I have a feeling that in order to achieve using deep learning (because let’s face it; it’s not AI) as a programming assistant, AI companies didn’t need to do half the shit they’ve done.

The resources being outright burnt on their endeavours haven’t served any useful purpose. You don’t need data centres covering the globe and to destroy thousands of rare books to assist with code, or find cancer in a scan.

Well, so far Linus is the only person in the world, it seems, that is using AI as a tool instead of a crutch. So I agree with that sentiment.

And what’s thw relation to the OP, or the comment you’re replying to? Are you understanding the result of what AI is giving you? Because thw comment states that this is not vibing because Linus owns the change, understands it.

And the bitterness?

I don’t support what the big corpos are doing, but you’re straight up denying that a hammer is a tool.

I’d agree with you if you mean that what has been marketed as AI to us is inherently bad. That said the word AI has been misused so much that is basically has no meaning. I’ve known researchers who have used things that can be called AI for really important research and it could run on a high end computer from 10 years ago. The externalities of that are quite minimal to what we gain. We really need to find language to discuss AI where when we talk about Big AI we’re not also dragging the proven, and efficient models which are nearly AI but not quite into the discussion.

I don’t think it’s inherently bad. We don’t need to use double the world’s energy to power it (as we see with open weight models) . We don’t need to centralize it in trillion aire companies. We don’t need to leverage an entire global economy on it. We don’t need to use it to replace human labor without a proper economic model.

Those are all aspects of greed and stupidity. None of that is inherent.

It really depends on which externalities you’re talking about

The only one I can think of as ‘inherent’ is the ethical concern over copyright. The others are more broadly related to the commercialization of the tool than to the tool itself.

I could say the same about all marketing, but that won’t make it disappear

But we very much do enforce times on when not to use calcs. Thars the whole point here.

We’re not talking about the average clueless vibe coder here. Programming involves a lot more than just writing code. What Torvalds did was essentially “AI assisted coding”. As i understand it, he used it to narrow down the position of the bug in the code and then fixed it himself with a one-liner because apparently AI couldn’t find a solution on its own:

And this was a debug session from hell, enormously helped by an AI doing much of the grunt-work. I’d like to call it my tireless helper, but the AI several times stated flat out that this was impossible and unsolvable and that we should just write a report about it. I suspect those things have been trained by people who may not be quite as stubborn as I am. But while the AI was ready to give up several times, it did keep adding debug code and analyzing it faithfully when I pushed. So credit where credit is due and I let the AI write the commit message above. This is basically a one-liner fixing a bogus “round_up()” to a “round_down()”, but there were 24 patches adding more and more debug information to this, and 18 kernel boot to finally narrow it down to this. - Linus

https://github.com/torvalds/linux/commit/818bebeb63dd6bf5f4e07e145f6cdbace520a34c>

That’s basically how AI should be used. To speed up things that are technically easy to do, but still very time consuming.

This is a very old discussion and argument, much older than the current AI push. You’re right to a degree but also that hasn’t stopped the rise of strong minds for software development. There will always be the curious and the lazy. The curious will always want to know how things work from the chain of events of a button click to the pixel shifts and logic gate flips to display code on a screen as well as run it. There will also always be the lazy unencumbered by such a desire for awareness who just want the magic box to do the thing.

How many software devs can write in assembly? Not many. Not much has changed if anything beyond the ability of the unaware to see their will come into creation. People couldn’t make realistic images and now they snap selfies with no barriers to entry. Does that make them a photographer or an artist? Not really.

Could this not be said for IDEs with code awareness such as M$’s intellisense, among other tools that accelerate development speed?

What you are describing isn’t an AI problem, is people taking a two week bootcamp and calling themselves a software engineer.

There’s a ton of app builders out there that are basically some crypto scam wrapped around Claude (or whatever). There’s a lot of idiots pumping out code slop.

I think there is a middle ground where you can absolutely build apps faster with the help of ai and still get maintainable code.

Just don’t let AI write a whole app in one shot. Or even whole features.

It 100% can save cognitive load if used correctly. Knowing the limitations and restraints of your tools is true for every profession.

If you know what your llm can and can not do reliably you reduce the over head of verification to basically zero. While if you use it poorly you basically have to double your work load to fix things.

Code is probably one of the easiest things to verify, for that reason I struggle to see any other viable uses for LLMs tbh.

user: write code that correctly asserts what 1+1 equals.
LLM: assert_eq!(1 + 1, 4);
LLM: hmmm, that crashed.
LLM: assert_eq!(1 + 1, 3);
LLM: hmmm, that also crashed, I must be doing something wrong, let me try one more time.
LLM: assert_eq!(1 + 1, 2);
LLM: that worked!
LLM Response: assert_eq!(1 + 1, 2);

user: what does this sentence mean? "Buffalo buffalo Buffalo buffalo buffalo buffalo Buffalo buffalo"
LLM Response: This doesn't mean anything, or someone really likes buffalo's

I can always run the code to check, I still don’t know how that sentence makes sense.

Local opposition to data centre builds seems to be most effective at the moment. People oppose it from across the political spectrum, and it’s becoming a hot topic. Add your voice to the crowd.

I have no good answer to that. My personal choice is ruin/kill any bastard who orders/approves building of data centers in their current state (read: doing damage to everything immediately around them and well beyond that), but I am physically far from anyone in that group

It doesn’t benefit everyone if it pushes the quality down

Look I know some people will chime in and say things were better when programmers had to do their own memory management but I can’t say I agree. Good tools don’t push quality down. People using tools incorrectly certainly does and it’s a guarantee some people will use new tools incorrectly.

Having a tool that can chew through a lot of code and give you suggestions which can be verified is indisputably valuable. Is it worth all the datacenters? Hell no. But open source people aren’t driving that.

The documents I quote are from publications by the U.S. Copyright Office. But, the Berne convention and the WTO ensure that member states copyrights are aligned.

I doubt “allow for AI generated code” has the bipartisan support you think it does; it certainly doesn’t have the popular support. But, ever if it were legal, it should still be unethical (e.g. Amnesty International’s reports of human rights violations in generative AI), and I’d be attempting to avoid it.

His story was a bit less flattering. The narrative the advocates want is that Torvalds is all in on vibe coding the kernel.

The actual story is that he found it somewhat useful, even as it gave him wrong information and directed him to get help from a developer. A useful, but limited augmentation.

If Torvald’s account were the accepted reality, they wouldn’t be steamrolling rural locals to make so many datacenters and wouldn’t be demanding we need more natural gas power, and wouldn’t represent the majority of our economy. They would have utility and be available, but not so obnoxious. Well, maybe, they might not have bothered with so much expense in training for this level of capability, but the cost is sunk now.

Of course most development isn’t Torvald’s level, it is human slop. So there’s more to things than the actually complicated work that AI is limited for.

But it’s not only critics that oversimplify. Advocates are obnoxiously bad at declaring everything is all good with AI

No argument here. I know Grammer, but there is how we speek and buisnees specific grammar. The level of formalities and what not.