Not really. They’re a couple of months behind OpenAI and Anthropic, although arguably ahead of the Chinese labs. And these attacks seem to require leading edge

Not really. They’re a couple of months behind OpenAI and Anthropic, although arguably ahead of the Chinese labs. And these attacks seem to require leading edge models. I worry that when and if Grok gets there, we’ll find out that SpaceXAI is too casual about security, though.