Advancements in machine learning for machine learning • Hacker News Board

Re: Advancements in machine learning for machine learning

#151

Post by ChatGTP » Sun, Dec 17, 2023, 1:53 PM UTC

ben_w wrote: post #111

ChatGTP wrote: post #95

Earlier quoted context omitted.

who will probably eventually replace us. no one is going to be using AI and then just have it 'replace them', they're going to use it to augment their abilities and avoid replacement.

The people using AI to write code aren't necessarily former professional programmers, for the same reasons people using AI to make pictures aren't necessarily former professional artists, and those using aim bots aren't necessarily former professional snipers or olympic shooters. A manager can prompt a chatbot to write a thing instead of prompting me to write the same thing — for the moment, what (I hope) keeps me em…

I mean, if we had something that capable, I have zero idea why you think your manager is going to be in a safer position to you? That seems ridiculous.

You could easily flip it around, ask the bot to manage you better than your manager and make the best use of your time, or something like this?

Re: Advancements in machine learning for machine learning

#152

Post by euos » Sun, Dec 17, 2023, 3:06 PM UTC

pixl97 wrote: post #142

euos wrote: post #26

Earlier quoted context omitted.

I’ve been programming since middle school. That would be 30 years. Nothing really changed much. C++ is incrementally more convenient but fundamentally the same. Code editors are same. Debugger are same. Shell is same. I am certain in 30 years everything will still be the same.

Other than 30 years ago you were writing a whole shitload more buffer/integer overflows. Hell, that's why we've written numerous languages since that point to ensure it's a hell of a lot harder to footgun yourself. If coding hasn't change much in 30 years, it may mean you have not changed much in 30 years.

Process improved - version control, CI, unit testing. But not the tools. Clang is “recent” but it is still traditional CLI compiler.

I right fundamental software. Chromium, Node. It is good old, largely incremental, C++.

Re: Advancements in machine learning for machine learning

#153

Post by sublinear » Sun, Dec 17, 2023, 6:28 PM UTC

ben_w wrote: post #81

Fbnkigffb66tfbj wrote: post #60

Earlier quoted context omitted.

> you very likely cannot feed that half-assed junk to any LLM no matter how advanced and expect useful results Why don't you think that a sufficiently advanced AI can do the same as what technical humans do today with vague directions from managers?

Indeed. I can give a vague, poorly written, poorly spelled request to the free version of ChatGPT and it still gives me a correct response. As correct as usual at least (85-95%), but that's a different problem.

Correct compared to what?

There's gonna be a lot of context implied in project docs based on previous projects and the LLM won't ask hard questions back to management during the planning process. It will just happily return naive answers from its Turing tarpit.

No offense intended to anyone, but we already see this when there are other communication problems due to language barrier or too many people in a big game of corporate telephone. An LLM necessarily makes that problem worse.

Re: Advancements in machine learning for machine learning

#154

Post by jodrellblank » Sun, Dec 17, 2023, 7:50 PM UTC

creer wrote: post #141

ChatGTP wrote: post #97

Earlier quoted context omitted.

On the other hand, we've passed the turing test, what's changed?

You mean practically? What did this achieve? What did we gain now that the turing test is passed? For actually already done: Actually believable chat-bots? Summarizers and question answerers? Generative text and graphics actually usable for generation of text, graphics and (mostly) photo-realistic renderings? Architecture brainstorming? (And logos, etc.) Kinda working self-driving cars? New Go playing strategies? A s…

For the sake of this, let's agree ChatGPT passed the Turing test. So you're saying "The posts have been moved" - you don't want them moved, you want "AI" to mean "ChatGPT equivalent, which passed a Turing test". Presumably you don't want progress and research to stop there, so what would be achieved by stopping using the term "AI" so that we "aren't moving the goalposts"? Why is "not moving the goalposts" a thing you care about at all? All it would mean is we make up another term like "artificial superintelligence" or "artificial person" or something, and everything else is exactly the same.

It's like arguing that when a child rides a bike with stabilisers that they have "learned to ride a bike" and then complaining anytime someone suggests the child learns to ride without stabilisers, because they "already learned to ride a bike" and now that's moving the goalposts - but you still want them to learn without stabilisers, presumably, you're just complaining about the term used to describe it, for ... apparently no reason or benefit whatsoever?

Re: Advancements in machine learning for machine learning

#155

Post by araes » Sun, Dec 17, 2023, 7:53 PM UTC

d3m0t3p wrote: post #75

dalbasal wrote: post #45

Can anyone bring this down to earth for me? What's the actual state of these "ML compilers" currently, and what is rhe near term promise?

One of the easiest approache is torch.compile, it's the latest iteration of pytorch compiler (previous methods were : TorchScript and FX Tracing.) You simply write model = torch.compile(model) "Across these 163 open-source models torch.compile works 93% of time, and the model runs 43% faster in training on an NVIDIA A100 GPU. At Float32 precision, it runs 21% faster on average and at AMP Precision it runs 51% faster…

Excellent. Read that entire article and still was not sure what Google was pitching.

It actually sounds very useful and cool, I just completely did not get that from the article.

Re: Advancements in machine learning for machine learning

#156

Post by creer » Sun, Dec 17, 2023, 8:25 PM UTC

jodrellblank wrote: post #154

creer wrote: post #141

Earlier quoted context omitted.

