[Thought for 2 months] I have joined Google DeepMind.

Last night I taught nanochat d32 how to count 'r' in strawberry (or similar variations). I thought this would be a good/fun example of how to add capabilities to nanochat and I wrote up a full guide here: github.com/karpathy/nanoc… This is done via a new synthetic task

At the OpenAI Frontiers event in London this week, OpenAI demoed a new bidirectional speech model that was post-trained on translation data and can translate what you say into another language in real-time while you're still talking by waiting for complete verbs instead of going