Yes. There's a risk Anthropic AIs won't be able to hack or kill people. There's no risk that OpenAI AIs won't.