It's a bit refreshing to read updates from an AI lab without bullshit doomerism and overwhelming sense of self-importance from shit like OpenAI and Claude.

Also, DS is an excellent model.