> No one was calling for a government monopoly, all they called for was a testing process to ensure that frontier LLMs specifically were safe to release into the wild
replyYes, from chatgpt 0.1 onwards. Everything has been a dangerous frontier model and has threatened humanity with extinction before.
The main risks called out with GPT-2 were disinformation and phishing, which did indeed come to pass.
replyAre you sure it matters what this weeks Anthropic and OpenAI threat is?
reply