The assumption that it is an arms race is a shockingly anthropocentric view of something that's supposed to be 'intelligence' but is just a distillation of collected HUMAN opinion.
Not only that, it's a blindered take on what human opinion is. Humans are killer apes AND cooperative, practically eusocial apes. Failing to understand both horns of that dilemma is a serious mistake.
All technology is an arms race. People are hung up on OpenAI, OpenAI is just one of hundreds of AI companies. Military AI in drones is already to point where AI can fly an F-16 and beat humans.
What's the difference for the world between the Altman and the Sutskever approach for OpenAI? With Altman the bad stuff happens at OpenAI and everyone gets it at the same time. With Sutskever, the bad stuff happens two years later but it happens in random pockets all over the world and nobody can be quite sure what they're facing.
Not only that, it's a blindered take on what human opinion is. Humans are killer apes AND cooperative, practically eusocial apes. Failing to understand both horns of that dilemma is a serious mistake.