Google DeepMind team dropped more wild AI Agent videos. 3 examples: 1. Identify famous face and facts from drawings

2. Identifying iconic landmarks from journal

3. Reasoning about Formula 1 car from drawing

The new chat-gpt-4o does it flawlessly aswell

I thought it said, "Do you need a hand with your drawing?" 🤣

"Is that what you were going for?" 😂

They never stop

They got an amazing team with really promising products

Check out Google DeepMind's new AI videos, they're fascinating!

So it's not available to public yet?

It is fascinating and indeed impressive to see how multimodal training has contributed to AI's "thinking" abilities - and yet I can't help but wonder whether we're enthralled by the novelty of seeing these demos. We need to see more than "what the model can do." Better use cases.