
Yeah, likely a viscous liquid solidified slowly as it spread out. That can make little branches just like this picture.

Yeah, likely a viscous liquid solidified slowly as it spread out. That can make little branches just like this picture.

This is what I really hope comes out of the AI bubble. Totally dynamic characters in games could be a really exciting way to liven up a world, assuming they are performant enough.

A point in case why vibe coding is so very fucking dangerous.

Very clever design. Though the material engineering looks just about at the limits of what humans can do
C++ is kinky that way. You can consent to all manner of depraved programming patterns. Great for use in personal life, but maybe not appropriate for the office.

It’s a nice little game that reminds me how bad I am at kana
Maybe it was a mistake to use an entire generation of the world’s brightest minds to build better ad optimization.
We easily could have been to Mars already.
It seems clear that during training they weight Musk’s tweets (and those of accounts he likes) more heavily than everyone else’s.

Fully open is extremely good, it uses entirely open datasets and has all training and inference code. We need more things like that.
Did you try flipping the magnet over? If you put the two north ends together they’ll repel instead of attract.
China has a big problem with selling an identical product fifteen times through twelve different companies.
I think it’s an SEO strategy for Amazon, where they edge out any competition by being everywhere on the first page of search results. They also have the ability to game reviews by killing any products that get bad reviews and recreating them under a new brand.

There doesn’t seem to be much going on here. What’s the actual advantage of this technique? It seems they’re doing things more or less the same as everyone else.
I mean, the article talks about decoders as if they were some brand new innovation.

There’s almost nothing so hard that doing it in space doesn’t make it harder. Not sure what these guys are smoking.
He’s just trying to win some more bread for Europe. Eventually everything will be fully subsidized by fines on American companies.
I think we all know the answer is zero. No way Trump can run.

A neural network network?!

The human brain is a neural network that operates on less than 100 Watts of energy. So it’s obviously not physically impossible.
There’s no scaling law that invalidates general intelligence, only ones that invalidate the exact, unmodified techniques we use today. At any time someone could find a more efficient way to do things and turn the whole world upside down.

Ah that’s nothing in America. You should see some of the highways in Texas.
The government doesn’t want an AI that they can’t control, and grok is far too unhinged.