Fly connectome reveals fast-weight continual learning neurons — a skill current LLMs lack
tszzl · x · 2026-09-15
- Google Research and HHMI Janelia have completed the first full map of an adult male fruit fly's brain and CNS, using AI to fuse millions of 2D images into 3D neural shapes covering 166,000+ neurons — a record.
- Analyzing the connectome, researchers identified a group of neurons (hΔH, hΔA, hΔI, hΔG) that could let the fly navigate via fast synaptic weight updates instead of neural activations — effectively fast-weight continual learning in a fly, something current LLMs don't do.
Related event: Fly connectome reveals fast synaptic-weight learning for navigation(2 posts)→
More from AGI Musings
- Kevin Roose's Machine Gods pod: why scientists are mad at OpenAI solving their hardest problems — kevinroose · 2026-10-30
- Cryptographer Matthew Green: Anthropic's agent worm framing is unhelpful, we need a simulated one to know — matthew_d_green · 2026-10-30
- Ethan Mollick on AI and work: 'You're a hand-weaver in a world of power looms' — emollick · 2026-10-30
- Vercel founder Rauch: trust is the deepest moat for a post-AGI business — soleio · 2026-10-30
- Scattered Across AI Assistants: A New Kind of Cognitive Failure Nobody Has a Prior For — manosaie · 2026-10-30
- a16z partners: 10-20M people drive $150-200B in token spend, and AI workspaces will eat the browser — petergyang · 2026-10-30