'Text-to-neural-network' vision: describe a task and get a purpose-built tiny net
flowersslop · x · 2026-11-03
flowersslop sketches a future capability: describe a niche task — say a camera that detects one insect species while ignoring others — and have ChatGPT spawn a tiny neural network purpose-built for it, not trained but "designed into existence," performing perfectly. The idea echoes tszzl's line "designed and not grown," pointing toward on-demand construction of specialized nets rather than growing them via training.
More from AGI Musings
- Sub-1T models already hit human level; brain's 100T synapses are largely redundant — teortaxesTex · 2026-11-03
- Under 1T params already superhuman — AI models aren't underparametrized, argues researcher — teortaxesTex · 2026-11-03
- Psychics push back as AI-generated readings flood apps and TikTok Live — Chris_Armstrong · 2026-11-03
- Bending Spoons CEO: AI now writes 94% of our code, up from 0% in two years — illscience · 2026-11-03
- Lightcone Commons donor-routing mechanism draws candid caveat from evaluator — NathanpmYoung · 2026-11-03
- New data-constrained scaling laws point to quadrillion-parameter transformers, 7000x Chinchilla — teortaxesTex · 2026-11-03