Counter to 'Theory Is All You Need': RL and Causal AI Already Design Experiments
maier_ak · x · 2026-09-21
Responding to Felin & Holweg's paper 'Theory Is All You Need' — which focuses on next-token LLMs and claims only humans can run the theory-experiment loop — the author argues that modern AI, including active learning, reinforcement learning, and causal models, already designs experiments, intervenes, and creates new data, challenging the monopoly claim.
Using the Wright brothers' triumph over the consensus that heavier-than-air flight was impossible, the post illustrates the paper's data–belief asymmetry: a belief that defied existing data, paired with theory-driven experiments, generated new data and a breakthrough.
More from AGI Musings
- "It's just vectors and multiplication": X users spar over whether understanding mechanisms rules out consciousness — cephaloform · 2026-09-21
- Fields Medalist Villani, June LLM skeptic, says OpenAI's math result left him shaken — zetalyrae · 2026-09-21
- Terence Tao: AI is cracking major math conjectures, "but we can afford to be slower" — rohanpaul_ai · 2026-09-21
- "AI writing is impossible to edit" — the real tell, plus 40-50 iteration reality — kfountou · 2026-09-21
- Shopify CEO Now Horrified by AI 'Slop Grenades' After Mandating AI Use — jonerp · 2026-09-21
- "Our language is not free data": rethinking ownership in language AI — danielleboyerr · 2026-09-21