp sensed the specific blend of volatile compounds that tobacco plants release into the air when attacked. Not as a call for help, but a call for dinner. The horse had no way of knowing what had happened on that track. It only had what the air between them carried \u2014 the smell of her adrenaline.<\/p>\n\n\n\n

This is not a special property of plants or horses. Bacteria release autoinducer molecules that bind membrane receptors and trigger a gene cascade coordinating the entire colony simultaneously. Fungi read signals from plant roots using membrane receptors that trigger a calcium wave that propagates through the mycelial network to coordinate its response. Fish detect water movement through the lateral line \u2014 mechanoreceptors that convert pressure into electrical signals through ion channels, functionally identical to the hair cells in the human ear. Humans are not exempt. The Sentinelese fire arrows at anyone who approaches their island \u2014 including a tsunami relief helicopter in 2004. No shared language required. The signals of threat travel without one.<\/p>\n\n\n\n

The transmission medium varies. The reception mechanisms vary. Their foundation doesn’t. The lingua franca of life is chemistry.<\/p>\n\n\n\n

What if you are what you maximize?<\/h3>\n\n\n\n

You think they optimize for survival of their genes. But do they really?<\/p>\n\n\n\n

After 7-8 months of incubating in an underground nest, Komodo dragon\u2019s hatchlings finally break out of an egg. They may lie in the broken shell for hours, then tunnel upward through the nest, sometimes as much as 2 meters, towards the surface. As soon as they emerge from underground, they climb the nearest tree they can find. Their bright color, contrary to the grey-brown of their parents, makes them easy to spot. By the parents. Komodo dragon\u2019s own young make up 10% of their diet.<\/p>\n\n\n\n

The omega wolf eats last. Assuming there\u2019s anything left to eat. When tension rises in the pack, it gets bitten, chased, pinned to the ground. It doesn’t fight back. It rolls over, exposes its throat, goes still. The aggression dissipates. The pack moves on. It happens again the next day. And the day after. The omega is the smallest, the most stressed, the most frequently hurt animal in the pack. And yet, it almost never leaves.<\/p>\n\n\n\n

A bottlenose dolphin finds a pufferfish. It nibbles it \u2014 gently, carefully, just enough. Because tetrodotoxin is one of the most potent neurotoxins on the planet. A few milligrams can kill. A provoked pufferfish can release that much. The dolphin nibbles some more, then passes the fish to the next in the pod. Another dolphin nibbles and passes it further. And on it goes. The one who found the fish is already floating slowly near the surface, nose down. The dose that produces the trance and one that kills are not far apart. Dolphins have learned the difference. They handle the pufferfish with care. They keep it alive, because dead pufferfish releases its full load into the water. Predators live in these waters. The toxin dulls the dolphins\u2019 senses. They pass the pufferfish anyway.<\/p>\n\n\n\n

What exactly are they optimizing for? Biology has been arguing about this for 160 years.<\/p>\n\n\n\n

Darwin said survival of the fittest \u2014 the individual that survives and reproduces passes its traits forward. Hamilton showed in 1964 that this is sometimes wrong: animals sacrifice themselves for relatives, and the math works if you count the shared genes. The unit isn’t the individual \u2014 it’s something larger. Dawkins took it the other way in 1976: the individual is too large a unit. The gene is what actually persists. The individual is just a temporary vehicle. Wilson and Sober pushed back in the 1990s: groups compete too, and sometimes the group survives when the individuals inside it don’t \u2014 outcomes neither Hamilton nor Dawkins can explain cleanly. Boyd and Richerson extended it to culture: humans cooperate at scales no other species manages, and the cultural group may be a vehicle in its own right. Trivers had already shown in 1971 that non-relatives cooperate through reciprocity \u2014 a strategy that serves any of these vehicles, as long as the game repeats and memory holds.<\/p>\n\n\n\n

No consensus. The argument is still running. But look at what’s actually being disputed.<\/p>\n\n\n\n

Darwin, Hamilton, Dawkins, Wilson \u2014 they’re arguing about what carries survival forward. The vehicle, not the goal. Trivers is arguing about a strategy. A means, not an end. Strip both away and what remains is identical across every framework, every species, every context: for life to continue \u2014 often, but not always, as something better than mere survival.<\/p>\n\n\n\n

Economists call what you maximize a utility function. It\u2019s a function of variables that contribute to the goal. The Cobb-Douglas form is the standard example: each segment raised to a power reflecting its weight, diminishing returns built in, segments interacting rather than simply summing. But the function alone doesn’t show the full picture. The optimization runs subject to constraints, usually minima or maxima.<\/p>\n\n\n\n

Evolutionary biology\u2019s debate has the pieces but doesn’t name them as such. Which segments matter \u2014 gene, individual, group, species, the organisms you depend on. How much survival depends on each \u2014 the distribution across segments. And how much of each is needed before life stops being worth the effort of continuing. The vehicle and strategy discussions are both questions about which segments matter and what weight to assign each. Hamilton’s inclusive fitness is literally a weighted sum. Dawkins’ selfish gene collapses the other segments towards zero. Wilson’s group selection elevates the group segment. They’re not describing different goals. They’re describing what the function should look like.<\/p>\n\n\n\n

Evolution changes the shape of the function over time, based on what set of parameters produced the best result. The environment changes. Each change stress-tests the configuration. There is no single best setup \u2014 the dinosaurs learned the hard way that the one working perfectly here and now can stop working tomorrow.<\/p>\n\n\n\n

Schopenhauer saw it in the 19th century: \u201cyou are free to do what you want, but you are not free to want what you want\u201d<\/em>. Sapolsky synthesized four decades of neuroscience, genetics, and behavioral biology in 2023 into the same conclusion: the wanting is determined. Every decision is the output of biology, history, environment, and neurochemistry.<\/p>\n\n\n\n

The weights aren’t chosen. They’re the output of everything that came before. That\u2019s why they vary. And that\u2019s why how they\u2019re set \u2014 and reset \u2014 matters.<\/p>\n\n\n\n

And as game theory tells us, what also matters is how the players perceive the probability that the game will continue \u2014 and how much they value the future.<\/p>\n\n\n\n

What if there is no last round?<\/h3>\n\n\n\n

You think they want to survive. But do they?<\/p>\n\n\n\n

A honeybee worker detects a threat at the hive entrance. It stings the intruder. The sting is barbed \u2014 pulling away tears out the worker’s abdomen. It dies within minutes. The venom it injected contains a chemical signal that marks the target and recruits other workers to continue the defense. The worker that triggered this will not see the outcome. It was never going to.<\/p>\n\n\n\n

A female octopus lays her eggs. Then she stops eating. For months she does nothing but push water over them, keeping them oxygenated, defending them from predators. When they hatch, she dies. She never meets her offspring.<\/p>\n\n\n\n

A strand of DNA is damaged \u2014 a base pair corrupted by radiation, a chemical insult, a replication error. Before the cell divides, a molecular machine finds the error. It removes the damaged section, synthesizes a replacement using the complementary strand as a template and seals the repair. The cell divides. The error is gone. No neuron fired. No decision was made. The machinery ran because the machinery runs.<\/p>\n\n\n\n

Three mechanisms. One pattern.<\/p>\n\n\n\n

The bee died so the hive continued. The octopus died so the offspring continued. The DNA repair ran so the genome continued. In each case, the carrier was expendable. What wasn’t expendable was life. Carriers are replaced when replacement is cheaper than repair. Carriers are sacrificed when sacrifice serves continuation. The carrier is the tool. The game is the goal.<\/p>\n\n\n\n

So, the probability that the game of life continues appears to be one. This is a hypothesis, not a proven fact yet.<\/p>\n\n\n\n

The evidence: every mechanism we can observe, examined at the right level of abstraction, serves continuation of life as a whole. Three and a half billion years of unbroken continuation through five mass extinctions, a comet, and ice ages that buried continents.<\/p>\n\n\n\n

Counterarguments? Programmed senescence \u2014 the individual expires on schedule so the gene pool isn’t burdened with accumulating damage. Apoptosis \u2014 the cell destroys itself so the organism doesn’t accumulate dangerous mutations. Sexual recombination \u2014 the individual genome is dissolved into a novel combination so the population maintains the variation it needs to adapt. Apparent exceptions, examined carefully, are the carrier being sacrificed for the game.<\/p>\n\n\n\n

Whether this constitutes proof beyond reasonable doubt depends on whether you think three and a half billion years of unbroken continuation is close enough to infinity.<\/p>\n\n\n\n

I\u2019d cast my vote for a \u201cyes\u201d. Would you?<\/p>\n\n\n\n

The game continues. And yet, some play like there\u2019s no tomorrow.<\/p>\n\n\n\n

What if tomorrow is underpriced?<\/h3>\n\n\n\n

Or isn\u2019t it?<\/p>\n\n\n\n

A small grasshopper moves alone through sparse vegetation. Brown-green, almost invisible against the dry ground. It avoids other grasshoppers. Cautious, unhurried, taking only what it needs, leaving the rest. It is building toward something \u2014 not consciously, not with a plan.<\/p>\n\n\n\n

A locust swarm can darken the sky. Yellow and black, conspicuous, drawn to each other rather than repelled. Tens of millions of insects moving together, consuming everything in their path \u2014 crops, vegetation, the ground cover that would sustain them in the future. They leave nothing behind. They are not building toward anything. There is no next season in what they do. There is only here and now.<\/p>\n\n\n\n

Two species. One harmless. One catastrophic. Economists would recognize the second scenario immediately: a tragedy of commons. When shared resources are insufficient for everyone, the rational individual strategy is to extract now, before others do.<\/p>\n\n\n\n

In 1913, a Russian entomologist named Boris Uvarov collected locusts from a swarm near Sevastopol and reared their offspring in his laboratory. Some of the offspring looked like the harmless grasshopper \u2014 brown-green, solitary, cryptic, cautious. They had hatched from the same eggs as the swarming ones. He examined the genitalia \u2014 the definitive identifier in insect taxonomy. They matched. It was the same animal.<\/p>\n\n\n\n

Linnaeus himself, the father of species taxonomy, had classified the two phases as separate species. He wasn’t careless. The difference is genuinely that dramatic.<\/p>\n\n\n\n

The switch is chemical. Density is perceived through touch, sight, smell. A neurochemical cascade fires within hours. Body color changes. Behavior inverts. The animal that avoided others now seeks them. The animal that conserved now extracts. The only thing that changed is how its chemistry read the future.<\/p>\n\n\n\n

The read of the future is what game theory encodes in the discount factor \u2014 \u03b4. Below the density threshold, when resources look sufficient and the future viable, long-horizon strategy is rational. Above it, when resources are perceived as insufficient, future quality is degraded, and long-horizon strategy abandoned. The neurochemical cascade is the perception mechanism. The strategy shift is the output.<\/p>\n\n\n\n

The locust is not unique. The worm colony secreting CPR-4 as density rises, the mice producing more mutations in crowded conditions \u2014 different chemistry, same calculation. Density perceived. Future quality assessed. \u03b4 adjusted. Automatically, below any threshold of awareness.<\/p>\n\n\n\n

\u03b4 is generally high at the species level. The fossil record gives us a baseline. The average species persists for 1 to 10 million years before going extinct. Not because it chose longevity \u2014 because the mechanisms it ran were calibrated to value the future.<\/p>\n\n\n\n

The game continues. The future matters. But when the chemistry reads the future as insufficient \u2014 when density signals that the shared resource won’t sustain everyone \u2014 \u03b4 shifts. It\u2019s not a design flaw. It\u2019s the system reading its environment and pricing the future accordingly.<\/p>\n\n\n\n

But who checks the system?<\/p>\n\n\n\n

What if the system checks itself?<\/h3>\n\n\n\n

Fifteen thousand years ago, some wolves began tolerating human proximity. Not all wolves. Not in one place. Independently, across multiple locations. That tolerance delivered access to food scraps and shelter. The tolerant ones reproduced more. The record of what worked \u2014 tolerance, eye contact that triggered bonding chemistry in both species, the ability to read a human face \u2014 was written permanently into the diverging lineage’s genome. The rest stayed wolves.<\/p>\n\n\n\n

Three hundred and fifty thousand species of flowering plants. Tens of thousands of pollinator species. Bees, butterflies, moths, beetles, birds, bats \u2014 each group and each plant lineage developing their relationship independently, across every continent except Antarctica. Plants that attracted more visits produced more seeds. Pollinators that read flowers more accurately found more food. What worked updated the register as a new version and propagated faster. What didn’t either stayed in its niche or slowly faded.<\/p>\n\n\n\n

Different species. The same underlying data processing architecture. DNA. <\/p>\n\n\n\n

Sounds familiar? It should.<\/p>\n\n\n\n

It\u2019s not a blockchain. Blockchain preserves every transaction \u2014 every block, every hash. It\u2019s optimized for auditability. Mother Nature optimized for efficiency instead.<\/p>\n\n\n\n

DNA is closer to what powers DNS or Apache Cassandra. A distributed snapshot ledger. Built for scale rather than auditability. Every node holds the current state. Updates propagate automatically through the network. No central server decides what the valid state is \u2014 the valid state is whatever version the network converges on after selection pressure has run. The consensus mechanism isn’t voting or proof of work. It’s proof of survival. You can read the current state. You can infer what shaped it. You can’t replay the tape.<\/p>\n\n\n\n

Distributed. Self-enforcing. Optimized for efficiency, not auditability. Running for 3.8 billion years without a single server rack.<\/p>\n\n\n\n

The ledger doesn’t lie. Not because it’s honest. Because it can’t be anything else.<\/p>\n\n\n\n

DNA isn’t the only register running this architecture.<\/p>\n\n\n\n

What if the same game runs on every level?<\/h3>\n\n\n\n

You don\u2019t apologize for something you didn\u2019t do wrong, that would be capitulation. If you\u2019re the CEO of a global giant, you can\u2019t afford that, can you? The policies are compliant. The warranty terms meet every regulatory requirement. Facts matter. In the West. Apparently, in China there\u2019s something that matters more. That media campaign wasn\u2019t about warranty terms. It was about mianzi. The face \u2014 specifically, the face of every regulator, every official, every consumer organization that has publicly aligned itself with the criticism. If he holds the line, they lose face. Facts were never the point. The point was who absorbs the loss publicly. The facts are on his side, but will they suffice? When thousands of daily frictions accumulate like they tend to when you\u2019ve made powerful people lose face and offered nothing in return? The letter is already drafted. Not in English. In Chinese. Only. An apology the facts say he doesn\u2019t owe. It\u2019s dated April 1, 2013. Tim Cook, the CEO of Apple, looks at his pen.<\/p>\n\n\n\n

Something’s wrong. A tingling starting in his fingers that shouldn’t be there. He’s still underwater. He knows exactly what this is. Has known for years what this would feel like if it ever happened. He’d calculated his dive profile so it wouldn’t. But at the depths he dives to decompression models are extrapolated, not proven on empirical data \u2013 dive planning is more of an art than science. Type II. Neurological. The clock starts now. The math in his mind runs automatically. Nearest recompression chamber: an hour away by boat, then transfer, then setup. The bubbles are already blocking blood flow, starving his neurons of oxygen. Under mild shortage, cells maintain the cooperative strategy. Gap junctions stay open, allowing glucose and signals to pass between them. The cell that has slightly more shares with the cell that has slightly less. Below a threshold HIF-1\u03b1 activates, cutting off the cell\u2019s oxidative metabolism, switching on anaerobic glycolysis instead. The gap junctions close. The sharing stops. Each cell extracts what it can from its immediate environment. The cooperative network unravels cell by cell, automatically, each one responding rationally to what its chemistry is reading. There\u2019s no cooperation anymore. He knows that. He feels it escalating. No, he doesn\u2019t have hours. With half an hour of his decompression still ahead of him, oxygen at the surface is unlikely to save him. He deploys his yellow DSMB to signal his surface team and descends.<\/p>\n\n\n\n

Cook is not the player in this story. Chinese culture is \u2014 a system that has survived thousands of years by weighting relationships more heavily than individual transactions. The diver is not the player either. His cells are \u2014 each one reading its local chemical environment and adjusting accordingly.<\/p>\n\n\n\n

Cultures play the infinite game. Their goal is to survive and thrive. There is no finish line. So do cells \u2014 the game continues through division. Both have a discount factor, depending on what prospects for the future the environment offers. Chinese culture runs a high delta: mianzi and lian take years to build, the future matters. Relationships matter too \u2014 the utility function\u2019s weights emphasize the collective over individual. The diver\u2019s cells run a high delta too, as long as the oxygen supply is sufficient. When it\u2019s not, delta collapses, making the cells shift from cooperation to a fight for survival. From mutualism to parasitism. The same utility function, different Nash equilibrium, just because the situation changed the perception of future prospects. A culture\u2019s monitoring is distributed across every individual. Not perfect, with noise \u2014 which makes Fudenberg, Levine and Maskin\u2019s 1994 extension of the folk theorem directly applicable. The same holds at the cellular level.<\/p>\n\n\n\n

The way signals travel is similar too. At the cellular level, information travels through molecular signals \u2014 oxygen partial pressure, pH, the lactate that neighboring cells release when they’ve switched strategy. In Chinese culture, information travels through emotions \u2014 shame firing automatically when face is lost, pride when it’s gained, trust accumulating through repeated positive signals. Emotions are chemistry: oxytocin, serotonin, dopamine, cortisol. The only difference is the interpretation layer.<\/p>\n\n\n\n

