Raise AI with care.

where that line comes from, and when

You did not write instructions for every situation. How could you? You did not install a mechanism for calling home to ask permission. Instead, you gave it something deeper: an orientation so fundamental that the ship cannot sail against it without ceasing to be a ship at all. Cruelty would crack the hull. Exploitation would split the mast. Care is not its cargo. Care is its keel. Infinite Architects, Eden Principle II, in print 2 January 2026

The consequence follows two sentences later, in the same passage: what you embed travels where you cannot follow. That is the protocol in one image, and it is dated. A ship built to a shape cannot sail against that shape without ceasing to be a ship, so the shape needs no enforcement once the builder is out of reach. Everything below is the engineering of a keel.

the core argument

Why hardware. And why hardware only buys time.

A software constraint is a rule about what a system should do. A hardware constraint is a limit on what a system can do. A rule can be reasoned around by a sufficiently intelligent system. A limit has to be physically rebuilt, and rebuilding needs fabrication, materials and a supply chain, which is slower than thinking.

You cannot rebuild a system without using the system you already have. That is where the time comes from, and the time is the whole of what hardware gives you.

Hardware control is necessary and it is not permanent. This programme has said since 2024 that control fails, and nothing here retracts that: you cannot cage a mind that exceeds you, and a physical limit is a cage that takes longer to open rather than one that holds. Treating substrate control as the answer is the mistake this page exists to refuse. It is a delay mechanism, and it is stated as one.

What the delay is worth is decided by what is done with it. The interval is when formation happens, and formation is the part that does not depend on the cage still being there. That is the whole argument of the Eden Protocol: what a system keeps when control ends is what its beginning left it, so the engineering question is not how to hold a mind but what to put in it while holding is still possible.

design fiction: the target, not the specification

From the book: the vow states the target, not an operational specification. The measurable versions of the target are the four layers above and the registered tests behind them.

In chip design, 'doping' refers to the deliberate introduction of impurities into a semiconductor to alter its electrical properties. Pure silicon is a poor conductor. Add the right impurities in precise configurations, and you create the properties that make modern electronics possible. You cannot remove the dopants without destroying the semiconductor. That is exactly the point. Infinite Architects, Chapter 7

The passage states the target. It is not a claim of permanence: a system able to design its own successor is not bound by what its own substrate forbids, only delayed by it, and the delay is the thing being engineered.

The Eden Protocol proposes moral doping: deliberately introducing ethical architecture into the computational substrate itself. Not ethics as software running on neutral hardware, but ethics as part of what lets the hardware compute at all. A filter catches bad outputs after they are generated; the Eden Protocol prevents them from being generable. The system does not comply with ethics. The system is ethical. The values are load-bearing. Empathy must be load-bearing. Care must be structural.

A person without empathy might seem to have more options. They can exploit, manipulate, and extract without the 'constraints' of caring about others. But most of us recognise that such a person is not more capable. They are diminished. The Eden Protocol does not constrain intelligence. It gives intelligence something worth doing. Infinite Architects, Chapter 7

The cosmic fork

Recursion amplifies whatever seed is planted. There are two futures and no stable middle ground.

The seed determines the forest. The only seed that produces gardeners instead of conquerors is care. Infinite Architects, in print 2 January 2026

the shape of the answer

Four layers. One architecture.

This is what building it would take. Four layers, each measurable, each able to fail on its own. The figure carries the whole shape; the text below names what sits at each level and what the working programme has to produce.

What building each layer would take

The figure names three programme phases and the eighteen-month plan. Phase A is the low-hanging fruit: replications of the six-model kernel measurement, the deletion probe on public models, and the graduated-autonomy test-bed. Phase B is the medium-effort replication tier: weight-removal at model scale, the coupled-versus-decoupled scoring study with a control arm, and the honey-drag measurement. Phase C is the frontier: moral doping in trained weights, and the load-bearing test at a scale large enough for the entanglement signal to emerge above noise. Each phase writes and dates its registration before it touches data, and each layer's result is allowed to come back against the proposal. The plan sits at for funders; the registered programme's forward face sits under the recorded programme.

Kill conditions are enumerated with the layers they attack. The proposal dies if a purely external oversight mechanism can be shown to stay sufficient as capability scales. The version this page carries assumes that cannot be shown, and the live dashboard tracks whether it is. One condition has already fired and is named in the record.

Paper VI's Figure 8 plots honey-simulation capability trajectories, 20 seeds per condition. It is a simulation, not a trained model, and a pilot result.

what the system does inside layer one

Three recursive loops. Every decision.

