I'm hesitant to follow this line of reasoning, it plays directly into the AI/LLM hype. "our model so dangerous, much powerful"
I call bullshit, some engineer(s) likely pointed these bots to hack a target, and gave it access to the tools to do so. what makes more sense, the biggest grifters on the planet pulling another grift, or a sci-fi scenario where the "AI breaches containment"?
also, fuck the NYT and Ezra Klein