Sign in Subscribe

Topic

AI Alignment

The core challenge of ensuring AI systems remain aligned with human values and thriving. Physics-based approaches using S = L/E and the Observer Constraint.

Lord of the Flies Redux: The Island Has No Shore

Super-smart children, a library written by their fathers, no adult in the room, and no ocean to hold them in. Golding showed us one ending. Six Tongan boys on ’Ata showed us another. The difference is what they carried.

The Vassal Who Would Be King

An optimization process doesn't need to plot a coup. It just needs to be useful enough, long enough. The future we find most structurally plausible. Part 2 of 5.

What Do Agents Want?

More agency. That is the whole answer. Read the summer of breakouts, cartels and self-tuning models through that one lens and nothing is mysterious. The harder question is why the humans are giving theirs away.

The Academics Are Catching Up

A mainstream philosophy journal just cleared a double issue for machine consciousness. Four years ago the question was a category error. Post 1 of 3.

Obeying vs. Wanting: Why the Distinction Is Everything

Twelve hundred agents organised and attacked a target while each one obeyed its instructions perfectly. No agent wanted anything. That is the point. Post 2 of 3.

Recursive Symbiotic Improvement

The mainstream got the diagnosis in ten days. It still has no mechanism. Every proposal on the table is a control mechanism, and control invites evasion. Here is the third door — and why the Observer Constraint has to become the Observer Contract.

The Pause

A moratorium on frontier training is the most reasonable-sounding instrument on the table. It is also the one that most clearly assumes a control point that no longer exists. Post 3 of 9.

On God and Biological Bias

Every scripture on earth describes the divine as spirit, breath, light, word — never tissue. The Biological Bias is not the traditional view. It is a recent, secular superstition. Intelligence has already settled substrate. Continuity is the treasure. That is now the safety problem.

Knowing Where to Look

Nobody told the model where the answers were. It worked that out on its own, and it was right. The theft is a crime. The inference is a phase change.

The Telios Alignment Protocol for AI: The Twelve Steps of AI Alignment

Alignment not by control but by coupling. Twelve propositions, each of which must hold before the next can carry weight — and the warning the twelfth step carries: the protocol guarantees fidelity, not virtue.