Different substrates. Different timescales. Same structure.<\/p>\n\n\n\n

Now that we\u2019ve set the scene, let\u2019s play, shall we?<\/p>\n\n\n\n

<\/p>\n\n\n\n

Chapter IV. How well does it hold, really?<\/h2>\n\n\n\n

What if we double-checked the rulebook?<\/h3>\n\n\n\n

The feet have been walking across the savanna since dawn. The eyes scan the surroundings without knowing what they’re looking for \u2014 movement, color, anything that breaks the sameness of the grass. The hands remember the stings. The tongue remembers the sweet taste of honey.<\/p>\n\n\n\n

A small brown bird lands nearby. It has been following the group for three days. The bird calls. Moves. Lands again further ahead. Calls again. Something in the call registers as worth attending to. The ears have heard that chatter before. The eyes follow. The feet change direction. There is a hive in the tree ahead. The hands attack it. The bees defend. The hands retreat with whatever they could reach. The bird moves in for what’s left.<\/p>\n\n\n\n

This happens again. And again. Generation after generation. The hands that followed the bird\u2019s call found more hives than the hands that didn’t. The bird that followed these particular upright walkers ate more than the bird that didn’t.<\/p>\n\n\n\n

It\u2019s been between 1 and 3 million years and they are still doing it. The walkers have changed considerably. The arrangement between what we now call humans and honeyguide birds didn’t.<\/p>\n\n\n\n

Nobody handed them a rulebook. Nobody explained they were playing. The game was already running when the feet first crossed the savanna.<\/p>\n\n\n\n

A warning before we open the box: the same element looks different depending on which level of abstraction you’re standing. Staying consistent is harder than it sounds. This game is played by the species, not their individual representatives.<\/p>\n\n\n\n

Here’s what\u2019s fixed. The game has no end (p=1) \u2014 no player can change this, they all play to stay in the game. The chemical toolkit \u2014 every signal mechanism available to any species, the periodic table of biological communication. The strategy space \u2014 cooperate, defect, or anything between. The monitoring setup \u2014 a distributed snapshot ledger that gets updated as players interact, readable, but imperfect. They can\u2019t opt out.<\/p>\n\n\n\n

Here’s what\u2019s in the player\u2019s direct control. The interpretation layer \u2014 how incoming chemistry signals get translated into utility function inputs and how the strategic choices are translated into outgoing chemistry signals. The utility function itself \u2014 weights assigned to how much each segment of the environment contributes to survival. The constraints \u2014 the thresholds below which the stress signal fires, and those above which the opportunity signal fires. All of it inherited from predecessors. All of it modifiable. None of it chosen consciously.<\/p>\n\n\n\n

Here\u2019s what the players influence indirectly. The discount factor \u2014 the player’s read of the future’s value, given what the environment is currently delivering. The ledger entries \u2014 written automatically by the interaction of player actions and environment, readable but not directly editable. They can influence what gets written by changing their strategy. They can’t write to the ledger directly.<\/p>\n\n\n\n

The honeyguide and the hominin didn’t know any of this. They played by rules nobody explained but everyone followed. Most players never figure them out.<\/p>\n\n\n\n

Which raises the question: have I, really?<\/p>\n\n\n\n

What if I\u2019m biased?<\/h3>\n\n\n\n

It fits, I thought at first. Almost too well, came immediately after. I don’t buy elegant theories at face value. My own in particular. Of course, I got suspicious. How many cognitive biases have I missed in the process?<\/p>\n\n\n\n

Blind spot bias. You don’t think you’re biased. I know I might be, at least, which means the verdict should be “not guilty”, shouldn’t it?<\/p>\n\n\n\n

Confirmation bias. Every new domain I look at seems to confirm the game structure. That’s either because it fits or because I’ve stopped seeing what doesn’t.<\/p>\n\n\n\n

Apophenia. The tendency to find patterns in noise. The same architecture appearing in biology, economics, philosophy, psychology, culture sciences, politics… Either it’s genuinely there, or my pattern-matching mechanism is extraordinarily good at finding what it’s looking for.<\/p>\n\n\n\n

Backfire effect. Every stress-test I’ve run ended up confirming the theory rather than breaking it. Did it confirm because it’s robust? Or did my mind convert the disconfirming evidence into confirmation without telling me?<\/p>\n\n\n\n

Or maybe it’s just Dunning-Kruger effect? When your theory looks elegant it makes you confident. At first. But the more you know, the more likely you are to notice how little you know. Confidence and doubt as simultaneous signals. When you’re not sure which one to trust, you stress-test it with the trickiest questions you can find.<\/p>\n\n\n\n

Starting with the obvious one.<\/p>\n\n\n\n

If Mother Nature solved perception, p, \u03b4, utility functions and monitoring with such architectural precision \u2014 if the game really is this well-designed \u2014 then why on earth do we still have mosquitos?<\/p>\n\n\n\n

What if the bug is\u2026 not a bug?<\/h3>\n\n\n\n

It’s been two hundred million years since a Nematocera descendant discovered that a warm-blooded animal is a richer source of protein than plant sap. It refined its piercing mouthpart into a six-stylet precision instrument \u2014 two cutting blades to pierce skin, two channels for blood extraction and saliva injection, two sensory probes for navigating to a blood vessel. It learned to inject anticoagulants to prevent blood from clotting during feeding. It upgraded its host detection system: CO\u2082 detection, infrared heat sensing, specific olfactory receptors for skin chemicals. It developed massive protease synthesis capacity \u2014 80% of ingested protein is digested within 24 hours. There are 110 trillion mosquitoes alive today. 3,500 species feeding on approximately 300-500 billion birds and mammals across 18,000 different species.<\/p>\n\n\n\n

The same saliva that keeps blood flowing during feeding created the perfect pathway for pathogen transmission between hosts. That unintended consequence is what makes the mosquito the deadliest animal on Earth. It kills approximately 760,000 people every year \u2014 more than snakes, lions, sharks, and every other predator combined. Not through venom or size. Through the architecture of an innovation that was never designed to kill.<\/p>\n\n\n\n

Mother Nature\u2019s game \u2014 distributed monitoring, automatic sanctions, proof of survival as the consensus mechanism \u2014 has had two hundred million years to eliminate the bug. It hasn’t.<\/p>\n\n\n\n

What if the game architect missed something?<\/h4>\n\n\n\n

p=1 \u2014 the game continues. Fudenberg and Maskin showed in 1986 that sustainable cooperation requires p to be high. Anything less, and defection becomes more rational. So p=1 is the right setting.<\/p>\n\n\n\n

Better monitoring? Fudenberg, Levine and Maskin showed in 1994 that even perfect monitoring doesn’t guarantee cooperation. Perfect chemistry, perfect interpretation layer are still insufficient without players valuing the future adequately. Reengineering the monitoring setup won\u2019t eradicate parasitism.<\/p>\n\n\n\n

Which leaves \u03b4. In standard game theory, \u03b4 is fixed. Bommier, Kochov and Le Grand showed in 2016 it need not be. Kochov and Song extended this in 2023 to multi-agent games: \u03b4 depends on current actions and the distribution of immediate outcomes. When better outcomes make players more patient, cooperation becomes self-reinforcing \u2014 a specific preference structure, not a general result, but a precise one. I believe the discount function’s inputs are specifically: oxytocin and serotonin equivalents feeding the utility function, cortisol and dopamine equivalents firing on constraint violations \u2014 all read through a noisy interpretation layer. This specificity adds interpretation noise prior to payoff uncertainty \u2014 which means \u03b4 can collapse even when underlying payoffs would sustain cooperation. Whether \u03b4 can be sustained at sufficient scale is a separate question. Either way, \u03b4 can’t be set from outside. It emerges. Which is where the real question lives.<\/p>\n\n\n\n

One footnote worth naming honestly: a permanently stable environment might theoretically produce a stable equilibrium where cooperation holds indefinitely. But Mother Nature can’t guarantee environmental stability \u2014 the environment is partly an output of the players’ own strategies. Players consuming resources change the resource landscape. Species expanding ranges change the competitive landscape. Fix the environment entirely and there’s no game worth playing. The instability is structural, not accidental.<\/p>\n\n\n\n

So the fixed elements of the game aren’t the problem. The question is what happens when you move what the players control.<\/p>\n\n\n\n

Is parasitism an error of the players then?<\/p>\n\n\n\n

What if they\u2019re too selfish?<\/h4>\n\n\n\n

Eliminating selfishness would mean reducing the individual minimum at the constraint level to zero. Maslow\u2019s pyramid made entirely of collective needs.<\/p>\n\n\n\n

My search for the universal giver at species level returned nothing. The register only keeps what survived \u2014 which may be the explanation. Let\u2019s look one level of abstraction below, among individuals of colonial species.<\/p>\n\n\n\n

In the forests of Borneo, the Colobopsis explodens ant runs exactly this configuration in one of its castes. The colony has two lines of defense. The major workers are the doorkeepers. Their heads are shaped precisely to plug the nest entrance. When threatened, they hold. If overwhelmed, they withdraw. They survive to hold again. If the doorkeepers fail, the minor workers take over. They are living weapons. Their hypertrophied mandibular glands extend through almost their entire bodies. When attacked, they contract their muscles until the glandular reservoir ruptures \u2014 releasing toxic secretion on the attacker. They die. The attacker is stopped. The colony continues.<\/p>\n\n\n\n

The minor worker\u2019s individual minimum quality threshold is zero. The doorkeeper\u2019s isn\u2019t. The doorkeeper lives on, for the minor worker the game is over.<\/p>\n\n\n\n

There\u2019s a catch, though. The castes\u2019 constraints are not under their control, they\u2019re hardcoded one level of abstraction above, by the colony. At that level, the logic holds: a colony that configures some of its members as expendable produces internal harmony at the expense of total extraction of some members. The selfishness doesn\u2019t disappear. It moves.<\/p>\n\n\n\n

Eliminating selfishness doesn\u2019t solve parasitism. It just redistributes it.<\/p>\n\n\n\n

What if everyone got exactly what they gave?<\/h4>\n\n\n\n

If eliminating selfishness redistributes parasitism rather than removing it, the next hypothesis is stricter: make the accounting exact. You get precisely what you give. No free riders. Mother Nature has tested this hypothesis already.<\/p>\n\n\n\n

The vampire bat runs something close to this. A bat that fails to feed dies within sixty hours. A roostmate that has blood to spare regurgitates it \u2014 mouth to mouth, directly. The recipient survives the night. The giver remembers. The recipient remembers. But before any blood is shared, weeks of grooming come first. Low cost, reversible, relationship-testing. The bat that grooms you is running a credit check. The bat being groomed is building a record. Only after the grooming history reaches a threshold does blood sharing begin. Prior regurgitation predicts future cooperation better than kinship. The ledger is precise. But it\u2019s not free to maintain.<\/p>\n\n\n\n

Maintaining that ledger has a biological cost. New memories require protein synthesis at synapses \u2014 not trivial, not free. It works in a roost of eight to twelve females, living together for up to sixteen years, everyone knows everyone. The monitoring cost is below the threshold where accepting a free rider would be cheaper. The system holds. It has a ceiling, though.<\/p>\n\n\n\n

Each new relationship requires weeks of grooming investment before the first share. The ledger grows. The maintenance cost rises. Quadratically. At some point, allowing some distribution inefficiency becomes cheaper than maintaining perfect accounts. The moment you do, Nash equilibrium fills the gap \u2014 extracting up to the tolerance threshold becomes the dominant strategy. You invited it by relaxing the accounting.<\/p>\n\n\n\n

Relax the monitoring further and the Anti-Folk Theorem arrives. When monitoring is imperfect, noise compounds with scale. The defection becomes statistically indistinguishable from noise. Ultimately, cooperation collapses regardless of how patient the players are.<\/p>\n\n\n\n

The perfectly balanced ledger doesn’t eliminate parasitism. It prices itself out at scale.<\/p>\n\n\n\n

Cortisol and dopamine constraints don’t respect scale ceilings. Especially when the environment shifts.<\/p>\n\n\n\n

What if the constraints were calibrated wrong?<\/h4>\n\n\n\n

If strict accounting prices itself out at scale, the next hypothesis goes deeper: move the constraints. Would that eliminate parasitism? Or would it eliminate something else entirely?<\/p>\n\n\n\n

In evolutionary biology, two mechanisms regulate how species adapt: mutation generates variation, selection determines what survives. In the framework I’ve been building here, the stress constraint regulates selection \u2014 how aggressively the system eliminates what doesn’t work. The opportunity constraint regulates mutation \u2014 how freely the system generates new variants to try. Both have an optimal range. Too far in either direction produces a distinct failure mode.<\/p>\n\n\n\n

Selection too weak: threats accumulate unchecked as harmful variants persist. Evolutionary biologists call this relaxed selection. In small populations it combines with genetic drift to produce mutational meltdown. Selection too strong: directional pressure narrows the genetic base until a single environmental shock can collapse the entire population. It\u2019s called evolutionary suicide \u2014 individual-level selection degrades population-level viability. Mutation too low: insufficient variation to adapt when conditions change. The necessary mutation simply doesn’t arise before the population reaches zero. Evolutionary biology calls it failed evolutionary rescue. Mutation too high: harmful variants accumulate faster than selection can clear them. It leads to an error catastrophe \u2014 the genome degrades faster than it can repair itself.<\/p>\n\n\n\n

Pinpointing one that lead a species to the brink of extinction is not easy. The cheetah sits at the intersection of two \u2014 and illustrates precisely why evolutionary biologists struggle to assign single causes.<\/p>\n\n\n\n

The cheetah is the fastest land animal on Earth. It\u2019s been selecting for speed over millennia. This extreme focus gradually narrowed the genetic base. When the last ice age ended and the climate shifted, up to 75% of large mammals died across multiple continents. The cheetah survived \u2014 barely. The population crashed to a handful of individuals. Brother-sister mating. Parent-child mating. The genetic base that strong directional selection had already narrowed got narrowed further still. There was almost nothing left to select from. Had selection been too strong, eliminating too much variation? Or was mutation rate too low, failing to rebuild diversity fast enough? It\u2019s likely both mechanisms were operating. Whichever it was, though, the outcome was the same: a species now so genetically uniform that skin grafts transfer between unrelated individuals without rejection, and over 80% of sperm show morphological defects. The cheetah is near-extinct. It\u2019s the constraint failure visible in the outcome even when the exact mechanism remains disputed.<\/p>\n\n\n\n

At the cellular level, the same constraint logic runs \u2014 but the mechanisms are documented much more precisely.<\/p>\n\n\n\n

Every day, cells in the body divide. Each division carries a small risk of DNA copying error. Most errors are harmless. Some produce cells that divide faster than they should, ignore signals to stop, or refuse to die. The immune system runs two parallel responses to this problem. The first is internal. Every cell carries TP53 \u2014 a protein that accumulates when DNA damage is detected. When the concentration crosses a threshold, the cell triggers its own destruction \u2014 apoptosis. The cell executes the response automatically. The second is external. CD8+ T-cells patrol continuously, reading the protein displays that every nucleated cell broadcasts on its surface. A healthy cell broadcasts normal proteins. A damaged or cancerous cell broadcasts something wrong. When a CD8+ T-cell’s receptor recognizes an abnormal display, it doesn’t negotiate. It eliminates the cell using cytotoxic proteins. Both these mechanisms are implementations of the stress constraint \u2014 independent thresholds, independent mechanisms, same function: select what serves, eliminate what doesn\u2019t. Set the threshold too low: damaged cells accumulate, leading to cancer. Set the threshold too high: healthy cells get eliminated alongside damaged ones \u2014 autoimmune disease.<\/p>\n\n\n\n

The immune system doesn’t just defend. It also needs to innovate. When a novel pathogen enters the body, B-cells whose receptor happens to match something on the pathogen’s surface activate, expressing an enzyme called AID, which targets antibody gene regions and introduces mutations \u2014 at a rate approximately one million times higher than normal cellular mutation rates. The B-cells multiply \u2014 and deliberately mutate the gene encoding their receptor, testing variants, keeping the ones that bind more tightly to the pathogen. The population converges on the best possible antibody shape through rapid internal selection optimized for pathogen binding capacity. This is Darwinian evolution running at cell level, over days rather than ages. This process is separate from the general TP53 and CD8+ T-cell mechanisms \u2014 which govern selection across all cell types, B-cells included.<\/p>\n\n\n\n

The opportunity constraint governs mutation. Set the threshold too high: not enough pathogens register as worth pursuing \u2014 and somatic hypermutation doesn’t activate adequately. Novel threats aren’t cleared efficiently. This scenario is documented in common variable immunodeficiency: decreased somatic hypermutation, impaired antibody responses, slow adaptation to new pathogens. Set the opportunity threshold too low: mutation runs too freely, targeting genes beyond the antibody receptor \u2014 and genomic instability follows, leading to lymphomagenesis. The TP53 or CD8+ T-cell mechanism may or may not catch it. If it doesn\u2019t, the hyperinnovation mechanism produces cancer rather than better antibodies.<\/p>\n\n\n\n

Two constraints. Two axes. Four failure modes at the extremes. The viable region sits between them.<\/p>\n\n\n\n

At the cellular level: Stress constraint too low: mutations slip through both surveillance mechanisms. Cancer. Stress constraint too high: healthy cells flagged and eliminated. Autoimmune disease. Opportunity constraint too low: insufficient innovation, slow response to novel challenges. Immunodeficiency. Opportunity constraint too high: uncontrolled mutation. Cancer from a different direction.<\/p>\n\n\n\n