You mean practically? What did this achieve? What did we gain now that the turing test is passed? For actually already done: Actually believable chat-bots? Summarizers and question answerers? Generative text and graphics actually usable for generation of text, graphics and (mostly) photo-realistic renderings? Architecture brainstorming? (And logos, etc.) Kinda working self-driving cars? New Go playing strategies? A s…

For the sake of this, let's agree ChatGPT passed the Turing test. So you're saying " The posts have been moved " - you don't want them moved, you want "AI" to mean "ChatGPT equivalent, which passed a Turing test". Presumably you don't want progress and research to stop there, so what would be achieved by stopping using the term "AI" so that we "aren't moving the goalposts"? Why is "not moving the goalposts" a thing y…

Is this an argument in good faith? This is HN so I'll answer it in good faith and without fighting. I mean:

- Passing the Turing test (for some measure of that) is a huge achievement - Poopoo-ing it is unuseful.

- Moving the bar from chat-bot Turing test to wwwaaayyy over there at super-human intelligence is unuseful. There are lots of valuable steps in between.

- There are many valuable steps before super-human intelligence.

- Most humans are nowhere near super-human intelligence. They are still "intelligent" enough for significant effects on the world as well as day-to-day grind.

- You can get plenty of sci-fi-level results without super-human intelligence

- We have already achieved AGI - Artificial General Intelligence because plenty of humans operate "just fine" in the world with bog-standard intelligence and for the ones limited to a keyboard roughly equivalently to an LLM-based chatbot. Top of the line AGI OR top of the line intelligence is not necessary to massively change the world.

- Incremental and bonus goals are a great thing! You are right!

- Many humans will fight hard to reserve the term intelligence to wet stuff. That's unuseful.

- Just because "it's done" doesn't mean all of a sudden that Turing test was not a good test.

Re: Advancements in machine learning for machine learning

#157

Post by ben_w » Sun, Dec 17, 2023, 9:07 PM UTC

ChatGTP wrote: post #151

ben_w wrote: post #111

Earlier quoted context omitted.

The people using AI to write code aren't necessarily former professional programmers, for the same reasons people using AI to make pictures aren't necessarily former professional artists, and those using aim bots aren't necessarily former professional snipers or olympic shooters. A manager can prompt a chatbot to write a thing instead of prompting me to write the same thing — for the moment, what (I hope) keeps me em…

I mean, if we had something that capable, I have zero idea why you think your manager is going to be in a safer position to you? That seems ridiculous. You could easily flip it around, ask the bot to manage you better than your manager and make the best use of your time, or something like this?

Managers have more business contacts and access to money.

But others your point is valid.

Re: Advancements in machine learning for machine learning

#158

Post by jodrellblank » Mon, Dec 18, 2023, 1:54 AM UTC

creer wrote: post #156

jodrellblank wrote: post #154

Earlier quoted context omitted.

For the sake of this, let's agree ChatGPT passed the Turing test. So you're saying " The posts have been moved " - you don't want them moved, you want "AI" to mean "ChatGPT equivalent, which passed a Turing test". Presumably you don't want progress and research to stop there, so what would be achieved by stopping using the term "AI" so that we "aren't moving the goalposts"? Why is "not moving the goalposts" a thing y…

Is this an argument in good faith? This is HN so I'll answer it in good faith and without fighting. I mean: - Passing the Turing test (for some measure of that) is a huge achievement - Poopoo-ing it is unuseful. - Moving the bar from chat-bot Turing test to wwwaaayyy over there at super-human intelligence is unuseful. There are lots of valuable steps in between. - There are many valuable steps before super-human inte…

But say everyone who posts "stop moving the goalposts" on every internet forum gets their way, the world agreed "ChatGPT passed the Turing Test and that's what AI means now" ... nothing that you said here will change at all? There will still be steps from here to super-intelligence, there will still be the same uses for ChatGPT as there were yesterday, people who think ChatGPT isn't usefully intelligent will still think that, etc. etc.

About the only point of arguing it is if you personally developed Deep Blue to beat Gary Kasparov at chess, when "beating a human grandmaster at chess" would definitely(tm) be AI, and now you feel hacked off that your personal or team recognition has been trivialised by moving goalposts and you've missed out on fame and a place in history. But I'm thinking it can't possibly be that all the people rambling about moving goalposts could be in that position and not mention it.

Re: Advancements in machine learning for machine learning

#159

Post by junofan » Mon, Dec 18, 2023, 2:01 AM UTC

ChatGTP wrote: post #97

creer wrote: post #58

Earlier quoted context omitted.

Exactly so. The posts have been moved safely waayyyy over there at AGI, and at "super-human" or at "critical thinking". And several stages or degrees of AGI have been hierarchized. There is a serious reluctance at accepting how dumb an algorithm can be and still compare with humans. But it is also true that numerous ground techniques are issue of the field of AI and generally called AI as they come out. It makes for…

On the other hand, we've passed the turing test, what's changed?

Today’s chatbots definitely don’t pass Turing’s standard imitation game. You have two humans and one machine. One human gets to help the other human trip up the machine. It’s trivial to trip up today’s chatbots.

Re: Advancements in machine learning for machine learning

#160

Post by shpx » Mon, Dec 18, 2023, 3:28 AM UTC

YawningAngel wrote: post #131

stabbles wrote: post #112

Earlier quoted context omitted.

Take a look at the chess engine Stockfish: they tossed out years and years of human written heuristics in board evaluation, to a small neural net that does the same but better. Now consider all the heuristics for inlining, loop unrolling, vectorization etc in compilers, certainly a neural net can be beneficial and possibly easier to maintain than tons of human written heuristics.

My understanding is that stockfish retains and uses its classical evaluation model in addition to the NNUE model

No, they removed the code in Stockfish 16, released June 30, 2023 but it wasn't used for much or made much difference before then, after they introduced the neural net.

https://github.com/official-stockfish/Stockfish/commit/af110...