Prof. Mollick: Lack of Verifiability in LLMs Is Real but Overstated
emollick · x · 2026-08-02
Ethan Mollick, a professor at the Wharton School, argues that while the lack of verifiable answers in many fields is a genuine issue for LLMs, it is often overstated.
He observes that as models improve in formal domains like math and coding, they are simultaneously getting better at handling less-verifiable, complex domains.
More from AGI Musings
- Experts Criticize AI Companies for Downplaying Security Incidents — Miles_Brundage · 2026-08-02
- AI Does Everything From Idea to Design, But Founders Still Pay Humans to Fix Adoption — oykun · 2026-08-02
- AI Safety Advocates: Concerned Doesn't Mean Anti-Usage — dhadfieldmenell · 2026-08-02
- Zuckerberg Mocked for Building Doomsday Bunker While Pitching AGI Utopia — danfaggella · 2026-08-02
- Podcast: Why Liberal Institutions May Survive Even After AGI — willmacaskill · 2026-08-02
- Business Processes Don't Need Perfection: Automate 80% with LLMs, Human-Check the Rest — Franc0Fernand0 · 2026-08-02