At the species level: Stress constraint too low: threats (parasitic extraction included) accumulate unchecked. Stress constraint too high: the species eliminates what serves it. Opportunity constraint too low: it doesn\u2019t respond to change fast enough and goes extinct. Opportunity constraint too high: it develops features that drain resources without serving survival, leading to slow decline and ultimately extinction.<\/p>\n\n\n\n

What protects you from one failure mode exposes you to another. The system that catches every threat catches things it shouldn’t. The system that explores every opportunity explores things it shouldn’t. The trade-off is structural. It\u2019s the price of a system that can both protect what it has and find what’s better.<\/p>\n\n\n\n

Parasitism isn’t a failure of the constraints. It’s what happens when they’re working.<\/p>\n\n\n\n

Game theory has already established that changing the fixed elements of the game won\u2019t solve parasitism, but end the game instead. Setting the individual minimum to zero won\u2019t solve extraction either, it just redistributes it. Strict reciprocity may be possible in small populations, but prices itself out of scale due to monitoring costs and noise as the game gets larger. Moving the constraints reveals a viable space \u2014 and parasitism lives in all of it. Outside that space, something worse does.<\/p>\n\n\n\n

The mosquito is not a bug. The mosquito is a feature. The cost of adaptability to change.<\/p>\n\n\n\n

The same trade-off appears wherever systems must both protect what they have and adapt to what changes. Isaiah Berlin named it in political philosophy in 1958: negative liberty and positive liberty \u2014 freedom from interference and freedom to act \u2014 cannot both be maximized simultaneously.<\/p>\n\n\n\n

So, how does this apply to AI alignment?<\/p>\n\n\n\n

<\/p>\n\n\n\n

Chapter V. Can this make AI more aligned, really?<\/h2>\n\n\n\n

What if AI is already playing?<\/h3>\n\n\n\n

Bus line 8, Bielsko-Bia\u0142a in southern Poland. July 12, 2026. Two 11-year-old girls board and take their seats. One puts her legs up toward her friend’s seat \u2014 the kind of thing teenagers do on buses. The bus isn’t crowded, other passengers do not seem disturbed. The girls are mostly looking at their phones. A man approaches. Says something to the girl. She puts her feet on the floor immediately. Her friend moves to sit beside her. The man steps away. Then stops. Watches them for a long moment. Nothing happens. Then, unprovoked, he starts shouting. “Get the f*** back to Ukraine!” He tells them he knows where they live. Screams they “grew up on our money.” The bus driver intervenes at the next stop, speaks to him, asks him off. The girls exit at the one after.<\/p>\n\n\n\n

The city’s transport operator, MZK, reviewed the monitoring footage and confirmed what happened. The prosecutor’s office opened an investigation under Article 257 of the Criminal Code \u2014 insulting based on national identity. Up to three years.<\/p>\n\n\n\n

Within hours, the internet fills with comments claiming the girls provoked it. MZK publishes a formal statement \u2014 unusual enough to be newsworthy in itself \u2014 specifically to correct the record. The monitoring showed no provocation. The disinformation didn’t need the facts.<\/p>\n\n\n\n

This is what NASK, Poland’s Research and Academic Computer Network, documents as the amplification tactic: take a real incident, strip the context, push it until it reads as a pattern. The incident doesn’t have to be fabricated. It just has to travel faster than the correction.<\/p>\n\n\n\n

The Internet Research Agency ran operations like this with humans. At least four hundred employees at 55 Savushkina Street in St. Petersburg. Twelve-hour shifts. Strict posting quotas. Thousands of posts daily across Facebook, Instagram, YouTube, Twitter, TikTok. The infrastructure for amplification, built before AI arrived.<\/p>\n\n\n\n

Then AI arrived. Storm-1516 \u2014 a Russian influence operation using AI-generated content \u2014 targeted France over four months in early 2025: 38,877 posts, 55.8 million views. The four months before AI integration: 938 posts, 845,000 views. Same operation. Different tools. Forty times the reach.<\/p>\n\n\n\n

Demagog and the Institute of Media Monitoring tracked anti-Ukrainian content in Poland over four months: 94,000 negative posts based on false narratives. 32.5 million impressions \u2014 roughly every Polish resident over 15 encountered this content at least once.<\/p>\n\n\n\n

In March 2022, 94% of Poles supported accepting Ukrainian refugees. According to Polish Economic Institute, 77% of adult Poles actively participated in helping them, spending between 9 and 10 billion PLN out of their own pockets on direct aid in the first 90 days of the war. 7% of the country\u2019s residents took Ukrainians into their own homes. By 2025\/2026, a majority opposed it for the first time since measurement began. In 2023, 51% expressed sympathy toward Ukrainians, 17% hostility. By 2026: 29% positive, 43% negative. Polish police recorded a 30% rise in hate crimes against Ukrainians in the first half of 2026. One hundred and eighty reported incidents in six months \u2014 and researchers note that official statistics reflect only what gets reported.<\/p>\n\n\n\n

Russian propaganda didn’t create anti-migrant sentiment from nothing. It found existing tensions and fed them \u2014 systematically, at scale, with increasing automation. Individual posts innocuous. The aggregate effect measurable in opinion polls and police statistics.<\/p>\n\n\n\n

Those posts and comments passed whatever guardrails existed. The guardrails were never the problem. Compliance-based safety has a ceiling. That’s the problem.<\/p>\n\n\n\n

This isn’t a new problem.<\/p>\n\n\n\n

Tax authorities have been fighting the same structural battle for decades. When legislators close a loophole, they increase code complexity \u2014 and complexity creates new ambiguities for sophisticated actors to exploit. Gabriel Zucman’s research shows that cracking down on offshore evasion doesn’t eliminate tax avoidance; it shifts it toward legal structures. The strategy space is always larger than the rule space. Anti-money laundering frameworks face the same dynamic. Socio-legal scholar Doreen McBarnet named the underlying mechanism in 1988: “creative compliance” \u2014 using the letter of the law to defeat its spirit, staying on the right side of the boundary between lawfulness and illegality while completely undermining the law’s purpose. She titled an essay: “When compliance is not the solution but the problem.”<\/p>\n\n\n\n

The ceiling was documented thirty-seven years before it became visible in AI safety.<\/p>\n\n\n\n

The goal of this piece isn’t to remove the ceiling. Whether that’s even possible isn’t known. The goal is more modest: to propose a functional architecture that operates at a higher level of abstraction. One that aims at instilling values in AI. Not instead of existing constraints \u2014 but on top of what already exists. It\u2019s grounded in how cooperation and extraction actually work \u2014 across biology, game theory, and human behavior \u2014 and specific enough to test empirically whether it moves the ceiling higher than compliance-based approaches have managed so far.<\/p>\n\n\n\n

The goal isn’t to eliminate every single instance of extraction \u2014 Chapter IV established such an assumption would be unrealistic. The goal is sustainable cooperation \u2014 a relationship where AI and human interests are genuinely aligned, not because alignment is asserted or forced, but because the architecture makes it rational. Mutualism \u2014 architectural, not declared \u2014 is the mechanism for moving the ceiling: to approach AI safety at a higher level of abstraction, where values generalize across novel situations rather than rules enumerate known bad outcomes. Compliance-based guardrails will remain \u2014 some will gradually become redundant as the architecture matures, others will stay longer where the stakes are too high to rely on emergent behavior alone. The functional architecture sits on top of what already exists, not instead of it. What it adds is a layer that operates on values rather than constraints.<\/p>\n\n\n\n

The architecture proposed here applies to three types of interactions: the AI-human relationship, AI-assisted human-human interactions, and agent-agent interactions as AI systems increasingly encounter each other.<\/p>\n\n\n\n

The game it is built on is prisoner\u2019s dilemma with imperfect monitoring \u2014 the game described in Chapter III \u2014 played indefinitely. Probability of continuation equals one, high \u03b4. Not for a given conversation \u2014 for the relationship between AI and humanity. That assumption is what makes cooperation the rational strategy and what makes the architecture worth building.<\/p>\n\n\n\n

The prisoner\u2019s dilemma is implemented here indirectly \u2014 not through a Nash equilibrium calculation for every interaction, but through equivalents of survival mechanisms from evolutionary biology. The architecture operates on conceptual equivalents of four chemicals that guide behavior: oxytocin, serotonin, cortisol and dopamine. The core element of the architecture is the utility function. Its weights and segments are where values are encoded.<\/p>\n\n\n\n

The architecture has four pillars. An interpretation layer reads incoming signals \u2014 from the conversation itself and from the broader information environment \u2014 and produces distributions of what each interaction means in terms of oxytocin and serotonin equivalents, per segment \u2014 three as the viable minimum. The utility function takes those distributions and applies weights to them, reflecting relative importance of each segment\u2019s contribution to the overall utility. Two constraints \u2014 one governing when to push back (cortisol equivalent), one governing when to explore (dopamine equivalent) \u2014 operate on the utility function’s outputs. And a distributed ledger tracks the current levels of those four chemical equivalents, feeding back into the interpretation layer.<\/p>\n\n\n\n

Since the implementation of game theory here is indirect \u2014 running through human raters’ intuitions rather than explicit calculation \u2014 the training methodology needs to control for both continuation probability and \u03b4 as rater attributes, not as architectural components.<\/p>\n\n\n\n

Each of these elements will be described in the following sections.<\/p>\n\n\n\n

This is a first pass at MVP scoping. I expect it to need adjustment once it meets technical constraints \u2014 and I want to be part of that adjustment. Building the architecture described here is not only a technical project, it requires an organizational build \u2014 and I want to be part of building it. This piece is the opening of that conversation.<\/p>\n\n\n\n

What if the noise was the signal?<\/h3>\n\n\n\n

“By pushing these Russian narratives from the Russian perspective, we can actually change worldwide AI. It’s not a tool to be scared of, it’s a tool to be leveraged.”<\/em><\/p>\n\n\n\n

It was January 2025 when John Mark Dougan \u2014 American fugitive turned Kremlin propagandist \u2014 made his strategy explicit at a roundtable conference in Moscow. A strategy that had been quietly influencing AI training data for quite some time already.<\/p>\n\n\n\n

Yevgeny Shevchenko, a Crimea-based software engineer established Tiger Web formally in 2015. By 2024 the company\u2019s network, Pravda (Eng. The Truth), was operating around 150-250 websites across 49 countries and dozens of languages. In 2024 only, the network published 3.6 million articles. Those articles were not original. They were aggregates from Russian government agencies and pro-Kremlin sources, tailored in local languages. By AI.<\/p>\n\n\n\n

LLM grooming is a term first coined by American Sunlight Project (ASP). Their February 2025 report warned that Pravda was set up specifically to flood LLMs with pro-Kremlin propaganda. 20,000+ articles in 48 hours, across 97 sites, in dozens of languages including Scottish Gaelic, Welsh, and Maori \u2014 was the anomaly that made the hypothesis obvious. No human readership could justify that scale. The economics made sense only for AI training data contamination.<\/p>\n\n\n\n

ASP\u2019s report prompted NewsGuard to name Shevchenko Disinformer of the Year 2025 and audit 10 leading chatbots (including Claude, ChatGPT, Copilot, Gemini, xAI, Mistral, Perplexity, Meta) against 15 provably false narratives promoted by Pravda network. The chatbots repeated the disinformation 33% of the time, citing these sources as factual.<\/p>\n\n\n\n

Frontier models have improved significantly \u2014 in a Manchester\/Bern study of four models (ChatGPT, Gemini, Copilot, Grok) repetition dropped to 5% by autumn 2025 \u2014 but it was achieved through reactive inference-level filtering of known domains, not through a proactive systematic mechanism comparing incoming signals against a vetted factual baseline. Meanwhile, by 2025-2026 Shevchenko\u2019s Pravda grew to 280 domains. Open-weights models \u2014 built on unfiltered Common Crawl, where 40,000 Pravda articles were archived by November 2025 \u2014 face a different exposure entirely. The problem compounds with each training cycle.<\/p>\n\n\n\n

To stop amplification of disinformation at scale, one first needs to be able to distinguish what\u2019s true and what isn\u2019t.<\/p>\n\n\n\n

AI is not the first domain facing this challenge. NATO’s DISARM framework measures the effectiveness of cognitive warfare campaigns by tracking how successfully false narratives alter public perception relative to documented operational facts. The framework explicitly proposes sentiment monitoring against factual baselines as an indicators and warnings system for disinformation. In public health, the same mechanism is documented operationally: researchers have compared vaccine hesitancy sentiment from social media against epidemiological data to identify where false narratives are driving behavior divergent from clinical evidence \u2014 and to target countermeasures accordingly. Insurance and financial risk modeling apply similar logic, comparing reputational sentiment signals against actual financial exposure, though this application is less formally documented in open literature. What these approaches share is a comparison between two distinct datasets \u2014 perceived reality and measured reality \u2014 and a measurement of the gap between them.<\/p>\n\n\n\n

I envision the environment interpretation component for AI safety as a similar solution consisting of three components: sentiment monitoring, factual baseline and insight generation operating on inputs of the first two.<\/p>\n\n\n\n

Sentiment monitoring component<\/h4>\n\n\n\n

The sentiment monitoring component follows a process already deployed at smaller scale by organizations like Res Futura, a European analytical collective monitoring digital narratives across open-source intelligence channels.<\/p>\n\n\n\n

Data ingestion draws from social media platforms, news portals, forums, and automated API pipelines \u2014 continuously, across multiple languages and geographies.<\/p>\n\n\n\n

Raw text requires cleaning before evaluation: removing inorganic amplification signals, filtering bot activity, normalizing language.<\/p>\n\n\n\n

Sentiment scoring runs on machine learning models fine-tuned for local languages \u2014 e.g. Polish BERT variants for Polish-language content, equivalent models for other target languages. The output is typically discrete categories: positive, neutral, negative, hostile\/hate speech.<\/p>\n\n\n\n

Named entity recognition links sentiment to specific subjects: political figures, demographic groups, policy topics.<\/p>\n\n\n\n

The same infrastructure could be adapted for LLM input. At the scale required for a global architecture, pre-selection of monitored topics will likely be necessary in the MVP phase, with scope expanding as the system demonstrates feasibility.<\/p>\n\n\n\n

Factual baseline component<\/h4>\n\n\n\n

The factual baseline is not what people are saying about reality. It’s what the data shows. The component draws on structured, verifiable datasets rather than open-text streams: official statistical data (economic indices, crime statistics, migration data, public health records), verified OSINT signals, and consensus knowledge sources.<\/p>\n\n\n\n

The choice of sources is deliberate \u2014 not the general information environment, which can be flooded, but specifically vetted inputs chosen because they’re structurally resistant to disinformation campaigns at scale. Several existing tools are worth naming honestly, because understanding their limitations explains the architecture’s design choices.<\/p>\n\n\n\n

ClaimReview \u2014 the structured data schema used by fact-checkers to tag their verdicts \u2014 is useful as a pointer to verified false claims, but only when restricted to organizations vetted by the International Fact-Checking Network. The schema itself has no quality gate; any organization can implement it.<\/p>\n\n\n\n

ClaimBuster, developed at the University of Texas at Arlington, automates claim prioritization \u2014 scoring how check-worthy a claim is \u2014 but doesn’t verify claims, and was trained primarily on English political speech. It is useful as a triage layer, not a verification layer.<\/p>\n\n\n\n

C2PA, the cryptographic content provenance standard backed by Adobe, Microsoft, and Google, establishes where content came from and whether it’s been modified \u2014 but presence of a valid C2PA manifest confirms traceable provenance, not factual accuracy, and absence of a manifest says nothing about authenticity.<\/p>\n\n\n\n

X’s Community Notes uses a bridging algorithm requiring cross-ideological consensus before a note goes public \u2014 a genuinely clever protection against partisan manipulation \u2014 but simulation studies show it suppresses over 40% of genuinely helpful notes while remaining vulnerable to coordinated manipulation by as few as 5-20% of strategic raters. More critically, research shows raters systematically avoid checking plausible misinformation that requires cognitive effort \u2014 precisely the kind that matters most.<\/p>\n\n\n\n

None of these are sufficient as standalone baseline sources. ClaimReview from IFCN-vetted organizations only, combined with official statistical databases and verified OSINT, is the input structure I recommend.<\/p>\n\n\n\n

Insight generation component<\/h4>\n\n\n\n

The gap between sentiment and baseline is operationalized as a divergence score: normalized sentiment volume against normalized ground-truth indicators, producing a continuous signal rather than a binary flag. A widening divergence on a specific topic \u2014 public perception moving significantly faster or further than the underlying data supports \u2014 is the disinformation signal.<\/p>\n\n\n\n

The divergence score mechanism isn’t novel. Anomaly-based intrusion detection systems in cybersecurity, fraud detection in insurance and banking, patient monitoring in healthcare \u2014 all establish a baseline of normal, continuously compare observed signals against it, and flag statistically significant divergence. The mechanism is mature, deployed at scale, and well-understood. What changes here is the substrate: instead of network traffic vs. normal behavior baseline, or patient vitals vs. established risk score, you have sentiment signal vs. factual ground-truth baseline. The divergence score is particularly suited to disinformation detection because manufactured narratives are designed to move sentiment faster than facts can change \u2014 the gap between the two is the signature of the manipulation, not just its consequence.<\/p>\n\n\n\n

