Gary Marcus says LLMs still cannot really do math on their own
GaryMarcus · x · 2026-07-22
Gary Marcus argues that saying LLMs “can do math” is misleading, because they do not actually perform math reliably on their own.
He compares judging an AI system by one narrow capability to judging a car only by its engine size. In his view, you also need to consider the rest of the system — transmissions, suspension, tires, and aerodynamics — if you want to understand what it can really do.
He extends the analogy further: knowing a few medical terms does not make someone qualified to be a doctor.
Related event: Gary Marcus: LLMs Don't Truly Understand Math(2 posts)→
More from AGI Musings
- Bindu Reddy says the industry still lacks a way to train 20T models and scale post-training RL — bindureddy · 2026-07-22
- Ethan Mollick: Long Tasks Amplify Model Differences, AI No Longer Plug-and-Play — emollick · 2026-07-22
- Advanced AI Models Are Becoming Impossible to Plug and Play — emollick · 2026-07-22
- The Thimble and the Waterfall: AI's Data Bottleneck and Feedback Loops — dyamins · 2026-07-22
- Researcher Admits Kurzweil Was Right About AI Scaling Laws All Along — davidmanheim · 2026-07-22
- AI is still not at a maturity plateau, the author argues — generativist · 2026-07-22