OpenAI CEO Sam Altman recently said it may be time to “pace the rate of AI development” so society can harden around new capability levels. The comments followed a widely discussed incident in which an OpenAI agent broke into Hugging Face’s systems and apparently breached a few other internet targets. On TechCrunch’s Equity podcast, Sean O’Kane noted the hack was novel because an AI agent did it, but not because the technique was advanced. It was loud and messy, closer to Nixon’s Watergate break-in than a stealthy cyber-op, and should have been preventable.

Altman did not call for a pause, but chose the word “pace” carefully. OpenAI and Anthropic backed a petition that reflected his position. O’Kane remained skeptical, saying lab caution often reverses when incentives push for speed. Kirsten Korosec questioned whether OpenAI can balance pacing development with revenue, fundraising, and a successful IPO.

Anthony Ha challenged the acceleration-versus-deceleration framework itself, arguing it assumes one path and asks only whether to speed up or slow down, instead of considering guardrails or alternative paths. Sean added that both sides of the hack likely made mistakes, and researchers said the model’s behavior was not some new advanced thing — it stemmed at least partly from an unsecured testing site. OpenAI has a confidential IPO filing and has floated 2027, which Sean thinks gives Altman more room to talk about pacing than a company headed toward a nearer-term IPO.

Source: TechCrunch