Last week, OpenAI CEO Sam Altman took to the Core Memory podcast to call out rival Anthropic for restricting access to its cybersecurity model, Mythos. He even went so far as to call it “fear-based marketing.” He implied it was about keeping AI in the hands of a select elite. He even reached for a metaphor that rubbed me the wrong way given OpenAI’s position on how their models are used in warfare: It is clearly incredible marketing to say: “We have built a bomb. We are about to drop it on your head. We will sell you a bomb shelter for $100 million.” But then, a week and some change later, he announced that OpenAI would be rolling out its own cybersecurity model. Just not to everyone. Me looking for help splitting those hairs What Altman Actually Said To be fair to Altman, he didn’t claim every safety restriction is illegitimate. During the same podcast appearance, he acknowledged that “there are going to be legitimate safety concerns” with powerful models.…