Stanford AI account mocks 'recursive self-improvement' vs 'repetitive slop injection'

bengoertzel · x · 2026-11-18

The Stanford deep learning account (dileeplearning) reposted two old news screenshots (United's widespread system outage, Chase resuming operations after a software bug) with the question: 'Do these look like recursive self improvement or repetitive slop injection?' — a jab at circular, AI-generated pipeline content. Ben Goertzel amplified the quip.

The underlying debate — whether current AI self-replication loops amount to capability evolution or just slop injection — is circulating in AI circles.

Original post →

More from AGI Musings

AGI Musings channel →