Andy Halliday

I'm charging it up to see if I ever find the time to try that out.

But I think they're basically abandoning this particular form factor and going on to the cyber deck, which would give you a bigger screen.

And I guess it's a folding keyboard.

I'm going to knock off some things here quickly that I've been floating in my list of things to update our audience about.

One of them is that we talked months ago about The Thinking Game, the documentary about Demis Hassabis and DeepMind.

That's the Google division and the one that has accomplished tremendous advances in reasoning.

language models, and they used and developed, really, the whole concept of reinforcement learning and put that into practice with AlphaGo, AlphaFold, et cetera.

So those are models that learn by self-reinforcing, not getting reinforcement from human feedback.

Okay.

Okay, so one of the players in that movie, Thinking Mind, who's very notable if you watch that documentary, which I totally recommend you do, is David Silver, who is kind of the right-hand person to Demis in DeepMind as they went through this progression, developing reinforcement learning.

Well, David Silver has decided to leave DeepMind

to pursue superintelligence.

And so he's going to be funded, I'm sure, dramatically, and probably with funding from Demis and others at Google and so on.

And AI, he says, needs to get beyond just having human knowledge.

It'll need to discard that and be able to learn

from scratch, working from first principles.

And so this is an advancement of the ideas behind reinforcement learning, like the model itself, figuring out what the right objective functions are and how to reinforce those.

And he's going to do this with a company that he's calling Ineffable Intelligence.

which will build an endlessly learning superintelligence that self-discovers the foundations of all knowledge.

This is related in my mind.