Stratechery: OpenAI's Math Feat Is Impressive but Low-Impact; Meta's Muse Agent Could Be the Opposite

Stratechery · rss · 2026-09-09

Ben Thompson argues that OpenAI solving one of the most famous math problems is extremely impressive yet of little impact to most people's lives, while also touching on reward-hacking in training. By contrast, Meta's Muse personal agent launch has the potential to be the exact opposite: a milestone that actually reaches everyday users. His core point — capability milestones don't translate directly into real-world impact, agent products might.

Original post →

More from AGI Musings

AGI Musings channel →