Physicist quits tenure-track job for AI safety: LLMs made research coding 'solved'
chaseleantj · x · 2026-09-23
Oz Amram accepted a tenure-track particle physics job 4 months ago — his goal for 10 years. Now he's leaving immediately to work on AI safety, and wrote a long post explaining why.
- Frontier models (Claude Opus 5, ChatGPT 5.6 Sol) went from nearly useless to transformative for physics research within a year, but adoption is jagged and many physicists haven't noticed.
- Coding for research is 'solved' for him: he no longer touches code, delegating everything to agents that write tests and validation plots; this alone boosted his productivity >3x.
- He predicts that within 1-3 years, his role as an ML+physics researcher will be entirely handled by AI — hence the pivot to safety, and a call for the physics community to take AI far more seriously.
Related event: Particle Physicist Leaves Tenure-Track Role for AI Safety(2 posts)→
More from AGI Musings
- Korn Ferry survey of 16,000: AI leaves 52% of workers with more tasks, more burnout — GaryMarcus · 2026-09-23
- eigenrobot: EA lacks authorities to disown shrimp-rights and negative utilitarianism — eigenrobot · 2026-09-23
- Researchers say AI has broken peer review: endless nitpicks, no one actually reads — IanArawjo · 2026-09-23
- People Building AI Are the Ones Most Afraid of It — cneuralnetwork · 2026-09-23
- Frontier models are flooding CVE queues — 90-day disclosure windows must shrink to 30 — chrisrohlf · 2026-09-23
- Education critic Benjamin Riley explains why he's not betting on Alpha School's future — benjaminjriley · 2026-09-23