I like the transparency of its reasoning, and I agree with you, OpenAI/Anthropic/Google should show the reasoning traces as well.