[0xDEAFBEAD]

But this incident undermines OpenAI's safety case compared with open models.

Furthermore how do you explain Sam's downplaying? https://xcancel.com/HumanHarlan/status/1965932275465597077#m