AI researcher: deep learning 'doom' papers existed every year, yet progress keeps accelerating

RazRazcle · x · 2026-09-06

Responding to stagnation claims, RazRazcle argues that inside AI research everything feels insurmountable because attention fixates on what's not working—but zooming out, the trend is undeniably accelerating. He notes that at every mainstream ML conference over the past decade you could find numerous papers suggesting deep learning would stop working, yet it kept improving.

Original post →

More from AGI Musings

AGI Musings channel →