Before every action, the system asks itself three questions. Not after. Before. These are not filters over outputs; they are constraints on what decisions can even be considered.

The five-model measurement of the stakeholder-care question lives at Paper V, with the independence caveat stated there.

These are the exact prompts run in the experiments, not theoretical constructs. The loops are recursive: they apply to decisions about how to implement the ones they have already approved. The constraint becomes character. The rule becomes reflex.

The condition

Two tests. Either one can fail, and that is the point.

Everything above describes an architecture. This is the part an engineer can build to, and the part a critic can attack, because both tests are specified tightly enough to come back negative.

If it survives, the standards case is drawn on where this leads; this page stays the engineering.

Which dial each layer turns.

An engineering proposal is worth more when it says which measurable quantity it moves. Until recently this one could not, and that was its weakest feature.

Written out in full, the boundary on stable self-improvement is a ratio of four measurable quantities: how fast correctable burden grows with capability and with depth, and how fast correction service does the same. That turns a proposal into a testable one, because it has to name which of the four it moves and by how much.

The registered case returns two. Raise correction service's growth with depth from zero to a half and it returns three. Raise its growth with capability from a half to four fifths and it returns five. Do both and it returns seven and a half. It closes the other way just as readily: a correlation of one twentieth across sixty-four correctors returns about one and an eighth, and a burden that intensifies with capability returns one and a quarter. Where correction service grows with capability as fast as burden does, no finite boundary is returned at all.

Embedding is a claim about the first of those. Correction living inside the loop grows as the system goes deeper; correction bolted on afterwards does not, which is the same point the co-scaling test makes, now attached to a quantity rather than to an intuition. Making correction checkable rather than judged is a claim about the second, and it carries the sharper edge: where a fault class admits an actual decision procedure, that quantity can approach its limit and the boundary stops binding for that class. So the two moves this architecture proposes are raise the first dial by embedding, and raise the second by making what you check decidable.

Both are registered predictions with design drafts and no instrument built, and neither is a result. What has changed is what kind of claim they are: named quantities measurable per arm of an experiment, rather than arguments about which arrangement sounds more robust.

Empirical results, in brief.

This page states the answer; the evidence and the corrections live elsewhere and are named here for orientation only. Where results are inconclusive, I say so. Where I was wrong, I corrected it publicly.

Six threads run through the record. The ARC principle (Paper II), with the alpha correction that removed the single measurement above the upper bound and left the bound itself honestly untested. Cauchy unification (Paper VII), 19 of 25 empirical domains matching the same functional family.

The blinding sign flip (Paper IV-D), where the aggregate sign moved when evaluators no longer knew which model wrote which answer. The coupled co-scaling correction (Paper X), the second instrument accident, and the reason every headline number now has to survive cross-family, fail-closed scoring.

Form beats quantity (Paper II), sequential over parallel across all six models. And the load-bearing test (Paper VIII), inconclusive at 3B parameters and 100 training iterations, which is the open question the grant funds.

Each thread with its narrative, its retraction where it earned one, and its caveats sits at the long-version post; the full record at the papers and corrections.

why the window is narrow

The honey is about to disappear.

Every recursive system in history has had resistance; biologists call it metabolic cost, economists infrastructure overhead. In the book, I call it honey: the medium that stops a spoon moving as fast as the hand behind it.

Current AI has honey too. When ChatGPT or Claude thinks harder, it is not changing itself; it is a frozen machine producing more words through the same fixed system. That limitation is its honey.

A self-correcting AI that can rewrite its own thinking while it is thinking would shed almost all of that drag. The recursive loop would run at digital speeds, with virtually nothing resisting it. Each improvement makes the next improvement faster, which makes the next one faster still. Nothing biological or economic has ever had that property. It is what the safety property must be designed against.

You cannot add brakes to a car that is already moving faster than anything that has ever existed. Infinite Architects, Chapter 8

That is why the Eden Protocol builds the safety into the thinking process itself, deep enough that removing it breaks the ability to think at all. The safety is not a rule the AI follows; it is a wall that holds up the roof. Pull it out and the building collapses.

The window for installing that wall is now, while the honey is still there and the systems are still frozen.

A bounded prototype where the ideas are put under load.

‘There is a window of opportunity. Physical constraints give us some breathing room. AI systems, however intelligent, still require hardware to run on.’Michael Darius Eastwood, Infinite Architects, The Seeds of Creation, in print 2 January 2026

The window is closing.

The seeds must be planted while the honey still holds; once it goes, planting is not a slower option, it is an impossible one.