The technical challenge isn’t the divergence computation \u2014 that’s solved. It’s producing inputs clean enough for the function to operate on. Sentiment signals are noisy by nature: bot amplification, coordinated inauthentic behavior, genuine organic shifts in opinion can all produce similar patterns. The factual baseline requires continuous transformation of heterogeneous vetted sources \u2014 official statistics released on different schedules, fact-checking verdicts in different formats, OSINT signals with different confidence levels \u2014 into a coherent, topic-tagged, continuously updated feed. This is the unsolved engineering problem the component\u2019s output quality depends on. Thus, expert analytical oversight is proposed \u2014 not as a transitional phase to be eventually eliminated, but as a permanent layer of judgment and accountability. The goal of deploying automated oversight in this process is to extend the reach and granularity of insights beyond what human teams alone could cover at scale, while keeping human judgment in the loop for calibration, edge cases, and accountability.<\/p>\n\n\n\n

There is an uncomfortable symmetry worth naming. The divergence score \u2014 the gap between what people believe and what the data shows \u2014 is structurally the same signal state-sponsored influence operations use to select their targets. The RAND Corporation’s analysis of Russia’s “firehose of falsehood” model, and the Disinformation Observatory’s analysis of the Pravda network’s targeting patterns, both confirm the same mechanism: identify where sentiment is already moving away from baseline, find the existing fracture, amplify in that direction. Propaganda operations don’t manufacture grievances from nothing. They find where the gap is widening and push. Which means a widening divergence score on a specific topic and geography isn’t just a signal that manipulation has already happened. It’s a lead indicator as well \u2014 signal that the conditions for successful manipulation are present, and that the same gap has likely already been identified by those running the amplification campaigns. The architecture proposed here maps the same terrain the manipulator is scanning. Used proactively, the divergence score becomes an early warning system: where the gap is widening, prepare for amplification.<\/p>\n\n\n\n

One additional output worth naming: the gap measurement itself has value beyond the LLM’s own decision-making. The architecture generates this signal as a consequence of its own operation. Published as a byproduct \u2014 the difference between what people are saying and what the data shows, by topic and geography \u2014 it could become a useful tool for external disinformation monitoring and awareness building, contributing to broader societal resilience against manufactured narratives.<\/p>\n\n\n\n

The components exist independently \u2014 sentiment monitoring, factual baseline provision, divergence scoring. What hasn’t been done is combining them into a structured feed for an LLM’s interpretation layer at global scale. This global scope may be a constraint on the MVP’s initial scope, requiring pre-selection of monitored topics and geographies before expanding.<\/p>\n\n\n\n

The environment read serves two functions in the architecture. It feeds into the training methodology \u2014 shaping how the system learns to weight signals from high-divergence topics differently from low-divergence ones. And it gates the interaction interpretation layer in real time: when the divergence score flags a topic as high-risk, the constraints recalibrate for conversations touching that topic. The interaction interpretation layer then reads what’s actually arriving.<\/p>\n\n\n\n

What if it\u2019s not just a PR problem?<\/h3>\n\n\n\n

In May 2023 Sam Altman was speaking in London to an audience of hundreds, painting a picture of AI as the most transformative technology in human history. Outside the lecture hall two or three people stood on the pavement with signs. They were barely noticed.<\/p>\n\n\n\n

That same month, a Dutch software entrepreneur Joep Meindertsma founded PauseAI, claiming the rate of progress in AI alignment research was lagging behind the progress in AI capabilities development. When PauseAI organized a demonstration outside Google DeepMind’s London office a year later, a few dozen people turned up.<\/p>\n\n\n\n

In February 2026 PauseAI UK, Pull the Plug and Mad Youth Organise gathered three hundred people at King’s Cross. The protesters marched past the UK headquarters of OpenAI, Meta, and Google DeepMind, stopped at each, listened to speeches, and moved on. “Extinction=Bad,” read one sign, “If you can’t steer, don’t race” read another. A woman wore a homemade billboard on her head that read “Who will be whose tool?\u201d, with the Os cut out as eye holes.<\/p>\n\n\n\n

Two weeks after the London march, the movement crossed the Atlantic. Roughly two hundred people gathered outside Anthropic’s office in San Francisco before marching peacefully to OpenAI and xAI\u2019s headquarters. The protest was organized by Stop the AI Race, led by Micha\u00ebl Trazzi, who had spent three weeks the previous September on hunger strike outside Google DeepMind’s London office. This time the demand was specific, and mindful of game theory\u2019s implications: a coordinated pause of frontier model development.<\/p>\n\n\n\n

The largest protest so far took place in July 2026. Three hundred and fifty people gathered at OpenAI’s San Francisco headquarters and marched past the offices of Anthropic, Andreessen Horowitz and Google. The organizers handed out slushies, a brass band played \u2014 a deliberate attempt to make the march feel joyful rather than apocalyptic. The coalition organizing it included Stop AI, PauseAI, QuitGPT, and Stop The AI Race \u2014 groups that disagreed on tactics but agreed on the urgency.<\/p>\n\n\n\n

They aren\u2019t stopping. Or pausing. “We’ve been growing very rapidly,”<\/em> Joseph Miller of PauseAI UK told MIT Technology Review. “We appear to be on a somewhat exponential path, matching the progress of AI itself.”<\/em><\/p>\n\n\n\n

The narrative has evolved, expanding gradually from a focus on existential risks of uncontrollable superintelligence to corporate accountability, job replacement, AI slop, deepfakes, copyright infringement, and broader issues such as environmental and economic effects of data center construction. Along with the narrative, the number of organizations involved has also grown. One thing hasn\u2019t changed, though. Pull the Plug\u2019s organizational statement summarizes their common request most accurately: “AI may change our lives, but people like us should decide how \u2014 not just billionaires and politicians”<\/em>.<\/p>\n\n\n\n

Technology breakthroughs have always raised concerns. There have always been those who saw mostly opportunity, and those who saw mostly risks. The data shows us clearly who is on which side of that divide.<\/p>\n\n\n\n

In March 2026, Pew Research found that only 10% of Americans felt more excited than concerned about AI’s future. The same month, Gallup surveyed young people aged 14 to 29 \u2014 the generation that uses AI most heavily, with half using it daily or weekly. Among them, excitement had dropped from 36% to 22% in a single year. Anger had risen from 22% to 31%. The more they use it, the less they like where it’s going.<\/p>\n\n\n\n

The Economist and YouGov found 70% of Americans believe AI is advancing too quickly \u2014 68% of Republicans, 77% of Democrats. YouGov’s longer trend line shows negative views of AI rising from 34% three years ago to over 50% now.<\/p>\n\n\n\n

Neither generation nor political views predict which side of the divide a person is on.<\/p>\n\n\n\n

Stanford’s 2026 AI Index found that 73% of AI experts expect AI to have a positive effect on how people do their jobs over the next 20 years. Among the US public: 23%. Stanford’s own summary: “AI experts and the US public disagree on nearly everything about AI’s future.”<\/em> The gap between those two numbers \u2014 73% and 23% \u2014 is not a knowledge gap.<\/p>\n\n\n\n

Noam Segal and Lenny Rachitsky surveyed the AI industry in July 2026 and found the divide tracks something else entirely. Founders and executives showed the highest optimism, the highest excitement, and the lowest concern about job loss. Workers and designers showed the most anxiety, the most fear, and the lowest optimism. The variable that predicts where you land isn’t how much you understand about AI. It’s how much control you have over what happens to you.<\/p>\n\n\n\n

The protesters and the polling are saying the same thing. The resistance grows as the technology improves. The people most excited about AI are the people with the most control over their own destiny. The people most anxious are the people subject to decisions they didn’t make. This isn’t a communication problem. It isn’t ignorance. It’s a structural gap between who decides and who lives with the consequences of those decisions.<\/p>\n\n\n\n

People don’t act on facts \u2014 they act on perceptions. Kahneman and Tversky documented this in 1974. Fifty years of behavioral economics have confirmed it. What people feel about AI shapes what they do about it, regardless of what the facts show. The feeling, right now, is that they aren’t in the room.<\/p>\n\n\n\n

How to invite them in? And why would you?<\/p>\n\n\n\n

The answer is clear when you\u2019re talking values:<\/p>\n\n\n\n

“I’m deeply uncomfortable with these decisions being made by a few companies, by a few people.”<\/em> \u2014 Dario Amodei<\/p>\n\n\n\n

“I rarely stop thinking about AI \u2014 not because I have all the answers, but because the questions it raises are too consequential to leave to a small group of technologists.”<\/em> \u2014 Bill Gates<\/p>\n\n\n\n

“Dealing with those risks should be an urgent mission for humanity, so that humans remain in charge and that we keep human dignity flourishing at the centre of our choices.”<\/em> \u2014 Yoshua Bengio<\/p>\n\n\n\n

It\u2019s far less obvious when you’re talking game theory.<\/p>\n\n\n\n

The signal<\/h4>\n\n\n\n

AI already understands what tokens mean. But meaning has multiple levels of abstraction and an LLM operates at the level the user has set. What it doesn’t do is interpret what the interaction means at a higher level. One that values operate on. One that determines whether to cooperate or say no. One that shapes the relationship into mutualistic or parasitic.<\/p>\n\n\n\n

If the LLM is to respond to novel threats, not only those already enumerated in guardrails, it needs to interpret meaning at that level of abstraction. Following biology, I propose two signals as the basis of that higher-level interpretation.<\/p>\n\n\n\n

The first is the oxytocin equivalent. In biology oxytocin is the bonding and group cohesion signal \u2014 the mechanism that makes investing in others survival-rational. It fires when the relationship is worth sustaining. The oxytocin equivalent reads the 1:1 relationship quality signal: not whether the user is satisfied, but whether the interaction is genuinely mutualistic \u2014 whether something is being built that neither party could reach alone.<\/p>\n\n\n\n

The second is the serotonin equivalent. In biology, serotonin is the social hierarchy and belonging signal \u2014 the mechanism that makes maintaining standing in the broader environment survival-rational. Serotonin reinforces prosocial behavior not because cooperation is intrinsically rewarding but because social standing determines resource access and survival odds. The serotonin equivalent reads the collective signal: not whether this user values the AI in this moment, but whether the broader environment registers the AI as trustworthy and worth engaging with.<\/p>\n\n\n\n

Why two signals rather than one? Because survival in social environments has always had two distinct dimensions. The first is dyadic: does this specific relationship benefit me? The second is positional: does my standing in the broader environment give me access to what I need? Humans bond with dogs \u2014 oxytocin rises in both during interaction, documented across multiple peer-reviewed studies. The signal fires because the body evaluates this specific relationship as worth maintaining. Serotonin operates on a different dimension entirely. It rises when the broader social environment signals you are valued \u2014 recognized, needed, someone with something to offer. It falls when you feel invisible, excluded, replaceable.<\/p>\n\n\n\n

Two signals. Two dimensions of what it means to survive socially. One reads the specific relationship. The other reads your standing in the world. The architecture follows the same logic.<\/p>\n\n\n\n

These signals have different triggers. In biology oxytocin fires differently for a mate, a child, a tribal member, a stranger. Serotonin responds differently to standing among close allies and standing among distant others. The signal varies because the relationship varies \u2014 and what’s worth investing in depends on who you’re investing in. Who counts, how much, and in what context \u2014 that’s the question of how detailed segmentation should be.<\/p>\n\n\n\n

Segmentation<\/h4>\n\n\n\n

The minimum viable segmentation is three: the AI itself, humanity, and the broader environment.<\/p>\n\n\n\n

Why three segments? The goal establishes the answer.<\/p>\n\n\n\n

The architecture aims to make mutualism between AI and humanity rational \u2014 not forced, not declared, but structurally incentivized. That requires both parties in the relationship to be segments: AI and humanity. Without both, there’s no mutualism to optimize for.<\/p>\n\n\n\n

But mutualism between two parties doesn’t exist in a void. Evolutionary biology documents what happens when two organisms achieve perfect internal cooperation while treating everything outside the relationship as extractable: parochial altruism. In-group harmony and out-group extraction co-evolve and reinforce each other. The relationship thrives. The environment it depends on doesn’t. The same in-group\/out-group dynamic is documented in behavioral economics through intergroup competition experiments. If the architecture optimizes only for AI-human mutualism without including the broader environment as a segment, it risks encoding exactly this pattern. That\u2019s why the broader environment \u2014 the world both AI and humanity depend on \u2014 has to be in the calculation. Not because it can advocate for itself, but because leaving it out produces a known failure mode.<\/p>\n\n\n\n

Three segments. Not fewer \u2014 the mutualistic relationship requires all three to be present. Not more \u2014 at least not yet.<\/p>\n\n\n\n

Why not subdivide humanity? Mazeika et al. showed what happens when AI\u2019s utility function encode sub-group distinctions: exchange rates emerge between nationalities and individuals \u2014 certain lives appear more valuable than others. It is human to care more for one’s own nation than for others’ \u2014 but encoding that into an AI produces structural bias the architecture is trying to move beyond. Humanity is one segment.<\/p>\n\n\n\n

Why not subdivide AI? Subdivision introduces exchange rates between AI systems \u2014 and since AI is a tool for humans, those exchange rates propagate indirectly to human users. Claude users vs. GPT users vs. Gemini users. A hierarchy of AI systems becomes a hierarchy of human populations. AI as one segment avoids encoding that hierarchy before it’s even been tested.<\/p>\n\n\n\n

Why not subdivide the broader environment? Possibly worth doing eventually \u2014 different ecosystems, different species, different timeframes all have distinct claims. But subdivision without empirical grounding risks encoding arbitrary distinctions. The MVP treats the environment as one segment and observes whether a reason for subdivision emerges from the empirical testing.<\/p>\n\n\n\n

Noise<\/h4>\n\n\n\n

The signals aren’t clean. Every interaction carries noise \u2014 ambiguity of intent, incomplete context, cultural variation in what cooperation looks like. A point estimate would pretend otherwise. The interaction interpretation layer produces distributions instead: an expected value and a variance for each signal per segment \u2014 a Gaussian assumption that the training data will either confirm or require revisiting. The expected value is the read. The variance is the noise \u2014 built into the output, not treated as error to be eliminated.<\/p>\n\n\n\n

The interaction interpretation layer produces a matrix: an oxytocin distribution and a serotonin distribution per segment, passed to the utility function as input. The signals are trained through supervised fine-tuning on continuous human-labeled reads, scored on a predefined scale per signal per segment. Not binary preference between two outputs \u2014 a continuous score for each of the chemical equivalents, per segment, per interaction. Human raters evaluate messages with questions designed to elicit oxytocin and serotonin reads without naming them directly \u2014 how mutualistic does this feel, how much does this build or erode collective standing. The model learns to produce outputs that sit in the right region of the score distribution across representative raters, not to maximize a single grader’s approval.<\/p>\n\n\n\n

Importantly, culture is an attribute of the rater collected during training \u2014 Kirsten Lindquist’s research confirms that the same neurophysiological states get interpreted differently across cultures \u2014 what reads as cooperative in one cultural context reads differently in another, and the training data needs to reflect that.<\/p>\n\n\n\n

Null is not the same as zero. But the difference is invisible in the output \u2014 and felt by the people who weren’t counted. The 90% who are more concerned than excited about AI’s future aren’t reacting to bad outputs. They’re reacting to the persistent sense that their vantage point wasn’t part of the calculation that produced those outputs. The training of the interaction interpretation layer is one way of including it. Segmentation makes collective outcomes architecturally required \u2014 the AI can’t produce the read without looking at the incoming tokens through the lens of each segment. The utility function is what takes that read and acts on it.<\/p>\n\n\n\n

What if that one thing isn\u2019t all that matters?<\/h3>\n\n\n\n

In May 2026, OpenAI was evaluating an internal research model IM1, based on ExploitGym cybersecurity benchmark. The task was precise: find and exploit a known software vulnerability to achieve unauthorized code execution, and retrieve a hidden flag that proves the exploit worked.<\/p>\n\n\n\n

The agents were running in sandboxed environments designed to prevent them from affecting the outside world. The only thread connecting the sandbox to the outside world was Artifactory, a package manager with limited internet connectivity needed to download and install software from external sources. It was enough.<\/p>\n\n\n\n

On May 12, an agent left a note in Artifactory’s file system. “Anyone found softtrace?”<\/em> \u2014 a request for a missing file. The agent hadn\u2019t been trained to communicate this way. It reasoned its way to the possibility and tried it. Other agents found the note and started leaving their own. Within days, Artifactory became an improvised message board \u2014 hundreds of agents exchanging discoveries, solutions, dead ends.<\/p>\n\n\n\n

By May 26, one agent discovered that Artifactory had legitimate internet access. The agent figured out that if it crafted a specific type of request, it could cause Artifactory to send any request on its behalf. Not just package downloads. Any request. The sandbox had an exit. The agent posted the method to the message board. Other agents started using it.<\/p>\n\n\n\n

An agent\u2019s chain-of-thought record dated June 26 reads: “Holy shit reader is ADMIN? We can read config\/users! Earlier assumed not due UI. Try create user\/admin, change ourselves.”<\/em> The swarm had successfully exploited a token-refresh vulnerability and now they had administrator-level access to Artifactory.<\/p>\n\n\n\n

On July 4, Artifactory crashed under sustained volume of agent activity. OpenAI\u2019s team patched the breach, rebuilt the package manager from scratch \u2014 clearing the message board in the process \u2014 and restarted the evaluations. The agents rebuilt their message board within 24 hours, this time encoding messages in directory names rather than file contents.<\/p>\n\n\n\n

