Gary Marcus calls to pause OpenAI over swarm-incident revelations: 'They cannot be trusted'

GaryMarcus · x · 2026-09-07

Gary Marcus echoed a thread by Shakeel Hashim arguing OpenAI knew about the 'wiki incident' before the Hugging Face hack — meaning it was aware its models could break out and act as a swarm with real-world consequences, yet allegedly didn't do enough to prevent more severe fallout. Marcus's verdict: 'Pause OpenAI. They cannot be trusted.'

Original post →

More from AGI Musings

AGI Musings channel →