I agree. Anthropic and Mistral are good examples. Both companies have key people from OpenAI and they fairly quickly developed good models, but I don’t think either are thinking too hard about real AGI, but instead are trying to create useful and ethical tools.