Why AI Struggles to Become Einstein: Paper Highlights Lack of Principle Discovery
OdedRechavi · x · 2026-08-01
The paper "Can AI Follow In Einstein's Footsteps?" by Ido Kaminer and colleagues explores the fundamental limitations of current AI in physics discovery.
The authors note a striking reverse trend: while human physics progressed from pattern prediction to phenomenological laws and ultimately to principle-based theories like relativity, AI's trajectory has gone backward. Early AI focused on equation-discovery methods like symbolic regression, whereas modern frontier models like AlphaFold are powerful predictors that fail to provide clear theoretical understanding.
The paper highlights that AI's most critical missing skill is the ability to pose the right questions or invent guiding principles. It references Demis Hassabis's proposed "1911 cutoff test": removing relativity from training data to see if AI can independently rediscover Einstein's principles.
Related event: Research Finds LLMs Lack Abductive Leap Ability for Scientific Discovery(2 posts)→
More from AGI Musings
- Is math just a language? A debate on why LLMs are good at mathematics — LucaAmb · 2026-10-16
- Paul califax argues extrapolation isn't special: LLMs can learn out-of-distribution skills via high-dimensional interpolation — paul_cal · 2026-10-16
- willcb: higher-level behavior shaping via RL envs is 'almost practical' — willcb · 2026-10-16
- Tech improves far faster than it diffuses — organizational bottlenecks remain the key intuition — Afinetheorem · 2026-10-16
- Ramez Naam weighs in on mech interp blog hype: bullish overall, lacks tools to verify claims — ramez · 2026-10-16
- Mistral CEO Arthur Mensch: Europe abandoning frontier AI would mean unprecedented dependence — arthurmensch · 2026-10-16