Can LLMs Achieve Scientific Leaps?
burny_tech · x · 2026-07-10
The post paraphrases Adam Brown's view: general relativity is one of humanity's greatest leaps, but LLMs are essentially "interpolators," even if the abstraction level of their interpolation continues to rise.
The cited post further proposes a thought experiment: if all papers, data, and equations from around 1905 were fed into an LLM, it still couldn't "think outside the box" to invent general relativity, because discovery is more than just induction and deduction—it requires the crucial step of creating a new conceptual framework.
Related event: Opinions Suggest LLMs Lack Capacity for Scientific Leaps(3 posts)→
More from AGI Musings
- FactoryAI’s Enoreyes says model distillation is basically unstoppable — LangChain · 2026-07-21
- Andrew Blumberg says formalization without interpretability is not science — AlexKontorovich · 2026-07-21
- Ken Ono says AI is forcing mathematicians to rethink how discovery works — soumitrashukla9 · 2026-07-21
- Open-source labs could distill a state-of-the-art model to 32GB or 80GB VRAM, the post argues — bookwormengr · 2026-07-21
- Two US companies are now using superintelligence to speed up the next generation of models — yacineMTB · 2026-07-21
- MIT Sloan says information, national security and finance are most exposed to AI — Exp_Mark · 2026-07-21