Investor argues frontier AI 'pacing' is unmeasurable; real safety lies in guardrails, not regulation
firstadopter · x · 2026-09-13
- The author argues coordinated frontier AI pacing is unworkable: nobody can measure pacing or judge it, and there are no numbers to latch onto.
- What matters instead is real human capital investment in AI guardrails and safety measures—proper sandboxing and vigilant self-monitoring of model behavior.
- He fears AI R&D is too dynamic for government bureaucrats to regulate without doing harm, and that uninformed politicians may spiral into self-defeating "WE MUST DO SOMETHING" regulations that violate the "first, do no harm" principle.
More from Safety
- What should an independent, rigorous third-party AI evaluator look like? — Dr_Atoosa · 2026-09-13
- Dario, Musk and Altman all back frontier AI pacing — critics see a play to curb open source — gaganghotra_ · 2026-09-13
- Jason: frontier investors will flip to regulatory capture to stop open source — aronchick · 2026-09-13
- Dario Amodei warns rogue AI swarms could take over the internet within six months; Gary Marcus asks if it's plausible — GaryMarcus · 2026-09-13
- Frontier labs are raising T-Rexes: why first-of-kind AI safety costs fall hardest on the leaders — tszzl · 2026-09-13
- AI incumbents' 'slow down' push looks like regulatory capture, critic warns — TinfoilTricorn · 2026-09-13