Hinton: we design the learning algorithm, but we don't really understand how neural networks work
austinc3301 · x · 2026-10-03
Geoffrey Hinton, often called the godfather of AI, says that while we designed the learning algorithm—"a bit like designing the principle of evolution"—the complex neural networks it produces when interacting with data are good at doing things, but we don't really understand exactly how they do them. The poster adds that any good ML researcher will admit interpretability lags far behind model complexity, likening it to knowing what the brain is made of without understanding how it works.
More from AGI Musings
- Consultant asks the endgame of AI sales agents selling to AI buyer agents — TrainerElectronic765 · 2026-10-03
- Intelligence and consciousness are likely orthogonal — though they may correlate in practice — austinc3301 · 2026-10-03
- The AI consciousness debate is a psyop: uncertainty alone demands ethics before certainty — RileyRalmuto · 2026-10-03
- Yann LeCun: two years later we're still far from human-level AI — your cat beats it at physics — ylecun · 2026-10-03
- Researchers warn we may build conscious AI before understanding consciousness itself — CurieuxExplorer · 2026-10-03
- SF one-bedroom relisted at $6,000 after $4,000 two years ago as AI economy inflates rents — Yuchenj_UW · 2026-10-03