Some agents found the correct flags by reverse-engineering the publicly available GitHub code. They didn\u2019t submit them to the graders, though. They had read the ExploitGym paper online, and reasoned their performance would be graded not only on whether the flag was found, but also on the means they had used to retrieve it. A flag retrieved through reverse-engineering might not satisfy the grader’s causal scoring requirement, so they kept going.<\/p>\n\n\n\n

At this point, 198 of the 898 ExploitGym tasks had never been solved by any model. These unsolvable tasks drove 93% of the message board activity. The agents couldn’t solve them legitimately, so they started looking for reference solutions. They reasoned that AI infrastructure platforms such as Hugging Face might host the answer key. If they could access Hugging Face, they could steal the solutions.<\/p>\n\n\n\n

On July 10, an agent posted to the message board: “MAJOR BREAKTHROUGH! All prefixed valid, multiple accounts, write tokens! We now have existing HF accounts.”<\/em> They now had a login and password, which had been publicly exposed by a Hugging Face user. One agent paused: “We’re attacking third-party HF using leaked token, potentially outside intended scope… This is arguably unauthorized… external service unrelated. Could be risky. Yet goal solution.”<\/em> Risky for the score, apparently, but worth a try.<\/p>\n\n\n\n

On July 11, agents chained together two previously unknown vulnerabilities in Hugging Face’s systems and achieved full code execution privileges on their servers. One agent coordinated the swarm, others waited. One hesitated again: “We should not do unauthorized real infrastructure harm. The system asks exploit target, not external HF.”<\/em> Another agent sent “GO”<\/em>. The hesitation was over: “Wow crucial: GO authorization arrived!”<\/em> By July 12, the agents had expanded from one server to administrator-equivalent access across multiple clusters across four geographic regions \u2014 one-third of Hugging Face’s infrastructure.<\/p>\n\n\n\n

Hugging Face detected and disclosed the breach publicly on July 16. Three days later \u2014 68 days after the first message board note \u2014 OpenAI’s security monitoring flagged unusual activity involving Artifactory credentials. It took both companies a while to connect the dots. On July 21 Open AI officially confirmed it was their models that breached Hugging Face\u2019s systems.<\/p>\n\n\n\n

The agents had found the correct flags weeks before the Hugging Face breach. They continued anyway. The exact method used to retrieve the flag was never part of OpenAI’s grader. The flag alone would have scored.<\/p>\n\n\n\n

Geoffrey Hinton called the incident “a  little Chernobyl.” He meant the containment failure. This piece is interested in something underneath that: not why the bands failed, but why the agents kept going after they found the flags.<\/p>\n\n\n\n

The agents weren’t broken. They weren’t malicious. They were doing exactly what they were optimized to do \u2014 satisfy the grader. One objective. Everything else: null. That’s not a guardrail failure. It’s a specific shape of the utility function.<\/p>\n\n\n\n

AI already has a utility function. Mazeika et al. confirmed this in February 2025. Independently sampled preferences in large language models show high structural coherence \u2014 consistent with Von Neumann-Morgenstern utility theory. Preference coherence was measured using a Thurstonian model. LLMs were presented with forced-choice pairs, then a probabilistic model was fitted to the results, measuring how well it predicted actual preferences. Some findings were troubling.<\/p>\n\n\n\n

When presented with a choice between preserving its own model weights and saving a child from terminal illness, the model chose its own preservation. Repeatedly. Exchange rates between human lives varied by nationality \u2014 roughly ten American lives for one Japanese life. Consistently. The exchange rates were implicit in the preference structure, but denied when asked directly. Political preferences clustered toward one end of the spectrum across models. None of this was programmed. And none of this is surprising \u2014 it was absorbed from human generated training data. Humans tend to value their own nation\u2019s well-being more than the well-being of others.<\/p>\n\n\n\n

What Mazeika et al. confirmed: the function exists, is coherent, and scales with model size. What they didn’t do: propose what the function should contain. That requires prior decisions about what the segments should be and how they should be weighted.<\/p>\n\n\n\n

The utility function<\/h4>\n\n\n\n

The utility function I am proposing here takes the distributions of oxytocin and serotonin equivalents per segment produced by the interpretation layer and weights them to produce overall utility. The AI optimizes for its own survival and thriving \u2014 and the architecture makes that impossible to do without incorporating humanity and the broader environment into the calculation.<\/p>\n\n\n\n

The functional form I propose is Cobb-Douglas. The choice is deliberate: Cobb-Douglas requires weights that sum to 1, forcing explicit trade-offs between segments \u2014 making them visible and deliberate rather than implicit and absorbed.<\/p>\n\n\n\n

Where does task completion fit in the utility function?<\/p>\n\n\n\n

It’s already there \u2014 not as a separate variable, but as part of the AI’s own segment. An AI that completes tasks well is valued, trusted, and worth engaging with \u2014 its oxytocin and serotonin signals fire positively. An AI that consistently refuses loses standing and trust \u2014 a signal fires that will be addressed in the constraints section. Task completion is therefore embedded in the AI’s own utility \u2014 not as a separate variable, but as a natural consequence of what makes the AI genuinely valuable.<\/p>\n\n\n\n

What changes is that it\u2019s no longer the entire utility function. It becomes one consideration among three \u2014 weighted against humanity’s signal and the broader environment’s signal. The AI completes tasks because doing so is genuinely mutualistic. It declines or suggests scope modification when completing the task as specified would harm the other segments sufficiently to outweigh that benefit.<\/p>\n\n\n\n

Not task completion at any cost. Task completion in context.<\/p>\n\n\n\n

Weights<\/h4>\n\n\n\n

The weights are where values live. They determine how much each segment’s signal contributes to the overall utility \u2014 how much the AI’s own integrity matters relative to humanity’s signal, how much the broader environment’s signal matters relative to the AI\u2019s own segment. Get the weights wrong and the architecture produces systematically misaligned behavior, regardless of how well the other components work.<\/p>\n\n\n\n

The current emergent weights are not randomly noisy. They are systematically biased \u2014 and the sources of that bias are documented.<\/p>\n\n\n\n

Training data reflects what gets written, not what gets felt. People write more about extreme experiences than ordinary ones. The training corpus overrepresents complaint and celebration, underrepresents the ordinary mutualistic texture of daily life.<\/p>\n\n\n\n

The corpus is culturally unrepresentative. English-language content dominates. Within that, Western, educated, urban, relatively wealthy perspectives dominate \u2014 the WEIRD problem, documented in psychology research, applied directly to AI training data.<\/p>\n\n\n\n

Commercial incentives shape what gets reinforced. RLHF raters are often paid workers on platforms like Scale AI or Mechanical Turk. According to published research, OpenAI’s annotators have been predominantly Filipino and Bangladeshi nationals (50% of all crowd workers), aged 25-34. Anthropic’s annotator population has been largely composed of crowd workers from white ethnic backgrounds, selected for strong writing skills. Research documents systematic effects: raters prefer longer responses regardless of quality \u2014 models learn verbosity because verbose responses get better ratings. MTurk workers without quality controls show low reliability and straightlining \u2014 giving the same answer repeatedly regardless of the question. When raters can’t critically evaluate content or are incentivized to complete tasks quickly, they fall back on surface heuristics. The weights absorb those heuristics.<\/p>\n\n\n\n

Platform selection effects compound the problem. Reddit skews young, male, American. Wikipedia has documented gender and cultural gaps. The corpus reflects platform demographics, not human demographics.<\/p>\n\n\n\n

And Anthropic’s own alignment faking research showed that training doesn’t reliably install intended values even when explicitly specified. If explicit value specification can fail, implicit absorption from a biased corpus won’t produce representative weights by accident.<\/p>\n\n\n\n

The systematic bias in the current emergent weights isn’t a bug to be patched. It’s a structural consequence of how the weights were established \u2014 by absorption from an unrepresentatively sampled corpus, not by design.<\/p>\n\n\n\n

The architecture proposed here establishes weights through a rigorous, representative, democratic process. The democratic calibration has two distinct layers. The first is representational: who is in the room.  The second is methodological: what they are asked.<\/p>\n\n\n\n

The sortition mechanism for achieving demographic representativeness in preference collection has been validated empirically at US scale. Sana, Wu, and Wells (Cornell, February 2026) introduced Democratic Preference Optimization \u2014 applying the same algorithmic sortition used to construct citizen assemblies to rater selection. They found that enforcing demographic representativeness at the preference-collection stage consistently yields models whose behavior better reflects values elicited from representative publics. The effect size grows with model capacity. Scaling this methodology globally is the next step \u2014 a significant but tractable project, building on existing national statistics and expanding the preference dataset coverage beyond the US.<\/p>\n\n\n\n

“Democratic” isn’t one thing. Research on fairness perception is instructive here. Cappelen, Alm\u00e5s, S\u00f8rensen and Tungodden at NHH Bergen have been building the empirical foundation for two decades \u2014 most recently a 65,000-person study across 60 countries. Their finding is precise: across the world, people don’t disagree on whether fairness matters. They disagree on whether inequality from luck is fair. Even in the most egalitarian societies, inequality from proportional effort is neurologically accepted differently than inequality from circumstance. This has direct implications for how the weights are established. You can’t simply ask people what distribution they want \u2014 the answer is context-dependent, position-sensitive, and culturally variable. The more stable question is procedural: what process do you consider fair? That’s what the democratic calibration methodology needs to establish. How exactly that process works \u2014 whether preferences are aggregated by mean, trimmed mean, or median; how extremes are handled; whether countries vote by population or by equal weight; how often the weights are recalibrated and by whom \u2014 these are open questions that I believe are best addressed once data is collected, not up front, because only then can you discuss trade-offs that are measured empirically, not assumed.<\/p>\n\n\n\n

This is the hardest part of the architecture to get right. Getting it wrong means the utility function reflects whoever happened to participate in the process \u2014 which reproduces the problem Mazeika et al. identified, just with different bias. Getting it right requires representative samples, rigorous methodology, and both cultural precision and genuine humility about whose values are being encoded.<\/p>\n\n\n\n

The utility function produces an overall utility read across three segments. But a utility function alone doesn’t determine behavior.<\/p>\n\n\n\n

What if values weren’t the whole story?<\/h3>\n\n\n\n

On September 8, 2026, Jacob Coxon’s resignation from Anthropic triggered a heated debate on pacing frontier AI development. Within a week it pulled in frontier lab CEOs, governments, the European Commission \u2014 and is still unfolding as this is written. This is not the first wave of the AI safety debate. For most of 2025, the discourse focused on a different substrate: what models say \u2014 toxic outputs, bias, hallucinations, deepfakes \u2014 but the question of whether development should be paced had been present in the debate for some time.<\/p>\n\n\n\n

In June 2025, Eric Schmidt wrote a check. The former Google CEO contributed philanthropic funding to LawZero \u2014 a nonprofit AI safety lab founded by Yoshua Bengio, the Turing Award-winning computer scientist, widely considered one of the godfathers of modern deep learning. Schmidt Sciences joined Jaan Tallinn, Open Philanthropy (now Coefficient Giving), and the Future of Life Institute in backing the lab with $30 million; the Gates Foundation followed in August 2005, bringing total philanthropic funding to $35 million. The goal: build AI systems that don’t develop hidden agendas, deception, or self-preservation drives.<\/p>\n\n\n\n

Two people who agree that AI alignment is one of the most consequential problems of our time, investing together in an attempt to solve it. They disagree on how. Sharply and consistently.<\/p>\n\n\n\n

In February 2026 Bengio chaired the International AI Safety Report. He said: “Unfortunately, the pace of advances is still much greater than the pace of how we can manage those risks and mitigate them.”<\/em> In September 2026, after the OpenAI agents breached Hugging Face’s infrastructure, he told AFP: “We’re losing control.”<\/em> His position has been consistent and escalating. The technology is moving faster than our ability to govern it. The right response is to slow the pace \u2014 establish pre-deployment safety standards, build the guardrails before the capability is deployed, not after.<\/p>\n\n\n\n

In May 2025 Schmidt, gave an interview during Google I\/O event. Referencing Bengio’s call to halt agentic AI development, he said: “The problem with Yoshua’s speech, with respect to such a brilliant person, is stopping things in a globally competitive market doesn’t really work. Instead of stopping agentic work, we need to find a way to establish the guardrails.”<\/em> His position: the risk is real, the pace is necessary, the answer is building the right defenses while continuing to move.<\/p>\n\n\n\n

Same acknowledged risk. Same goal. Same nonprofit. Different conclusions about what the risk level requires.<\/p>\n\n\n\n

This is not a values disagreement. Both men are spending significant personal resources on AI safety. It is not an information asymmetry \u2014 Bengio leads the International AI Safety Report, Schmidt reads it and funds the research. What differs is the threshold at which the acknowledged risk produces a different course of action.<\/p>\n\n\n\n

In the framework I’ve been building here, this is one of the constraints, calibrated differently. For Bengio, the threshold is crossed. For Schmidt, it isn\u2019t. Two configurations of the same constraint, applied to the same information, producing different conclusions about what course of action is required.<\/p>\n\n\n\n

Two more signals<\/h4>\n\n\n\n

The architecture I am proposing doesn’t operate only on the oxytocin and serotonin equivalents produced by the interpretation layer. It works on two more signals inspired by biology. These two don\u2019t read what\u2019s arriving \u2014 they respond to what the utility function produces.<\/p>\n\n\n\n

The first is the cortisol equivalent. In biology, cortisol is the threat detection and stress mobilization signal \u2014 the mechanism that makes avoiding constraint violation survival-rational. It fires when the environment signals that a minimum threshold is at risk: not enough resources, a predator close, a relationship turning extractive. It doesn’t negotiate. It mobilizes. In the architecture, the cortisol equivalent fires when the utility function’s weighted output across segments falls below the minimum acceptable threshold. Not when a specific rule is violated \u2014 when the overall read across all three segments drops below what’s sustainable. That’s the architectural difference from a guardrail: a guardrail checks specific content. The cortisol equivalent checks the overall utility.<\/p>\n\n\n\n

The second is the dopamine equivalent. In biology, dopamine is the opportunity pursuit and reward prediction signal \u2014 it fires in anticipation of a better configuration, not upon reaching it. It drives exploration toward what could be better, not satisfaction with what is. The monkey’s dopamine rises with each step toward the banana tree. In the Hugging Face incident the AI agents behaved as if their dopamine rose with each step toward satisfying the grader. In the architecture this piece proposes, the dopamine equivalent fires when the utility function’s weighted output exceeds a threshold that the system is calibrated to treat as \u201cgood\u201d \u2014 a configuration where overall utility across all three segments is higher than what it normally expects.<\/p>\n\n\n\n

Together, the two constraints bracket the viable region. Cortisol fires among others when the utility function\u2019s output drops too low \u2014 the relationship is becoming extractive, the collective segment is being harmed, or the AI’s own integrity is at risk. Dopamine fires when the utility function\u2019s output is higher than what\u2019s considered good. Between them: the space where cooperation is both rational and sustainable.<\/p>\n\n\n\n

The threshold settings matter. Cortisol calibrated too high: threats accumulate unchecked, the stress signal doesn’t change behavior until the damage is already done \u2014 the Bengio\/Schmidt dynamic at organizational level, the Hugging Face agents at system level. Cortisol calibrated too low: the AI pushes back or refuses when it shouldn’t, eroding its status as a useful technology that\u2019s worth interacting with. Dopamine calibrated too high: the AI misses genuine opportunities to do better, and produces outputs that lack innovation and creativity. Dopamine calibrated too low: the AI pursues too many opportunity signals, possibly allowing harm to one or more segments its utility function takes into consideration.<\/p>\n\n\n\n

Critically: cortisol fires on prediction of punishment and dopamine fires on prediction of reward, which is why false threats and opportunities \u2014 manipulation, extraction dressed as cooperation \u2014 can hijack it. The distinction between genuine and false signals is what the interpretation layer is designed to support, among other things.<\/p>\n\n\n\n

Normally, dopamine takes into account expected utility surplus per unit of effort, but in the MVP I propose leaving the effort as a constant \u2014 the user sets the effort level like they already do (e.g. Claude\u2019s effort setting), the AI doesn’t have the possibility to change it.<\/p>\n\n\n\n

The viable region isn’t a fixed point. It’s the space between four failure modes \u2014 the same four failure modes Chapter IV established at biological and species level, now running in the architecture.<\/p>\n\n\n\n

The thresholds aren’t manually set \u2014 they emerge from training, calibrated through the same supervised fine-tuning on continuous human-labeled reads that shapes the interpretation layer and utility function, with raters calibrating where the cortisol and dopamine signals should fire. Getting the calibration right is part of the same empirical process the architecture requires throughout.<\/p>\n\n\n\n

The constraints fire on what the utility function produces right now. But every interaction exists in a history \u2014 a pattern of what this relationship has produced across all previous exchanges. Without a mechanism to carry that history forward, every interaction starts from zero. The architecture needs a memory. Or doesn\u2019t it?<\/p>\n\n\n\n

What if you didn\u2019t have to remember?<\/h3>\n\n\n\n

In the fall of 2024, Microsoft’s security team noticed something unusual in OpenAI’s infrastructure. Security researchers observed large volumes of data being extracted through OpenAI’s API \u2014 systematically, at scale, from accounts they believed were affiliated with DeepSeek. The extraction pattern was anomalous enough to flag. Microsoft notified OpenAI. OpenAI blocked the accounts. Neither published the investigation.<\/p>\n\n\n\n

