According to TechCrunch, Sam Altman and Greg Brockman stated at the DealBook Summit that humans should maintain continuous control over the future, opposing the gradual handover of decision-making power to AI. Brockman noted that as model capabilities improve, some computations no longer require writing out a full thought process, and relying solely on chain-of-thought monitoring may face limitations, thus necessitating new monitoring techniques. He also mentioned that following a previous security incident, the company reallocated 25% of its production engineers from their original development work to focus on technical safety hardening. The above statements reflect the views of the interviewees and the company's measures, and do not imply that the control problem for advanced models has been solved.