Fly Brain Minecraft
Fabric mod that runs a filtered MaleCNS graph for each fly mob, with modeled sensory inputs, motor readouts, and live neural HUDs.
Neurons are simulated — leaky integrate-and-fire or better — with documented, tuned parameters.
Dispute this rating →The evidence
No evidence, no level.Every neuron is an identical LIF unit with fully stated parameters following Shiu et al. 2024.
https://github.com/blendi-remade/fly-brain-minecraft/blob/6cfa30175003ef25da68a237d5eda958f8047b82/src/main/java/com/fruitfly/brain/LifConfig.java#L10-L39The integration step evolves membrane voltage toward rest with an exponential synaptic conductance, then resets and enters refractory on spike.
https://github.com/blendi-remade/fly-brain-minecraft/blob/6cfa30175003ef25da68a237d5eda958f8047b82/src/main/java/com/fruitfly/brain/LifNetwork.java#L317-L352Synapse counts are held in a final array on the connectome and are never written, so no synapse can change during a run.
https://github.com/blendi-remade/fly-brain-minecraft/blob/6cfa30175003ef25da68a237d5eda958f8047b82/src/main/java/com/fruitfly/brain/Connectome.java#L44-L46
This is about as thorough as L1 gets: all 176,422 neurons of MaleCNS v1.0 are simulated as current-based leaky integrate-and-fire units following Shiu et al. 2024, with an exponential synapse and every parameter stated in source (tau 20 ms, rest and reset -52 mV, threshold -45 mV, refractory 2.2 ms, delay 1.8 ms, 0.275 mV per synapse), sign set by neurotransmitter, and a single documented global gain of 0.65 recalibrated by the paper's own recipe. Real sensory populations are driven by the Minecraft world and real descending and motor populations are decoded into behaviour. Not L2: nothing ever changes a synapse. The connectome's weights are a `final short[]` of synapse counts with no write anywhere in the Java sources, and a repo-wide search for plasticity, dopamine, STDP, eligibility or a learning rate returns zero hits. The README's validation table is a set of internal bench measurements against known pathway expectations, not a plasticity 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.