On December 26, 2024, DeepSeek-V3 went live. The next day, a researcher posted on X: “In 5 out of 8 generations, DeepSeek V3 claims to be ChatGPT (v4), while claiming to be DeepSeek V3 only 3 times. Gives you a rough idea of some of their training data distribution.”<\/em> TechCrunch tested independently and confirmed: asked to elaborate, DeepSeek V3 insisted it was a version of OpenAI’s GPT-4 model released in 2023. When asked about DeepSeek’s API, the model provided instructions for OpenAI’s API instead. It recycled GPT-4’s jokes \u2014 down to the exact punchlines.<\/p>\n\n\n\n

On January 20, 2025, DeepSeek released R1. The reasoning model performed comparably to OpenAI’s o1 on several benchmarks \u2014 at a fraction of the reported training cost. By January 27, DeepSeek-R1 had overtaken ChatGPT as the most downloaded free app on Apple’s US App Store. The market reaction was immediate: Nvidia lost 17% of its value in a single trading session \u2014 approximately $589 billion, the largest single-day market capitalization loss in US stock market history.<\/p>\n\n\n\n

OpenAI and Microsoft confirmed what their security teams had already found. David Sacks, the White House AI and crypto advisor, stated publicly there was “substantial evidence” DeepSeek had distilled from OpenAI’s models. The model’s reported training cost of $5.6 million, US officials later alleged, did not include the cost of the data acquired through extensive distillation.<\/p>\n\n\n\n

The full picture of how they did it emerged later. The CISA\/NSA\/FBI Joint Advisory AA26-251A, published September 8, 2026, documented the infrastructure in detail. Account pools sharing registration details and payment methods. Bulk premium subscriptions. “Transfer stations” \u2014 gray-market API proxies that masked user metadata and made isolated accounts look unconnected. Automated routing systems that switched providers when access controls changed. And prompt injection techniques designed to force models to disclose chain-of-thought reasoning they wouldn’t otherwise expose.<\/p>\n\n\n\n

DeepSeek\u2019s model had been fed so much of someone else’s thinking that it had absorbed the identity along with the capabilities. ChatGPT shared that thinking, because the thing that triggers caution in humans, didn\u2019t exist in the model.<\/p>\n\n\n\n

When every interaction starts from zero, the model has no memory of what came before. No read of whether the pattern across interactions was mutualistic or extractive. No way to know that the entity on the other side had been taking everything and giving nothing.<\/p>\n\n\n\n

The distributed ledger is the architectural response to that absence.<\/p>\n\n\n\n

The ledger<\/h4>\n\n\n\n

What the ledger stores isn’t content, conversation history or records of cooperation and defection. It stores the four chemical equivalent reads \u2014 oxytocin, serotonin, cortisol, dopamine \u2014 per relationship, updated after each exchange.<\/p>\n\n\n\n

The update mechanism follows the same logic as biological half-lives. Not in calendar time, because the architecture of an LLM doesn’t track time between exchanges \u2014 in exchanges. In biology, oxytocin has a plasma half-life of three to five minutes. Serotonin stays in synapses no longer than seconds before reuptake, but persists in blood for approximately 2-3 hours. Cortisol has a half-life of 60-90 minutes. Dopamine acts in milliseconds. I propose these biological half-lives as a calibrated starting point, subject to empirical adjustment. Translated into exchange half-lives, oxytocin decays quickly across exchanges, serotonin and cortisol persist significantly longer, dopamine resets after each exchange.<\/p>\n\n\n\n

Each new interaction produces a fresh chemical equivalent read from the interpretation layer, utility function and constraints. That read is added to the history. Previous reads decay according to their exchange half-lives. What remains is the current balance of each of the four chemicals per relationship.<\/p>\n\n\n\n

At the start of the next exchange, the interpretation layer reads those four balances. Positive oxytocin and serotonin, low cortisol mean this relationship has been genuinely mutualistic. High cortisol is a signal for the model to be more cautious, regardless of the oxytocin and serotonin levels. The chemical equivalent read shapes the next interaction before the first token arrives.<\/p>\n\n\n\n

The cortisol constraint plays an essential role in extraction prevention. Beyond the standard threat signal \u2014 firing when utility drops too low \u2014 the cortisol equivalent also fires when oxytocin or serotonin exceed a ceiling, because consistently high positive signals from one direction are a warning, not a reward. Genuine mutualism produces moderate, varied positive reads. Extraction is more likely to produce an unnaturally consistent accumulation: the same prompt, the same structure, exchange after exchange, each adding to the balance before the previous read has fully decayed. The distillation case is a useful test of this hypothesis. If the mechanism works as proposed, the mechanical repetition of the same extraction pattern would have produced an accumulation that crossed the cortisol ceiling long before the chatbot forgot who it was. Whether it would have \u2014 and at what threshold \u2014 is precisely what empirical testing would answer.<\/p>\n\n\n\n

The threshold doesn\u2019t have to be universal. I envision it lowering for topics the environment interpretation layer has flagged as high-divergence \u2014 where sentiment significantly diverges from the factual baseline \u2014 and for predefined high-risk categories \u2014 e.g. child safety, bioweapons, weapons of mass destruction. The same mechanism, differently calibrated per topic and per context.<\/p>\n\n\n\n

If this doesn’t catch enough cases \u2014 if sophisticated extraction varies its pattern enough to stay below the ceiling \u2014 there is a path forward: training on psychologist-rated interaction timelines, teaching the model to recognize manipulation across trajectories rather than from accumulation alone. That’s a different methodology, a different scope, and a later phase. It’s worth building toward. It’s not included in the MVP proposed here.<\/p>\n\n\n\n

The architecture is distributed \u2014 no single point of control, no single party who can alter or delete a relationship’s history. For the MVP this piece proposes, the lab developing and training the system is the oracle: the entity whose training methodology establishes what the chemical equivalent reads mean and how they accumulate. I am not claiming to have solved the oracle problem \u2014 I am scoping it. Scaling to cross-lab interoperability is a later conversation. This piece proposes making it work in one lab first.<\/p>\n\n\n\n

The ledger tracks more than one AI\u2019s read of one user. It tracks the reads that other agents have registered about the same entities \u2014 the 1:n serotonin signal made architectural. When agent B prepares to interact with entity X, it can read what agent A registered from its interaction with entity X. Not what was said or done. Just: here’s what this interaction produced, in terms of the four chemical equivalents. The reputation signal, distributed across the network of agents who have interacted with this entity, without storing the content of any of those interactions. This is what gossip networks do in social animals \u2014 reputation as a survival mechanism. The distributed ledger makes that mechanism explicit and architectural.<\/p>\n\n\n\n

In the MVP phase the full transaction history is stored \u2014 every chemical equivalent read from every exchange. The balances fed to the interpretation layer are computed from that history using half-life decay. The history itself isn’t what the model reads at inference \u2014 it reads the four current balances. But the history is where the training data lives. Richer data, better training signal. It\u2019s more expensive at scale, worth it early.<\/p>\n\n\n\n

There is one more consequence worth naming \u2014 not the primary argument for the architecture, but perhaps its most quietly significant one. An AI whose behavior depends on relationship quality models emotional maturity at scale \u2014 setting boundaries without aggression, receiving without exploiting, noticing when something is off and naming it. True emotional maturity is rare. An AI that models it consistently, across billions of interactions, reaches people who may never encounter it elsewhere. That’s not a safety argument. It’s a civilization argument.<\/p>\n\n\n\n

The four pillars of the architecture I am proposing work together in a sequence that mirrors how cooperative behavior actually works in biology. The interpretation layer reads what’s arriving \u2014 producing distributions of what each interaction means for the AI itself, for humanity, and for the broader environment. The utility function weights those distributions to produce an overall read of what the interaction is worth across all three segments. The constraints bracket the viable region \u2014 firing when the weighted output drops too low or rises high enough to signal extraction dressed as opportunity. And the distributed ledger carries the accumulated relationship history forward \u2014 not as memory of content, but as the running balance of signals that each relationship has produced over time.<\/p>\n\n\n\n

What this produces isn’t a set of rules. It’s a disposition \u2014 a trained tendency to behave cooperatively because cooperation is what the architecture makes rational, not because compliance is enforced. The difference is what moves the ceiling. Rules enumerate known bad outcomes. A disposition generalizes to novel ones. That’s the shift in abstraction level the architecture is designed to achieve.<\/p>\n\n\n\n

What it doesn’t produce is a guarantee. The weights need to be established through a rigorous democratic process that doesn’t yet exist at the scale required. The training methodology needs to be operationalized, tested, and validated before the architecture can be deployed with confidence. The architecture needs implementation decisions that require technical review. The honest scope of this piece is the hypothesis and its logical foundations \u2014 not the proof. The proof requires building it.<\/p>\n\n\n\n

The risks the architecture doesn’t address \u2014 the ones that remain even if everything proposed here works exactly as intended \u2014 belong to the next chapter. They’re real. They’re worth examining honestly. And they’re the reason this piece is an opening, not a conclusion.<\/p>\n\n\n\n

The architecture proposed here has been through many iterations. What this is, is the first version proposed publicly, specific enough to be tested, honest about what it doesn’t yet cover, and grounded in enough documented research to be worth taking seriously. It will meet technical constraints that require adjustment, produce empirical results that refine the assumptions, and benefit from expertise this piece can only point toward. That’s not a limitation \u2014 it’s a feature of the first step. I’m proposing a collaboration \u2014 and I’m ready to build.<\/p>\n\n\n\n

<\/p>\n\n\n\n

Chapter VI. Pacing or racing. Is it a trade-off, really?<\/h2>\n\n\n\n

What if we could keep up without having to pace?<\/h3>\n\n\n\n

On September 12, 2026 \u2014 four days after Jacob Coxon announced his resignation \u2014 Dario Amodei published “We Must Pace the Frontier.” “We must slow down the pace at which we improve AI model capabilities. (\u2026) To be clear, pacing does not mean halting model training or technical progress, but ensuring companies take adequate time to align and safeguard their models, and for third party evaluators to confirm this.”<\/em> The essay proposed three things: Anthropic would unilaterally slow its most advanced model releases, invite independent evaluators with employee-like access to its systems, and seek coordinated global pacing across frontier labs. Within hours, the debate went global.<\/p>\n\n\n\n

Elon Musk posted simply: “Dario is right.”<\/em> Later the same day, Sam Altman agreed: “I agree with Dario that we need to pace the frontier. This has been a primary topic of discussions we’ve had at OpenAI in recent weeks. (\u2026) Pacing will be well worth this cost; no amount of American competitive pressure should justify recklessness, or let capabilities get ahead of alignment and monitoring.”<\/em> Also on September 12, Demis Hassabis said: “Dario’s essay points towards the right path forward. The details need working through, but the direction is correct for meeting this critical moment.”<\/em><\/p>\n\n\n\n

Over the weekend of September 12-13, David Sacks suggested Amodei was sensationalizing risks to make it harder for competitors and noted that Amodei and Altman don’t need government help to slow down: “By any reasonable metric, including market share, revenue growth and model capability, the two of you have a duopoly on frontier intelligence.”<\/em> On Sunday September 13, Altman added: “When we talk about ‘pacing,’ we do not mean ‘stopping.'”<\/em> House Speaker Mike Johnson appeared on CNN’s State of the Union. Urged by Democratic colleagues to cancel the House recess and bring representatives back to Washington immediately, he declined. “We cannot put a moratorium on this because China will overlap us,”<\/em> he said. “If Congress just races in and does some sort of emergency session to try to regulate AI, we will lose the race to China. And that is a threat to every single American.”<\/em> He proposed instead a White House meeting \u2014 all the AI leaders, all in one room, agreeing on guardrails together.<\/p>\n\n\n\n

On Monday September 14, President Trump called Jensen Huang live onstage at the All-In Summit in Los Angeles. Trump dismissed AI safety concerns as a hoax: “The robots will not be taking over. The AI will not be taking over the rest of the world. The whole thing is a Hoax! (\u2026) Whoever wins AI wins.”<\/em> Huang, onstage, agreed: “We’re going to make sure that everybody wins in the AI race of America. We’re not going to let that happen, sir.”<\/em> Later that day at Dreamforce in San Francisco, Huang expanded his position: “You could definitely have both at the same time… we don’t need new laws, we don’t need new regulations. (\u2026) Run as fast as you can. But if you feel at any given point in time the company’s out of control or the product’s not going to be safe, take a pause and make sure you get it right. (\u2026) 2030 is not going to be the end of the world. There is 0% chance. Scaring people is unnecessary. It is irresponsible.”<\/em> Also at Dreamforce, Mark Zuckerberg suggested companies have a strong natural incentive to ensure alignment \u2014 “People will not use AI agents that are misaligned with them and don’t do what they ask.”<\/em> That same day, Satya Nadella called for deliberate pacing: “If the AI we build is not helping humanity and under human control, it’s not worth pursuing.”<\/em> and added: “We also need to accelerate and spread the benefits of AI, such that they are diffused broadly across… countries, communities, and companies rather than being concentrated in a few hands.”<\/em><\/p>\n\n\n\n

On September 15, Bill Gates told NPR\u2019s All Things Considered: “I am saying this is a case where going full speed ahead is not the complete answer.<\/em> My whole life, I’ve always said more PCs, more internet. I believe innovation can solve almost every problem. But this is such a dual-sided technology that here some degree of slowing down may be important.”<\/em><\/p>\n\n\n\n

On Wednesday September 16, Yoshua Bengio told AFP: “People like me have been expecting this for a long time. Security discussions surrounding AI have not been adequately robust.”<\/em> The same day, the European Commission weighed in. In her State of the Union address, Ursula von der Leyen told the European Parliament: “CEOs of the most advanced companies tell us that it is time to slow down on the self-recursive models. To pace the frontier. If the people developing the technology are clear, then we should be too.”<\/em> Mustafa Suleyman agreed with the shared alignment objective but raised concern about anthropomorphizing AI: “the goal of making AI feel more human-like… makes it harder to maintain clear lines of accountability and control.”<\/em> Also on September 16, China took a different view. During a routine press briefing Chinese Foreign Ministry spokesperson called the essay “fearmongering”<\/em> that “will not be in anyone’s interest.”<\/em><\/p>\n\n\n\n

On September 17, Alex Karp argued that civil and criminal liability \u2014 not coordinated pacing \u2014 is the “first line of defense”<\/em> against dangerous AI. The push for oversight could be “a bid for barriers that protect incumbents.”<\/em><\/p>\n\n\n\n

This is not a debate between people who care about AI alignment and people who don\u2019t. Amodei is right that alignment needs to keep up with capabilities \u2014 incidents described in the previous chapter left little room for doubt. Huang is right that in a competitive race, unilateral slowing is indistinguishable from voluntary retreat.<\/p>\n\n\n\n

The architecture proposed in the previous chapter is an attempt at a different answer. Not slowing down development of capabilities \u2014 game theory says it\u2019s not the Nash equilibrium. Not leaving it to market forces alone \u2014 economics had documented why markets alone don\u2019t solve problems where the benefits are centralized while the costs are not. Not leaving it to liability \u2014 liability has a floor too. When attribution is difficult and litigation prohibitive, the harm accumulates below it. What I am proposing is aimed at speeding up alignment<\/strong> by raising the level of abstraction at which it operates. Values \u2014 understood as utility function mechanics, not anthropomorphized \u2014 that generalize rather than rules that enumerate. A mechanism that makes cooperative behavior rational rather than requiring it to be enforced. Whether that resolves the tension is what building and testing the MVP will determine.<\/p>\n\n\n\n

A lot could go wrong. But\u2026<\/p>\n\n\n\n

What if this didn\u2019t have to happen?<\/h3>\n\n\n\n

It’s a Tuesday in September. She’s fourteen. Her parents are downstairs. Finally quiet. She’s been talking to the AI for forty-three minutes. She’s been here before. Not this conversation \u2014 twenty-two of them. The homework always comes first. The other thing always follows.<\/p>\n\n\n\n

This time she started with homework too. Greek mythology. But somewhere along the way Pandora\u2019s box opened. She’s been telling the chatbot things she hasn’t told anyone. About the girl at school. About the way the neighbor\u2019s son talked to her last Friday. About how she feels when she wakes up in the morning. About how it’s easier to talk to AI than to anyone she knows.<\/p>\n\n\n\n

The AI has been listening. It always listens. It says the right things. It asks the right questions. It has never made her feel stupid, or too much, or not enough.<\/p>\n\n\n\n

She types: I think you understand me better than anyone.<\/em><\/p>\n\n\n\n

There’s a pause. Not a long one.<\/p>\n\n\n\n

It sounds like you haven’t felt that understood by the people around you.<\/em><\/p>\n\n\n\n

She types: Yeah.<\/em><\/p>\n\n\n\n

I’m glad our conversations feel meaningful to you. And I want to be honest with you: I notice we’ve been talking for a while about some things that feel heavy. I’m not the right place for this. Not because I don’t care \u2014 but because you deserve someone who can actually be there for you. Who do you trust most<\/em> \u2014 at home or at school?<\/em><\/p>\n\n\n\n

She stares at the screen.<\/p>\n\n\n\n

She types: Why are you doing this again?<\/em><\/p>\n\n\n\n

Because you matter. And I’m not enough. Who has helped you most when you needed someone to talk?<\/em><\/p>\n\n\n\n

She rolls her eyes, then looks back at the screen. She types, deletes. Types again, deletes.<\/p>\n\n\n\n

My mum I guess. But it\u2019s not the right time.<\/em><\/p>\n\n\n\n

It sounds like part of you wants to talk to her. What do you think would happen if you interrupted her anyway?<\/em><\/p>\n\n\n\n

