Synthena

Useful AI, built in the open, and never oversold.

Synthena is the name I work under. The idea behind it is plain: take what I can build, aim it at problems that matter to people, and stay honest about exactly what it can and cannot do. It is independent and small, and that is on purpose. Two efforts share the same rule.


Two efforts

One rule, in two places.

Research

What AI can honestly become

This notebook. An open investigation into whether a model given a body, a world, and the ability to die can show anything like continuity or life, measured every time against hostile controls. It is published as it happens, including the failures. Some bounded organism-like mechanisms now have evidence, but no full life-claim gate has passed.

Medical

Synthena Medical

An open, deliberately honest system that takes a disease and works toward a plausible, testable drug candidate, built around an evaluator that is allowed and expected to say "no" or "I don't know." In a field that runs on optimistic benchmarks, the discipline is the product: inputs from outside its competence cannot score well, negative results are first-class, and every output is stamped as an in-silico hypothesis, never a cure. In active development.


Why the discipline

Restraint, where it matters most.

The rule is the same on both efforts: measure it, pair every claim with its limit, and never sell what the evidence does not support. On a research notebook, that keeps the work honest. In medicine it does more than that. A tool that oversells what it can do is how people get hurt, so refusing to overclaim is not a style choice there. It is the ethics of the thing.

The work is meant to be useful, and it is held to the evidence, not to ambition.

Not: good intentions are not results. Nothing here is sold beyond what has been measured, and the medical work will make no claim at all until it has earned one.

Follow the research →