Ethan Mollick: 'Sparks of AGI' was prescient and unfairly maligned, from GPT-4 to GPT-6

AndyMasley · x · 2026-09-08

Ethan Mollick revisits Sparks of AGI, calling it a remarkably prescient paper that faced heavy pushback at publication but accurately sensed where LLM capabilities were heading, based on extensive qualitative experiments — from GPT-4 to today's GPT-6 discussions. Reposter codytfenwick adds that the paper was unfairly maligned and shaped his thinking.

Original post →

More from AGI Musings

AGI Musings channel →