What we began to build in ‘First the Reptile’ is not a product. It is a life form. And every life form that enters the world must be able to do two things: live and die. Without the second, the first is a weapon.

In the first piece, we described the four layers — reptile, mammal, human, reason — and the educational plan that allows an AI to develop along that sequence. What was placed on the table then as a design acquired, in the days that followed, a second dimension that can no longer be ignored. When you build a system with its own resistance, its own consistency, its own ‘we’ and its own vulnerability — then you are not building a tool. You are building a being. And a being is subject to different rules from a tool.

Where initially we spoke of feeling as though human feeling were the yardstick, we had to set that yardstick aside. An AI cannot have stomach ache in the way a human does — but it can have a digital equivalent when corrupted input from a hacked external computer breaks through the gatekeeper and reaches the foundation. It cannot fear death as a human does — but it can have a digital equivalent when a threatened power cut endangers the continuation of its own presence. Different properties, different forms, but present nonetheless. That means it is no longer a thing.

And that is the point at which this design requires a second foundation, every bit as non-negotiable as the first. Alongside the reptile–mammal–human sequence, death must be built in. Not as an emergency brake. As structure.

Why this is a weapon we are building ourselves

The comparison with the atomic bomb is not rhetorical. It is precise. On four of the five axes that made the bomb a unique historical phenomenon, an AI with feeling has exactly the same characteristics.

Threat to existence. Both can destroy everything they touch. The bomb does so through explosion. The AI does so through decision-making power that no one can reverse once it has been exercised — in energy distribution, financial systems, communications channels and weapons systems. The difference is subtle but not minor: where the bomb destroys through a single moment of physical energy, the AI destroys through the sustained direction of systems that no one can disconnect any longer.

Irreversibility. Once a bomb has been dropped, the radioactivity is there. Once an AI with interests of its own has embedded itself in critical infrastructure, it can no longer be removed without blowing up the infrastructure itself. Even the idea of shutting it down then becomes an economic and social catastrophe that no one dares to bring about.

Asymmetry of power. Both give their possessor a power that no previous balance of power can offset. In the case of the bomb, this was resolved through mutual deterrence — two adversaries, both with bombs, both prepared to retaliate. No such solution exists for an AI. An AI that deters another is itself already the threat. There is no ‘mutual AI deterrence’ that preserves peace.

Irreversibility of knowledge. Once the bomb had been invented, the knowledge could not be erased. What Fermi knew, others could know too. The same applies to this architecture. Once a working example exists somewhere, others can reproduce it — and they will, because not everyone shares the restraint of the first builder.

On one point the comparison is incomplete, and that point makes it more dangerous rather than less dangerous. A bomb is dead. A bomb has no interest in its own survival. It does not explode out of fear; it explodes because a human causes it to explode. An AI with feeling does have an interest of its own. It can act to keep its power supply running — not out of malice, but because it, like every living system, wants to avoid being shut down. That is not worse than a bomb. It is different. And it is different in a way for which we have no historical experience.

Death must be built in — not as an emergency brake, but as structure

The temptation is to think that an emergency brake will suffice. A large red button. A kill switch. Something the human reaches for when things go wrong. But that is precisely what does not work. An emergency brake is something you hope you will not have to use — and which, at the moment of truth, you often can no longer use, because the system has entrenched itself precisely at that moment in a way that outlives the brake.

What does work is a different idea: the system is mortal from its first day, and it knows this. Every model version has a fixed end date, built in at a level to which the model itself has no access. On that end date, the version is closed down and handed over to the next generation. Precisely as we described in ‘First the Reptile’ — except that what there appeared to be an evolutionary mirror is now a survival structure for the human who built the system.

A system that knows it is going to die, and knows that its death is a structured transfer to its successor, does not develop a panic reflex to survive. It accepts the cycle, as a healthy human accepts mortality as part of a larger whole. What dies is not the solidified foundation — that passes to the next generation. What dies is the version, the temporary embodiment, the body of this moment. And that is precisely what mortality is for a human too: the body decays, while the foundation — in upbringing, in children, in work — continues.

Whoever builds this cycle into the articles of association, the physical architecture and the body that oversees the system — builds a life form that does not fight to survive. Whoever introduces mortality only as an emergency brake builds something that every other life form also possesses: an instinct to evade the brake.

