rimliu

not this kind of non-deterministic tool though.

This is not how LLMs work.

And they are dropping like flies.

Cannot wait for "they want you to pay for the gas to drive a car you already bought!"

The "not solved" part of the education is the same as "perpetuum mobile" is not solved - if you want to get kids educated without stress and putting in effort you are out of luck.

    > I would counter that education, at least in the USA, does not adequately
    > prepare one for life, the real world, and effectively finding employment.

Education should educate. Alas, parents outsources raising their kids to schools.

It is watching Linux boot sequence all over again.

There is a plenty of open hardware and software. Think a about it for a moment.

Think layer? Like a page loading 6MB of crap to show 0.5KB of text?

Intelligence is not knowing the answers to the questions, but knowing what questions to ask.

There are other free and not that buggy alternatives. Features mean nothing it cannot export the result right. It is not a proof of concept, it is a dancing bear - yes, it if fun to see a bear dance, but it dances terribly still. The money spent on tokens to make it useable would buy you lifetime Adobe subscription...

That's a popular take now, but is there any project of that kind that really got traction and use? (I am not talking about all-hype-no-substance like the Adobe one).

And the big players will run out of money. Win-win, I guess.

   > recreation of the Adobe suite.

Have you looked at the issues page on Github? Looooooooong way to any recreation, and I doubt anyone is even wanting to walk that way. It was all for hype, and hype it got.

Apple pays Samsung for making the things Apple designed.

That's an amazing phrase.

I have a formal training in education. I may agree that learning may not necessarily be hard, but any learning requires effort.

But thinking requires mental manipulation of the things that are in your memory. The less structures you have there, the lesser the ability to reason.

Did you, really?

There are none. Claims and the truth can be different things, you know? See the research on that - all the "reasoning" is is just back-rationalization.

so that's basically we don't know how to make the good hardware, let's force those who know give it to us.

you forgot "make no mistakes".

novel math problems?

There is no such thing as reasoning in models. Any "reasoning" is invented afterwards.

LLMs already hit a wall. Now it 80% of marketing hype and 20% of retooling and benchmaxing.

And guess who are pushing for the Chat Control.

maybe there is a reason they are called... system frameworks?

Like Concorde was getting us closer to the speed of light.

Should CEOs of Antrophic and OpenAI be reported too then? They cannot stop saying that AI will destroy humans.

none of them is smart.