The new PhD playbook: build a benchmark, earn frontier-lab recognition, then raise money for a data startup
teortaxesTex · x · 2026-11-11
A widely discussed X thread argues PhDs today need a "Data & Bench mindset" and effectively have only one viable path: building benchmarks rather than datasets.
- Building quality datasets is out of reach for individual PhDs unless they join a frontier lab's data team — which costs you your academic reputation and impact, so you might as well quit.
- Benchmarks are feasible since you can at least pay for API calls to test them.
- The summarized playbook: build data awareness → create a strong benchmark → earn frontier-lab recognition → raise investment → start a data company.
- Context: good data is genuinely expensive now, and people around the author are reportedly already following this route.
More from AGI Musings
- Viral post: five years of 'maybe next year,' then 2026 arrived — hey_abusiddik · 2026-11-11
- 'Chat with your data' misses the point — people want a smart coworker — vboykis · 2026-11-11
- Ben Todd: AI task time horizons appear superexponential in AECI, not just revenue — ben_j_todd · 2026-11-11
- When AI becomes an oracle, what happens to mathematics? — stevenstrogatz · 2026-11-11
- David Patterson: superintelligence will end profit and equity growth, only sales-tax-funded universal high income works — davidpattersonx · 2026-11-11
- Even after a 6-month recursive leap, AI progress stays 3x the 2024 rate — ben_j_todd · 2026-11-11