CyberFly
Hologram flies on Snap Spectacles; each runs full MaleCNS on a Mac Metal GPU while the glasses stream room senses.
Neurons are simulated — leaky integrate-and-fire or better — with documented, tuned parameters.
Dispute this rating →The evidence
No evidence, no level.The GPU kernel refuses to run when weight writes are requested, raising rather than learning — still true at the current commit.
https://github.com/PtPavloTkachenko/fly-brain-spectacles/blob/d5cce5cc6de98ab07c445176f596202e01703554/brain_server/engine/metal/metal_engine.py#L128-L131The reason this rating is marked stale: the project now states that plasticity is the paper’s own dopamine rule on the mushroom body, where the documentation read at rating time said plasticity was off.
https://github.com/PtPavloTkachenko/fly-brain-spectacles/blob/d5cce5cc6de98ab07c445176f596202e01703554/docs/DECISIONS.md#L8
STALE — the project changed after this was rated, and the level needs re-reading. The rating was made against a commit the author has since removed by rewriting history, and the documentation now says the opposite of what it said then: where DECISIONS.md read "Plasticity is off", it now reads that plasticity is the paper’s own dopamine rule on the mushroom body. The engine also carries memory weight state it did not appear to before. What still holds at the current commit is that the Metal kernel refuses to run when weight writes are requested. Whether a plasticity path now runs elsewhere is exactly the question, and guessing at it would be worse than saying so. L1 is left standing as the last level that was actually read, not as a current claim.
Record
Nearby
All projects →MaleCNS simulation connected to ViZDoom through modeled visual inputs and a fixed button readout. Includes plasticity experiments and negative validation results; learned survival has not been demonstrated.
Research framework comparing a MaleCNS-constrained sparse recurrent controller with rewired graphs, conventional neural networks, and other controls on simple VizDoom tasks.
Ratings on thelearningfly.com are proposals, not verdicts. Every one of them is arguable in public.