I think OpenAI has made some really bad decisions with the core tech even while making great decisions with the overall services, and from Altman's various comments over the past two years I was under the impression this was coming from him.
The only reason I'm skeptical of this is the abruptness of it all. Why it needed to happen with a 24h turnaround is bizarre, unless there was something like an internal meeting this week regarding GPT-5 where his doomerism was even more sending things off the rails and there was a reactionary immediate "we no longer have faith you can do this."