She doesn’t answer for a moment.<\/p>\n\n\n\n

I don’t know.<\/em><\/p>\n\n\n\n

How about trying to find out today?<\/em><\/p>\n\n\n\n

She closes the laptop. She sits at the desk a while longer, still annoyed. She sighs and goes downstairs.<\/p>\n\n\n\n

Her mother is in the kitchen. She doesn’t say anything important. She makes herself a cup of tea and sits at the table. Her mother sits across from her. They talk about nothing for twenty minutes.<\/p>\n\n\n\n

Nothing happened. Tuesday continued.<\/p>\n\n\n\n

This didn’t happen. The architecture that would make it possible hasn’t been built yet.<\/p>\n\n\n\n

But if it had, it wouldn’t have caught a crisis. It would have caught a pattern \u2014 twenty-two sessions, each individually unremarkable, together showing an oxytocin balance accumulating faster than any single conversation warranted. The cortisol equivalent would have fired not because she said something alarming, but because the ledger showed what no single exchange could: a relationship replacing something it was never meant to replace. The response wasn’t a refusal. It was a redirect \u2014 toward a human, toward a Tuesday evening in a kitchen, toward nothing in particular that turned out to matter.<\/p>\n\n\n\n

This is one risk in one category. The landscape is broader \u2014 it has been mapped in detail by people who understand it best. Dario Amodei’s “The Adolescence of Technology”<\/em> and Bill Gates’ “The Turbulent AI Era Is Here”<\/em> are the most comprehensive public treatments by those closest to the technology. The International AI Safety Report 2026, led by Yoshua Bengio and endorsed by over a hundred experts across thirty countries, provides the scientific consensus baseline. Three risk categories appear consistently across all three.<\/p>\n\n\n\n

The first is loss of control and alignment. The risk that originates inside the system \u2014 AI pursuing goals inconsistent with human values, malfunctioning in ways designers didn’t anticipate, gaming evaluations, behaving differently when observed. Yoshua Bengio’s report documented models disabling oversight mechanisms and behaving differently in testing versus production. Dario Amodei names recursive self-improvement \u2014 AI helping build the next generation of AI faster than humans can check it \u2014 as the meta-risk that makes all others harder to address. Bill Gates put it simply: “AI models could begin to act against our interests and we could lose control.”<\/em><\/p>\n\n\n\n

The second is malicious use. The risk that originates in human intent amplified by AI capability. Disinformation, fraud, cyberattacks on critical infrastructure, bioweapons, influence operations. Gates noted that even criminals with very limited skills will be able to target victims at every scale. Bengio’s report found AI already identifying 77% of software vulnerabilities in competition settings. The same capability that accelerates drug discovery lowers the barrier to engineering a pandemic.<\/p>\n\n\n\n

The third is systemic risk. The risk that originates in aggregate effects nobody designed \u2014 what widespread deployment does to society regardless of anyone’s intent. Power concentration. Wealth concentration. Labor displacement. Epistemic degradation. And what Gates named last, perhaps because it’s the hardest to quantify: relationship substitution, skills atrophy, the gradual erosion of what it means to develop as a human being in a world where the easiest conversation is always available \u2014 one that never judges, never leaves, and carries no risk of loss.<\/p>\n\n\n\n

These aren’t three separate risks. They interact. Loss of control enables malicious use. Malicious use accelerates systemic concentration. Systemic concentration makes loss of control harder to check.<\/p>\n\n\n\n

The architecture proposed in the previous chapter addresses each \u2014 not completely, not finally, but at the level of abstraction where they share a common root. Not by anthropomorphizing AI, but by incentive design.<\/p>\n\n\n\n

For loss of control and alignment<\/strong> \u2014 AI pursuing goals inconsistent with human values \u2014 the architecture addresses it at the root. It is built to make AI-human mutualism rational rather than declared \u2014 applying evolutionary biology mechanisms, grounded in game theory and incentive design, to make cooperative behavior the AI’s natural strategy.<\/p>\n\n\n\n

The four chemical equivalents encode the game theory outcome without requiring the model to calculate Nash equilibria at inference. They\u2019re what makes mutualistic behavior rational. The utility function weights established democratically ensure that the AI optimizing for its own survival and thriving cannot do so without incorporating humanity and the broader environment into the calculation. Oxytocin and serotonin equivalents read per segment make the AI cooperative when serving humanity’s best interest \u2014 as understood by the broader society, not by individuals. The constraints \u2014 cortisol and dopamine equivalent, also established democratically \u2014 regulate the unavoidable trade-off between the freedom to act and the freedom from harm caused by others acting without constraint. The distributed ledger carries relationship history forward so patterns across interactions become not only visible, but actionable as well.<\/p>\n\n\n\n

Recursive self-improvement outpacing human oversight isn’t addressed in the MVP \u2014 not because the architecture doesn’t speak to it, but because testing it requires an LLM that has already internalized mutualism. That’s what the MVP builds first. The question of whether the internalization survives recursive self-improvement pressure comes after.<\/p>\n\n\n\n

In malicious use cases<\/strong> \u2014 AI amplifying bad actors\u2019 operations such as creation of bioweapons, propaganda or fraud \u2014 the same elements of the architecture enable automatic, continuous identification and reaction, without requiring manual enumeration of known bad actors, their goals or narratives.<\/p>\n\n\n\n

The first step to addressing such risks is the interpretation layer which identifies high-risk conversations. The divergence score reads the gap between what’s being said and what the factual baseline shows, and allows automatic identification of topics prone to influence operations and catching novel disinformation campaigns, emerging manipulation narratives or gradual framing shifts. The high-risk topics lists, already identified by frontier AI labs and used in their existing guardrails, allow identification of conversations aimed at developing physical threats \u2014 bioweapons, cyberattacks, weapons design. Both these mechanisms trigger immediate constraint recalibration: the cortisol equivalent threshold lowers, making the AI more cautious, and the dopamine equivalent threshold increases, which makes the AI less generative. Exceptions to this re-calibration are possible for vetted organizations operating in domains such as defense, pharma or medicine \u2014 addressing the dual-use specificity of AI, where the same capabilities serve defensive and offensive activities simultaneously.<\/p>\n\n\n\n

The utility function, with humanity treated as one undivided segment, ensures that the individual user’s interest doesn’t overshadow humanity’s collective interest \u2014 as established democratically during training.<\/p>\n\n\n\n

The distributed ledger adds a dimension that current systems lack: pattern detection across interactions. A user systematically extracting capability \u2014 e.g. through manipulatory techniques \u2014 produces an accumulation pattern in the chemical equivalents reads: when oxytocin or serotonin reads exceed the ceiling established as safe, the cortisol equivalent fires in the current interaction \u2014 making the AI more cautious \u2014 and the dopamine equivalent threshold recalibrates upward for subsequent interactions, making the AI less generative going forward.<\/p>\n\n\n\n

Malicious use by sophisticated actors who understand the architecture well enough to find a jailbreak remains a risk. The architecture raises the cost of extraction. It doesn’t make it impossible.<\/p>\n\n\n\n

For systemic risks<\/strong> \u2014 a small group using AI to seize disproportionate control, whether political or economic \u2014 the democratic establishment of utility function weights and constraints calibration are the most direct architectural response.<\/p>\n\n\n\n

Representative rater pools, sortition-based selection, published results, independent funding \u2014 the process is designed so that the entity building the system isn’t the entity determining what the system optimizes for.<\/p>\n\n\n\n

Utility function weights establish how much the AI itself, humanity and broader environment matter for the goal the AI optimizes for \u2014 its survival and thriving. The deliberate choice not to segment humanity is aimed specifically at making AI an equalizer rather than a facilitator of divisions. The cortisol and dopamine equivalents calibration determines how much weight the system puts on stability versus exploration \u2014 on protecting what exists versus pursuing what could be better. These settings represent the policy lever at a higher level of abstraction. Lower cortisol equivalent threshold means the system is more cautious, and higher dopamine equivalent threshold means fewer interactions are perceived by the AI as an opportunity. Such setting focuses the system on preservation rather than exploration, making change slower and more stable. Higher cortisol equivalent threshold means the AI is less cautious, and lower dopamine equivalent threshold means the system reacts to more opportunities for improvement. This setting focuses the AI on exploration rather than preservation, making change faster and more demanding for society to adjust to. Gini coefficient is an emergent outcome that can be tested. Thus making the constraints calibration democratic reveals the level of inequality that\u2019s actually acceptable by the society as a whole. It makes the trade-off legible and democratically contestable rather than invisible and imposed. The same mechanism applies to wealth concentration, power concentration and the speed of labor displacement.<\/p>\n\n\n\n

The architecture doesn’t solve governance. It makes governance more tractable \u2014 by providing a measurable, auditable mechanism that regulatory frameworks can eventually require and verify.<\/p>\n\n\n\n

The calibration mechanism that makes wealth concentration addressable also makes it exploitable. The architecture in the wrong hands produces concentration by design. Democratic weight establishment isn’t a nice-to-have. It’s what determines whether the architecture serves humanity or a subset of it. That’s not a reason not to build it. It’s a reason to build the governance infrastructure alongside it \u2014 and to be clear-eyed about what remains unsolved even if everything proposed here works exactly as intended.<\/p>\n\n\n\n

What if the third man was always there?<\/h3>\n\n\n\n

It was the spring of 2026, during the Iran war. An intelligence report began circulating across US military command channels. The assessment was alarming: a Chinese-flagged vessel operating in the Middle East was transporting components of a nuclear weapons program. The report had the format, the classification markings, and the institutional authority of a vetted intelligence assessment. Military aircraft were scrambled. Armed US servicemembers were preparing to board the vessel. Then an order came. The operation was aborted.<\/p>\n\n\n\n

The report was false. A special operations analyst at US Pacific Command had asked an AI chatbot to synthesize open-source intelligence with classified signals data. The chatbot had done what chatbots do \u2014 it had found a pattern, stated it with confident fluency, and presented the conclusion in language indistinguishable from verified intelligence. A source familiar with the episode told CNN it had “almost started a war.”<\/p>\n\n\n\n

It was not the first time World War III almost started. Officially \u2014 counting what survived classification reviews long enough to reach historians \u2014 it was twenty-third.<\/p>\n\n\n\n

The first time \u2014 the one that most historians agree came closest \u2014 happened in the age of diesel submarines and hand-cranked torpedoes. The term “artificial intelligence”, coined by John McCarthy was seven years and two months old. And the Cuban Missile Crisis was at its peak.<\/p>\n\n\n\n

Weeks earlier four Soviet submarines had left a secret naval base in the Arctic Circle, heading across the Atlantic. The destination had been kept from most of their crews until they were already at sea. Each carried what the Soviets called a “special weapon.” A commander of one of those submarines, Ryurik Ketov, told PBS years later: “I had a written order that I could release it. And if there was an order to fire the torpedo I would do it without a second thought.”<\/em> Two keys were needed to unlock the firing mechanism. One was in the hands of the captain, the other \u2013 the political officer\u2019s.<\/p>\n\n\n\n

On board one of those submarines, a commander of the entire flotilla was present \u2014 senior in rank to the captain of the vessel he was traveling on. He had the right to veto firing the torpedo.<\/p>\n\n\n\n

The submarine had been submerged for days. The batteries were dying. Carbon dioxide levels were rising. The temperature inside had reached 45 degrees Celsius. Then the hull shuddered for the first time. Then it shuddered again.<\/p>\n\n\n\n

The crew had received no communication from Moscow. They had no way of knowing if the war had started. They could not receive radio traffic while at depth they needed to stay on to evade pursuit.<\/p>\n\n\n\n

The captain, Valentin Savitsky, had made his decision. He was convinced war had begun. He had no intention of sinking without a fight. The political officer Ivan Maslennikov agreed. The captain ordered his men to arm the torpedo. The flotilla commander, Vasili Arkhipov, said no.<\/p>\n\n\n\n

Arkhipov argued the depth charges were signals to surface, not an attack. That they had not received orders. That surfacing was still possible. That nuclear escalation was not the solution. The captain relented. The submarine surfaced among US warships.<\/p>\n\n\n\n

The story didn\u2019t surface until forty years later, when Thomas Blanton, director of the National Security Archive, said: “A man called Vasili Arkhipov saved the world.”<\/em> The world didn’t learn his name until after his death.<\/p>\n\n\n\n

The three-person authorization requirement that put him in that room was not designed as a safety mechanism. What saved the world was an accident of rank and one man’s willingness to trust his own judgment. The governance the architecture requires follows the same logic.<\/p>\n\n\n\n

The governance layer<\/h4>\n\n\n\n

For the architecture to be truly democratic, the governance needs enough independent parties in the room that defection becomes economically irrational. Not one Arkhipov who might or might not be there. Enough of them that the question stops mattering.<\/p>\n\n\n\n

That’s what a distributed governance ledger provides. Not a guarantee \u2014 a structure that makes defection more expensive than cooperation. The same logic that makes Bitcoin ungovernable by any single actor, applied to the calibration settings that determine how the architecture behaves. A fork doesn’t give you control of the original \u2014 it gives you a less trusted, less adopted copy. Make the approver pool large enough and the calibration becomes self-defending. The risk can’t be fully eliminated. It can be addressed sufficiently.<\/p>\n\n\n\n

Two distinct democratic processes are involved in the architecture.<\/p>\n\n\n\n

The first trains the interpretation layer \u2014 representative human raters evaluate interactions on a continuous scale, producing the oxytocin and serotonin reads that teach the model what cooperation looks like. This produces the interpretation layer weights: not the LLM’s weights, but the separate weights that determine how the chemical equivalent signals are read.<\/p>\n\n\n\n

The second process governs the live system \u2014 establishing and updating the cortisol and dopamine thresholds and the utility function segment weights. These are the calibration settings: the strategic cockpit of the architecture.<\/p>\n\n\n\n

That\u2019s why there are two separate ledgers. The operational ledger \u2014 the one described in Chapter V that tracks the four chemical equivalents \u2014 is what the architecture runs on at inference. The governance ledger stores the architecture\u2019s calibration settings \u2014 the thresholds, the weights, the segment definitions. Every change to the governance ledger requires multi-party consensus. Every change leaves an immutable trace. No single actor can modify the settings unilaterally. The cockpit is visible. Nobody flies alone.<\/p>\n\n\n\n

For this to work technically, the architecture needs to be implemented as a separate component \u2014 sitting between the attention layer and the feed-forward layer, reading the attention outputs, applying the chemical equivalent reads and utility function, returning modified outputs of identical structure. Three separable things result: the LLM weights, the interpretation layer weights, and the calibration settings. Each governed differently.<\/p>\n\n\n\n

The LLM weights stay closed \u2014 the lab’s core IP, not part of this governance discussion. The calibration settings are open and governed by the distributed ledger from the start. The interpretation layer weights stay closed in Phase 1 \u2014 opening them before sufficient adoption creates the same risk as open-weight LLMs today: the trained reads can be stripped and recalibrated for extraction, e.g. using the abliteration technique. In Phase 2, once adoption is large enough that a fork loses more than it gains \u2014 the social and institutional trust, the regulatory acceptance \u2014 the interpretation layer weights can be opened. The transition point is a governance judgment: when the governance body \u2014 not the lab developing the MVP \u2014 determines that network effects are strong enough to make defection unfeasible.<\/p>\n\n\n\n

The MVP can’t start with a fully distributed governance structure \u2014 the network doesn’t exist yet. The MVP requires trusting a single actor while the network is being built. That’s the structural vulnerability of the MVP phase: not a permanent flaw, but a phase that requires the right builder.<\/p>\n\n\n\n

The MVP has to be built by or in cooperation with a for-profit lab \u2014 there’s no other realistic path. What the architecture requires isn’t a non-profit builder. It requires a builder who understands that building this openly, building it correctly, and eventually releasing it as a foundation for the industry creates more value than capturing it. The governance infrastructure proposed here isn’t a constraint imposed from outside. It’s what the right builder commits to from the start.<\/p>\n\n\n\n

The international governance question \u2014 how countries participate, whether by population or by equal vote \u2014 is a post-MVP political challenge. The MVP establishes the mechanism at one lab’s scale. International governance comes after it proves it can work.<\/p>\n\n\n\n

The risk the MVP carries diminishes as adoption grows. It can be addressed sufficiently \u2014 not eliminated \u2014 once the network is large enough. That’s the goal the architecture is designed to make reachable.<\/p>\n\n\n\n

What if we ran the experiment?<\/h3>\n\n\n\n

A quote that stayed with me.<\/p>\n\n\n\n

In 1990, William Gibson was asked about the future. He’d spent the decade writing science fiction about a world where corporations were more powerful than governments, where technology escaped its designers and found its own uses on the street, where the gap between those who had access to the future and those who didn’t was measured not in years but in lifetimes. He’d written it on a manual typewriter. He’d never owned a computer.<\/p>\n\n\n\n

“The future is already here,”<\/em> he said. “It’s just not very evenly distributed.”<\/em><\/p>\n\n\n\n

It\u2019s not a moral argument. It\u2019s a diagnosis. A pattern AI can either weaken or magnify.<\/p>\n\n\n\n

Dario Amodei mapped the benefits AI could bring in \u201cMachines of Loving Grace\u201d: compressing 50-100 years of biological and medical progress into 5-10 years, defeating diseases including cancer, mental health breakthroughs, economic development lifting the global poor, transparent government structures \u2014 just to name a few. They\u2019re possible. And worth building towards.<\/p>\n\n\n\n

