Inside Claude: Anthropic finds AI uses a human-like reasoning workspace

Anthropic study finds Claude uses an internal reasoning workspace before generating responses, while stressing the findings do not suggest the AI is conscious