MIT professor's essay: universities must prepare for AI smarter than all of us
MIT News AI · rss · 2026-10-09
Sasha Rakhlin, director of MIT's Statistics and Data Science Center, published a long essay on how AI is reshaping academia, especially graduate research and education:
- Why progress is so fast: The pace of AI in a discipline tracks the speed and reliability of verification — formal proofs and runnable code let systems self-improve. A model hit IMO gold last year; models now produce new research results. Universities should prepare for a near future where AI is smarter than most of us in most intellectual work.
- Rethinking credit and training: A polished paper is becoming a weaker signal of individual expertise; good questions, replication, synthesis, informative negative results, and shared datasets deserve more recognition, with explicit intellectual-responsibility norms when AI does much of the work. Training must distinguish unnecessary friction from formative work.
- Universities' strategic asset: Tacit lab knowledge — failed experiments, abandoned directions — could make models far better at scientific exploration. He envisions MIT labs connected through shared AI research infrastructure, with agents cross-linking methods, benchmarks, and researchers, requiring major investment now in compute, secure data systems, and post-training expertise.
More from AGI Musings
- Stanford and Databricks use AI agents to reproduce medical findings and test analytical robustness — ChengleiSi · 2026-10-23
- ISMS Fellow Avi Goldfarb: AI disruption raises returns to exploring research directions — avicgoldfarb · 2026-10-23
- signull on AI-found PHP exploits: economic liability is the forcing function legacy software needs — signulll · 2026-10-23
- TheStalwart: Talking computers are sparking a new philosophical boom, echoing ancient Greece — toptickcrypto · 2026-10-23
- AI cyberattacks could paradoxically accelerate AI development in an offense-defense arms race — kimmonismus · 2026-10-23
- Economist on AI detector Pangram: not a filter, but a useful verification signal — paulnovosad · 2026-10-23