At a conference last week, Anthropic co-founder Jack Clark said that LLMs are approaching a technical depth that most non-experts cannot evaluate. As if to immediately prove him correct, his competitor OpenAI released a model specialized for complex math and programming problems, prompting a dozen people to ask me for review requests. https://openai.com/index/learning-to-reason-with-llms/ Newscard GPT o1 is the long-awaited strawberry OpenAI once again takes a substantial research lead o1 mak...