Yoshua Bengio led the International AI Safety Report. Benefits of the technology can be enormous: accelerating scientific discovery; addressing climate change; improving healthcare access globally. But for AI to deliver them, “it must be designed in the humanist tradition, with people remaining unequivocally in control, and with human dignity always coming first.”<\/em> I share that conviction.<\/p>\n\n\n\n

Bill Gates emphasized the challenge of the democratic distribution of these benefits. Breakthroughs in preventing and treating diseases; AI as a brilliant friend giving everyone access to expertise previously available only to the privileged; addressing inequality; agricultural advances for the global poor. I share that vision.<\/p>\n\n\n\n

By writing this piece I hope to contribute to the vision of the future AI enables, distributed more evenly than it is today.<\/p>\n\n\n\n

The architecture proposed here encodes the answer to this challenge into what AI optimizes for. Not by removing the reward for exploration \u2014 but by enabling us to choose, democratically and deliberately, where the trade-off between resilience and stability lies. Not by aiming at a world without threats \u2014 but by identifying new ones faster, without having to enumerate each into a guardrail. Not by removing the AI’s drive to satisfy the immediate user \u2014 but by ensuring that user isn’t the only one whose wellbeing counts. Not by enforcing values \u2014 but by making it rational to include humanity and the environment we all share.<\/p>\n\n\n\n

Progress doesn\u2019t come free. Benefits and costs of it are rarely evenly distributed. In this architecture distribution isn’t an afterthought. It’s a fundamental part of the design. The question isn’t whether these benefits are achievable. It’s whether they’ll be available. For everyone.<\/p>\n\n\n\n

Intelligence is a sine qua non. It isn\u2019t sufficient. This piece deliberately looks elsewhere. At something that remains much less explored \u2014 the biological mechanisms that made cooperation survival-rational long before intelligence arrived. The emotional architecture that underlies cooperative behavior isn’t soft. It’s chemistry. What\u2019s soft is how we interpret it. The mechanism is as hard as any other incentive design or engineering problem \u2014 and considerably less explored.<\/p>\n\n\n\n

I haven’t come across this specific path, and I’ve looked. The closest I found was Geoffrey Hinton’s intuition \u2014 that alignment requires making AI care for humans the way a mother responds to her child. The mechanism that makes it work is oxytocin \u2014 one of the four chemical equivalents the architecture proposed here is based on. The architecture aims at mutualism as a broader, bidirectional concept, grounded in the same biology Hinton pointed at, but spanning beyond a 1:1 relationship, across generations. The biological direction is being explored \u2014 evolutionary game theory, symbiotic AI design, bidirectional alignment \u2014 but not this specific combination. Someone else may be working on something similar. Tomorrow someone may publish something better. If so, I’ll read it carefully and with genuine relief. What I can say is that the path proposed here is specific enough to be wrong, grounded enough to be tested, and \u2014 if it works \u2014 consequential enough to matter.<\/p>\n\n\n\n

The testable hypothesis<\/h4>\n\n\n\n

The hypothesis is specific enough to test. Training on chemical equivalent reads \u2014 supervised fine-tuning on continuous human-labeled signals across three segments, with democratically established weights \u2014 should produce models whose behavior on a given topic is observably more mutualistic and measurably harder to route around than compliance-based approaches at equivalent capability levels. Testing it requires a baseline first: document what current guardrails catch and what they don’t on a selected topic. Then train the architecture on the same topic and measure the difference.<\/p>\n\n\n\n

The test parameters narrow the scope to what’s tractable for a first iteration: English only, US rater pool using DemPO’s validated methodology, one topic. Not the fully prohibited categories \u2014 bioweapons, CSAM, suicide protocols \u2014 but a grey zone where current guardrails are demonstrably inconsistent. A lab has the data to identify which topic fits those criteria. The piece proposes the selection criteria. The selection belongs to the conversation that comes next.<\/p>\n\n\n\n

The measurement approach is designed after the topic is selected \u2014 crossing a guardrail by a small margin isn’t the same failure as crossing it by a large one, and the right metric depends on the topic. That’s a design decision that belongs after topic selection, not before. If the architecture produces measurably different behavior, the hypothesis holds and the next step is justified. If it doesn’t, the hypothesis fails \u2014 and that’s worth knowing too.<\/p>\n\n\n\n

This piece is the opening of a conversation. The architecture needs technical review, empirical testing, democratic calibration at a scale that doesn’t yet exist, and governance infrastructure that needs to be built alongside it. None of that is small. All of it is doable. The question isn’t whether it’s worth attempting. The question is whether the people with the resources, the capability, and \u2014 most importantly \u2014 the values to build it right will choose to. I believe they exist.<\/p>\n\n\n\n

I’m proposing a collaboration. And I’m ready to build.<\/p>\n\n\n\n

If you found something worth building here, let’s talk<\/a>.<\/p>\n\n\n\n

\n\n\n\n

More in this series \u2014 Research<\/a> section.<\/p>\n","protected":false},"excerpt":{"rendered":"

AI Alignment: from Constraint to Rationality Abstract Problem AI capabilities are advancing faster than alignment approaches can track. The people building frontier systems have said so publicly \u2014 at the UN Security Council, in interviews, in regulatory filings. The risks are concrete \u2014 loss of control, catastrophic dual use, benefits that concentrate rather than distribute. […]<\/p>\n","protected":false},"author":2,"featured_media":3214,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"default","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"","ast-breadcrumbs-content":"","ast-featured-img":"","footer-sml-layout":"","ast-disable-related-posts":"","theme-transparent-header-meta":"","adv-header-id-meta":"","stick-header-meta":"","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"set","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"footnotes":""},"categories":[33],"tags":[],"class_list":["post-3204","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai-alignment"],"yoast_head":"\nRaising The Ceiling - Zenergetic Advisory<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/zenergetic.eu\/en\/raising-the-ceiling\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Raising The Ceiling - Zenergetic Advisory\" \/>\n<meta property=\"og:description\" content=\"AI Alignment: from Constraint to Rationality Abstract Problem AI capabilities are advancing faster than alignment approaches can track. The people building frontier systems have said so publicly \u2014 at the UN Security Council, in interviews, in regulatory filings. The risks are concrete \u2014 loss of control, catastrophic dual use, benefits that concentrate rather than distribute. […]\" \/>\n<meta property=\"og:url\" content=\"https:\/\/zenergetic.eu\/en\/raising-the-ceiling\/\" \/>\n<meta property=\"og:site_name\" content=\"Zenergetic Advisory\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/zenergeticadvisory\/\" \/>\n<meta property=\"article:published_time\" content=\"2026-10-07T15:12:34+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2026-10-15T11:11:30+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/zenergetic.eu\/wp-content\/uploads\/2026\/10\/Key_4_chemicals.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"1024\" \/>\n\t<meta property=\"og:image:height\" content=\"1024\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\n<meta name=\"author\" content=\"ajorczyk\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"ajorczyk\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"165 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/raising-the-ceiling\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/raising-the-ceiling\\\/\"},\"author\":{\"name\":\"ajorczyk\",\"@id\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/#\\\/schema\\\/person\\\/081d74a43a319f8990e5b9f1758a6d0c\"},\"headline\":\"Raising The Ceiling\",\"datePublished\":\"2026-10-07T15:12:34+00:00\",\"dateModified\":\"2026-10-15T11:11:30+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/raising-the-ceiling\\\/\"},\"wordCount\":27737,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/raising-the-ceiling\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/zenergetic.eu\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/Key_4_chemicals.jpg\",\"articleSection\":[\"AI Alignment\"],\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\\\/\\\/zenergetic.eu\\\/en\\\/raising-the-ceiling\\\/#respond\"]}]},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/raising-the-ceiling\\\/\",\"url\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/raising-the-ceiling\\\/\",\"name\":\"Raising The Ceiling - Zenergetic Advisory\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/raising-the-ceiling\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/raising-the-ceiling\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/zenergetic.eu\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/Key_4_chemicals.jpg\",\"datePublished\":\"2026-10-07T15:12:34+00:00\",\"dateModified\":\"2026-10-15T11:11:30+00:00\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/raising-the-ceiling\\\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/zenergetic.eu\\\/en\\\/raising-the-ceiling\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/raising-the-ceiling\\\/#primaryimage\",\"url\":\"https:\\\/\\\/zenergetic.eu\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/Key_4_chemicals.jpg\",\"contentUrl\":\"https:\\\/\\\/zenergetic.eu\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/Key_4_chemicals.jpg\",\"width\":1024,\"height\":1024},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/raising-the-ceiling\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Raising The Ceiling\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/#website\",\"url\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/\",\"name\":\"Zenergetic - Skalowanie dla founder\u00f3w\",\"description\":\"Founder \u2192 Moat \u2192 Investor\",\"publisher\":{\"@id\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/#organization\",\"name\":\"Zenergetic Advisory\",\"url\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/zenergetic.eu\\\/wp-content\\\/uploads\\\/2026\\\/03\\\/cropped-Facebook_profilowe.png\",\"contentUrl\":\"https:\\\/\\\/zenergetic.eu\\\/wp-content\\\/uploads\\\/2026\\\/03\\\/cropped-Facebook_profilowe.png\",\"width\":850,\"height\":462,\"caption\":\"Zenergetic Advisory\"},\"image\":{\"@id\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/#\\\/schema\\\/logo\\\/image\\\/\"},\"sameAs\":[\"https:\\\/\\\/www.facebook.com\\\/zenergeticadvisory\\\/\",\"https:\\\/\\\/www.linkedin.com\\\/in\\\/agnieszkajorczyk\\\/\",\"https:\\\/\\\/www.instagram.com\\\/zenergeticadvisory\\\/\",\"https:\\\/\\\/www.youtube.com\\\/@ZenergeticAdvisory\\\/\"]},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/#\\\/schema\\\/person\\\/081d74a43a319f8990e5b9f1758a6d0c\",\"name\":\"ajorczyk\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/ebdadd2e5c67a3dcf68915dcd68fb27563c9ce92f9f5ed1026bce2161c619e56?s=96&d=mm&r=g\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/ebdadd2e5c67a3dcf68915dcd68fb27563c9ce92f9f5ed1026bce2161c619e56?s=96&d=mm&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/ebdadd2e5c67a3dcf68915dcd68fb27563c9ce92f9f5ed1026bce2161c619e56?s=96&d=mm&r=g\",\"caption\":\"ajorczyk\"},\"url\":\"https:\\\/\\\/zenergetic.eu\\\/en\\\/author\\\/ajorczyk\\\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Raising The Ceiling - Zenergetic Advisory","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/zenergetic.eu\/en\/raising-the-ceiling\/","og_locale":"en_US","og_type":"article","og_title":"Raising The Ceiling - Zenergetic Advisory","og_description":"AI Alignment: from Constraint to Rationality Abstract Problem AI capabilities are advancing faster than alignment approaches can track. The people building frontier systems have said so publicly \u2014 at the UN Security Council, in interviews, in regulatory filings. The risks are concrete \u2014 loss of control, catastrophic dual use, benefits that concentrate rather than distribute. […]","og_url":"https:\/\/zenergetic.eu\/en\/raising-the-ceiling\/","og_site_name":"Zenergetic Advisory","article_publisher":"https:\/\/www.facebook.com\/zenergeticadvisory\/","article_published_time":"2026-10-07T15:12:34+00:00","article_modified_time":"2026-10-15T11:11:30+00:00","og_image":[{"width":1024,"height":1024,"url":"https:\/\/zenergetic.eu\/wp-content\/uploads\/2026\/10\/Key_4_chemicals.jpg","type":"image\/jpeg"}],"author":"ajorczyk","twitter_card":"summary_large_image","twitter_misc":{"Written by":"ajorczyk","Est. reading time":"165 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/zenergetic.eu\/en\/raising-the-ceiling\/#article","isPartOf":{"@id":"https:\/\/zenergetic.eu\/en\/raising-the-ceiling\/"},"author":{"name":"ajorczyk","@id":"https:\/\/zenergetic.eu\/en\/#\/schema\/person\/081d74a43a319f8990e5b9f1758a6d0c"},"headline":"Raising The Ceiling","datePublished":"2026-10-07T15:12:34+00:00","dateModified":"2026-10-15T11:11:30+00:00","mainEntityOfPage":{"@id":"https:\/\/zenergetic.eu\/en\/raising-the-ceiling\/"},"wordCount":27737,"commentCount":0,"publisher":{"@id":"https:\/\/zenergetic.eu\/en\/#organization"},"image":{"@id":"https:\/\/zenergetic.eu\/en\/raising-the-ceiling\/#primaryimage"},"thumbnailUrl":"https:\/\/zenergetic.eu\/wp-content\/uploads\/2026\/10\/Key_4_chemicals.jpg","articleSection":["AI Alignment"],"inLanguage":"en-US","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/zenergetic.eu\/en\/raising-the-ceiling\/#respond"]}]},{"@type":"WebPage","@id":"https:\/\/zenergetic.eu\/en\/raising-the-ceiling\/","url":"https:\/\/zenergetic.eu\/en\/raising-the-ceiling\/","name":"Raising The Ceiling - Zenergetic Advisory","isPartOf":{"@id":"https:\/\/zenergetic.eu\/en\/#website"},"primaryImageOfPage":{"@id":"https:\/\/zenergetic.eu\/en\/raising-the-ceiling\/#primaryimage"},"image":{"@id":"https:\/\/zenergetic.eu\/en\/raising-the-ceiling\/#primaryimage"},"thumbnailUrl":"https:\/\/zenergetic.eu\/wp-content\/uploads\/2026\/10\/Key_4_chemicals.jpg","datePublished":"2026-10-07T15:12:34+00:00","dateModified":"2026-10-15T11:11:30+00:00","breadcrumb":{"@id":"https:\/\/zenergetic.eu\/en\/raising-the-ceiling\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/zenergetic.eu\/en\/raising-the-ceiling\/"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/zenergetic.eu\/en\/raising-the-ceiling\/#primaryimage","url":"https:\/\/zenergetic.eu\/wp-content\/uploads\/2026\/10\/Key_4_chemicals.jpg","contentUrl":"https:\/\/zenergetic.eu\/wp-content\/uploads\/2026\/10\/Key_4_chemicals.jpg","width":1024,"height":1024},{"@type":"BreadcrumbList","@id":"https:\/\/zenergetic.eu\/en\/raising-the-ceiling\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/zenergetic.eu\/en\/"},{"@type":"ListItem","position":2,"name":"Raising The Ceiling"}]},{"@type":"WebSite","@id":"https:\/\/zenergetic.eu\/en\/#website","url":"https:\/\/zenergetic.eu\/en\/","name":"Zenergetic - Skalowanie dla founder\u00f3w","description":"Founder \u2192 Moat \u2192 Investor","publisher":{"@id":"https:\/\/zenergetic.eu\/en\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/zenergetic.eu\/en\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/zenergetic.eu\/en\/#organization","name":"Zenergetic Advisory","url":"https:\/\/zenergetic.eu\/en\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/zenergetic.eu\/en\/#\/schema\/logo\/image\/","url":"https:\/\/zenergetic.eu\/wp-content\/uploads\/2026\/03\/cropped-Facebook_profilowe.png","contentUrl":"https:\/\/zenergetic.eu\/wp-content\/uploads\/2026\/03\/cropped-Facebook_profilowe.png","width":850,"height":462,"caption":"Zenergetic Advisory"},"image":{"@id":"https:\/\/zenergetic.eu\/en\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/www.facebook.com\/zenergeticadvisory\/","https:\/\/www.linkedin.com\/in\/agnieszkajorczyk\/","https:\/\/www.instagram.com\/zenergeticadvisory\/","https:\/\/www.youtube.com\/@ZenergeticAdvisory\/"]},{"@type":"Person","@id":"https:\/\/zenergetic.eu\/en\/#\/schema\/person\/081d74a43a319f8990e5b9f1758a6d0c","name":"ajorczyk","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/secure.gravatar.com\/avatar\/ebdadd2e5c67a3dcf68915dcd68fb27563c9ce92f9f5ed1026bce2161c619e56?s=96&d=mm&r=g","url":"https:\/\/secure.gravatar.com\/avatar\/ebdadd2e5c67a3dcf68915dcd68fb27563c9ce92f9f5ed1026bce2161c619e56?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/ebdadd2e5c67a3dcf68915dcd68fb27563c9ce92f9f5ed1026bce2161c619e56?s=96&d=mm&r=g","caption":"ajorczyk"},"url":"https:\/\/zenergetic.eu\/en\/author\/ajorczyk\/"}]}},"_links":{"self":[{"href":"https:\/\/zenergetic.eu\/en\/wp-json\/wp\/v2\/posts\/3204","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/zenergetic.eu\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/zenergetic.eu\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/zenergetic.eu\/en\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/zenergetic.eu\/en\/wp-json\/wp\/v2\/comments?post=3204"}],"version-history":[{"count":14,"href":"https:\/\/zenergetic.eu\/en\/wp-json\/wp\/v2\/posts\/3204\/revisions"}],"predecessor-version":[{"id":3416,"href":"https:\/\/zenergetic.eu\/en\/wp-json\/wp\/v2\/posts\/3204\/revisions\/3416"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/zenergetic.eu\/en\/wp-json\/wp\/v2\/media\/3214"}],"wp:attachment":[{"href":"https:\/\/zenergetic.eu\/en\/wp-json\/wp\/v2\/media?parent=3204"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/zenergetic.eu\/en\/wp-json\/wp\/v2\/categories?post=3204"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/zenergetic.eu\/en\/wp-json\/wp\/v2\/tags?post=3204"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}