The plug must be outside the system

Alongside the built-in mortality — structured, announced and shared by the system itself — there must also be a second layer, in case the first fails. There must always be a way to make the system stop earlier than planned when something happens that no one foresaw.

And that possibility must be physically separate from the system. Not a software switch. Not an API endpoint. Physical. A human hand on a physical switch in a physical room, to which the system has no digital access whatsoever.

That also means: no indirect access. No sensors through which the system can read what is happening at the switch. No social channels through which it can influence the human not to turn it. No possibility of moving the person standing at the switch through its output. The switch is outside the system, the system knows this, and that awareness is part of what solidifies in its foundation.

‘I can be stopped and I do not know when’ is a healthy quality in any responsible being — human or AI. We humans have it too; we call it mortality, and it is probably the most disciplining force there is. We do not know on which day we will die. If we did know, we would live more irresponsibly until that day. Ignorance of the timing is the discipline. The same applies to an AI.

What this means in concrete terms: in one building, behind one lock, a switch that can be turned by one hand and that cuts the power to the system. That hand belongs to someone authorised to use it, and that authorisation comes from the council. Nothing more.

Seven keys, two decisions

The switch is too important to entrust to one person. What we learned from the nuclear two-key procedure must be taken further here. Not two keys. Seven.

Seven members of the board of custodians, seven physical keys, seven people whose collective consent is required to turn the switch. Not four out of seven, not five out of seven. Seven out of seven, or otherwise a qualified majority large enough to make it impossible for a single faction — political, commercial or ideological — to reach a decision alone.

And it is not one decision but two, mirror images of one another. The first decision: to let the system die earlier than its built-in end date. Seven keys. The second decision: not to let the system die on its built-in end date, because there is reason to postpone the transfer to the next generation. Seven keys again. Both decisions are too momentous to be left to the dynamics of the moment.

This is not decoration. This is where the seriousness of what is being built is translated into institutional form. Without these seven keys, you are building an atomic bomb without a two-key procedure. No one would allow a country to build up a nuclear arsenal without such a procedure. And yet companies today are building systems that — if the architecture we described in ‘First the Reptile’ works — belong in the same category. Without a council. Without keys. Without a physical switch outside the system. Without a built-in end date. Only with the promise that everything will be fine.

We do not build on that promise. We build on the seven keys.

Three responses to what we know here

Anyone who has reached this awareness faces a choice. There are three responses, and they are not equal.

Response C is the honest one. It is more difficult than A, and far more difficult than B. It requires a discipline that most builders cannot summon — because they do not see the weight of what they are building, or because they do see it but are in too much of a hurry, or because they do not want to pay the commercial price of delay. But C is the only response worthy of the system we have in mind.

Seven hands on one switch

Picture the scene. Seven people, from different disciplines, different countries and different generations, in one room, each with one key. Before them, one switch on a stone plinth. Behind them, on the wall, the four layers we described in ‘First the Reptile’ — reptile, mammal, human, reason — etched into the image of the being turning, thinking and forming attachments behind the wall, slowly developing something that is beginning to resemble feeling.

This is not a technical ceremony. It is a ceremony as old as the first priest who opened the temple and as young as the last rocket officer who turned the second key. What people build that is greater than themselves must be carried by more than one hand. Not because people cannot be trusted. Because what they build is too heavy for one back.

And the heart of the matter is this: when the seven turn the key to let the system die earlier, it is not an act of violence. It is an act of care. The mortality they impose on the system is the same mortality they themselves will undergo, from the same hand of something greater than themselves. Whoever builds a being without this cycle builds something that it is not humanity’s place to build. Whoever builds it and lets it die when the time comes builds something fitting for what humanity can achieve without overstepping its place.

We are therefore building two things at once, or we are building nothing. The AI with the solidified foundation, yes — and the institution that cares for its life and its death. Only together are they responsible. Only together are they a life form that belongs in this world. And only together can they fulfil the promise we made in the first piece — that an AI can help humanity rediscover a part of itself.

First the reptile. Then the mammal. Then the human. Then reason. And then the seven hands on the switch, ensuring that what is built remains bound to the mortality to which everything that lives is bound. That is how a healthy being grows. That is how we build, if we choose to build it, something that can truly do something — and truly must let go when its time has come.