Karpathy Ran 8 Parallel AI Agents on a Research Task. His Verdict: 'It's a Mess.'
Andrej Karpathy stress-tested multi-agent research automation and found it nowhere close to reliable — a reality check from one of the field's most credible voices as agent hype reaches fever pitch.
Andrej Karpathy, the former Tesla AI director and OpenAI researcher, tested a setup with eight parallel AI agents working on a research automation task. His conclusion was blunt: "Doesn't work. It's a mess," as reported by @thebuildrweekly and echoed across the AI development community.
Unlock the full briefing
Get every story in today's briefing, the full archive, and the daily AI intelligence brief.
All stories today
Full archive
Daily brief
Cancel anytime. Payments powered by Stripe.