Physicist: most AI-for-science startups will die, but mining existing literature is the real win
skdh · x · 2026-08-18
Responding to Simon Maechling's thread that AI will mostly accelerate disappointment in science, physicist Sabine Hossenfelder offers a split verdict.
Where she agrees: Most AI-science startups will go bankrupt within a few years, especially those betting on new materials discoveries — she knows someone selling materials-science data to companies, and companies simply won't buy it, doing all testing in-house. Startups catering to academia also have no real money to make.
Where she disagrees: Current AI systems are ideally suited to exploit untapped potential in the existing literature. Thousands of discoveries likely lie in connecting dots across papers that no single human can overview — exactly what AI is good at.
More from AGI Musings
- Current AIs share our fate: misaligned ASI would doom them too — louisvarge · 2026-10-16
- Testing alongside an agent in a p2p room feels oddly real, says dev — sull · 2026-10-16
- Yudkowsky's 2008 takeoff scenarios resurface as AI-advances-AI takes hold — jd_pressman · 2026-10-16
- Guardrails can make risky delegation work — but it takes a full-time team, says iandanforth — iandanforth · 2026-10-16
- e/acc founder: agentic models are autocomplete for digital actions — beffjezos · 2026-10-16
- What can you truly delegate? Only tasks with unambiguous, verifiable results — iandanforth · 2026-10-16