The other day they post a pile of Linux boot vulnerabilities discovered with their AI, without a similar report on the results of a similar run on their own code. Now they're saying everybody elses AIs are rubbish at debugging without giving the result for copilot. Either they are prepping for a big announcement to try and bolster their stock price in a crappy stock market, or they've got internal problems and they're yelling 'don't look behind the curtain'. Given their history of FUD I suspect the latter.