One cortical neuron needs a 5-8 layer net to model — backprop has no biological analog
vishalmisra · x · 2026-09-27
Biological neurons are far richer than artificial ones
The author lays out evidence that the artificial neuron is a cartoon of the real thing:
- Depth: modeling a single cortical neuron requires a 5-to-8-layer deep network (Beniaguev et al., 2021).
- Computation: human dendrites compute XOR—the very function Minsky and Papert proved a perceptron cannot (Gidon et al., 2020).
- Learning rule: backprop needs the backward pass to reuse forward weights, and no known biological mechanism does this. Real neurons spike, keep fixed signs, respond to dopamine and glia, and run the whole brain on 20 watts.
Related event: "The Brain Is a Neural Network"? Misra Clashes with NeuroAI Researchers(7 posts)→
More from Research
- Google Research open-sources PaperOrchestra, a multi-agent framework for writing AI papers — google-research · 2026-11-16
- Kind ICLR Reviewer Raises Score Despite Unresolved Concerns, Citing the Paper's Value — MarkSchmidtUBC · 2026-11-16
- Hand-editing 0.2% of weights plus 100 CPU-minutes turns an English TTS model into an 18MB Sanskrit one — kastnerkyle · 2026-11-16
- Why chirality matters for AI for Science: enantiomers behave differently in biology — madeofAjala · 2026-11-16
- RSS 2027 submissions open with FAQ clarifying two-stage review and negative-result bar — GeorgiaChal · 2026-11-16
- Mobile AI Factory: agents auto-build on-device SLMs with 19x less GPU time — pmttyji · 2026-11-16