juthen

It's a university research, not in-company. The tech is published.

I would want to see what the enemy is thinking, what is his question he got right Also some other minor things: - Maybe health should not go to negative - How long to charge my move? - Maybe there could be a timer for a special attack?

They temp ban by IP as fast or faster as any other google service.

The length went from 13mins to 37mins (!) with small prompts.

Almost all the AI models come from open research. Yes, open software hasn't caught up because they don't have the billions of research money to implement this models, but that doesn't mean it won't happen at some point.

He's not written books on the subject. You can setup a search yourself. There's also a blog with easier access than transcripts. https://www.robterhorst.com/post/apple-watch-ultra-2-vs-seri...

Speech. The speech-to-text pipeline is inherent in us. The convertion model relies on our education and cultural factors. The models can transcribe speech and do this conversion for new data generation. Have 10 mics at a public square and you'll have an infinite dataset (not a very smart one, necessarily...).

Maintenance and long-term development will still require grounding to reality. A super-CEO might build the app themself, but keeping it running 5, 10, 20 years is a completely different deal. I imagine developers will eventually start to act more like librarians, knowing their system very well but not necessarily in charge of making new content.

GANs are barely ten years old and already they have reached the classical ML algorithm status.