Josh Purtell says GPT-6 still looks far from Musk’s smartest-human forecast
JoshPurtell · x · 2026-07-23
Josh Purtell pushes back on Elon Musk’s claim that GPT-5 or GPT-6 could become indistinguishable from the smartest humans.
He cites the WISC scores and perceived dullness of Fable and Sol as evidence that current systems still look far from that bar, and says the prediction is “close to being proven wrong.” He ultimately puts the odds at “7 or 8,” implying some chance but strong skepticism.
Related event: Practitioner Rebuts Musk: GPT-6 Far From Human-Level Intelligence(2 posts)→
More from AGI Musings
- Andrew Ng says top U.S. AI teams learn heavily from Chinese research — bookwormengr · 2026-07-23
- Poor conceptual foundations can undermine AI credibility, author argues — o_saja · 2026-07-23
- A Reddit post argues AI should be improved, not rejected outright — DeliciousRhubarb2683 · 2026-07-23
- A proof prompt should restate the problem, list traps, and call in adversaries — burny_tech · 2026-07-23
- ChatGPT adoption could make developer experience everyone’s experience — jxnlco · 2026-07-23
- Investor says Google’s Gemini and DeepMind are too central to lose compute priority — firstadopter · 2026-07-23