They do not have thumbs.

The fact that all you can do is reply to a typo is exactly the issue, there’s nothing of substance here

Your example falls flat on two counts

Llm’s that only have one glyph per token can easily count this, and yeah we train in bursts because it’s computationally cheaper, this hardly means anything important.

So you’re going by the vibes then? Why even bother with your comment? What basis in neuroscience is this founded in?

Neurons are in fact scochastic get over it.

Can you prove that humans don’t?

When the next fairphone with dp alt mode comes out I’m going to be all over it, then I may never care again.

Actually if it was a truly free market it’d be worse, even fewer protections on privacy and wages!

The point is a complete and utter removal of social mobility

White power is a political phrase used by neo-nazis, yes

Please present the best, most clear evidence from that list, as well as why you think that proves it

Why do you think socialism relies on people being that?

People say this but I have never seen a coherent answer to when or why things fall apart because of this.

How could they disprove or discredit it?

https://en.wikipedia.org/wiki/Burden_of_proof_(philosophy)

the burden of proof lies with the one who speaks, not the one who denies

Also not with this claim? It’s impossible to disprove this. You can always say something for how it can be done. Science uses falsifiability but you’re doing it wrong.

You tried to disprove their existence rather than prove their existence? That’s not how anything works?

Gram with a modifed helix

https://www.phoronix.com/news/Linux-Mint-Wayland-Screensaver

It was never impossible or difficult it’s just nobody bothered because screensavers are actually bad.

Not all implementations are slower, only certain ones.

It does not lack screensavers and there is no reason it has to be more demanding. The protocol isn’t more demanding, the implementations are

Actually cetaceans should be taxonomically the only true fish

Do humans do better? By what objective metric? How many of these have humans even made in history?

Also you don’t really need a human to verify these outputs, they just tried them. How many of these have humans even made? Also for the first of its kind with no previous rlhf scheme a 5% success rate is actually insane, this rate has historically dramatically improved with successive versions.

The stochastic parrot paper was nonsense, neurons are entirely stochastic. Stochastic models can be made very accurate

“Too stochastic” is nonsense. Neurons are stochastic. Also they made 16 from scratch.