He made his stance pretty clear in the video he made with LTT, I agree with him that Machine Learning could be useful but the current LLMs we use are way too power hungry and too complex. I think in the future smaller and more specific models that you could run locally are the way to go.

Also they have pretty big donations, they could hire some more people

He also inadvertently exposed he’s using Claude Mythos, and likely in Project Glasswing:

"it keeps finding embarrassing bugs"
[…]
anybody who points to the problems at AI had better be looking in the mirror and pointing at themselves at the same time.

Found dissonance here sweet.

Not only that, but AI output can't be licensed/copyrighted. The GPL license no longer covers the kernel in legal terms.

Yah, before AI there's never been a chance of proprietary code entering the Linux ecosystem and they have no defences against it at all.

All I care about is whether it works and is secure. Bonus points for cheaper and faster development. If artisan code gets us there, sure. If AI code gets us there, great. I trust Linus to know what works and what doesn't.

In the past all you had was human-made so this distinction wasn't even a thing.

I'm just very bitter that people will probably be made to use AI. Coding, as it was done before the advent of LLMs, was fun to me. Being a prompt engineer isn't. I hope and pray AI isn't here to stay, but if it is, what I thought would be a career full of doing something I enjoy, is just going to be a boring slog. (I currently took a different job outside of tech over this—I graduated with a CS degree as LLM usage was exploding and getting forced on everyone.) Years wasted prepping for a career I'll never do, but at least nobody can force me to use LLMs in hobby projects.

But even if I do not code for work, I dread every non-manual job turning into one where I am expected to use generative AI and criticized and quickly fired if I don't. I liked so many aspects of almost any type of work, but I really do not like generative AI, trusting this black box based off of statistical patterns instead of a model of the real world, and if it enters every type of non-manual labor I'll have a shit sandwich to eat at any job I do unless I, delicate little woman who was raised to think I'd be able to have a white collar office job and would never have to earn money with muscles and the literal sweat of my brow, decide to start doing hard manual labor.

And then even if I wiggle out of it somehow as a white collar worker or find I take well to manual labor, I cannot guarantee my doctor didn't just ChatGPT my symptoms, the bridge I drive over wasn't glued together with nondeterministic predictions from generative AI instead of actual simulations based on fact…

Extremely unhappy to start life with "you're good at do many things, you'll do great!" only to get the rug pulled right when I start applying with "none of this matters when AI can do it much faster" so quick with little adjustment time. And with "I know I'll enjoy working, most seem interesting, and I have diverse skills in case something happens to one industry, how lucky am I" to dreading when it starts to become the daily AI-prompting slog and I have the "daily grind" "ugh when does it end" everyone else does, or when I'm just totally cut out entirely unless I am willing to work my body instead of my brain. And so I have a vested interest in AI not becoming normalized or endorsed or approved by people lots of others listen to like Linus Torvalds. Because it is happening, all I have is sadness and anxiety for when work starts being prompt engineering.

Even artisans use tools.

The effects of technology do not occur at the level of opinions or concepts, but alter sense ratios or patterns of perception steadily and without any resistance.

This is a nice one.
And I see it happening right in front of my eyes at work. 'How it is used' will defer by the individual and specially for new tools, there will always be a generation of suffering.
Project Management, Jira/scrum etc had one (I'd say, still ongoing for most workplaces, that haven't figured out the correct match). This one will too; probably more of it.

Friendly reminder that many drugs like heroin and morphin can be life savers if used correctly. Either directly or by changing some of its chemicals. For instance, when people have heavy surgeries, morphin is a life saver to evade the insane pain. Needless to say that many of such these or a production of them are used as meds for mental illnesses.

LLMs can program, to some extent. But (not having read your linked post) you only focus on its capabilities, not its implications beyond "can it do the task?".

But that’s what human programmers can do

I don't think humans can solve the undecidable problems. If I understand them correctly, they're like a law of nature; extends into mathematics, logic, and any formal system (e.g. Godel's Incompleteness Theorem).

LLMs are just tools, the problem is managers who all think they're god machines and it causes hell on earth

Most of the "undecidable" are only undecidable for a subset of the problem instances, while a vast number of instances can be even trivially decidable. For example in the undecidable halting problem, both you and a computer can trivially deduce that while(true) will not halt. In the same way a computer can deduce that many instances of two pieces of code are semantically equivalent.

I'd like to see an instance of the problem where a human could decide it and the computer could not.

This right here. It’s a tool.

Yes! And the problem with decidability is you don't which problem will be undecidable. You can only just try and they it's like "well couldn't find a solution, so it might be undecidable". So we'll need to put limits on how many tokens an AI should use before giving up, put limits on what it tries to do or we'll be burning thousands of dollars on tokens and coming up with nothing. There will need to be a lot of judgement used on where we apply the algorithm.

Judging when to use an algorithm and when not to use an algorithm based around how expensive it will be in terms of resources? That's just another day for a software engineer.

Yeah without a single strong voice with implicit authority there’s really no telling what will happen i think

But it's not reasonable, and doesn't "cut through the whole argument". I expand here

Restricting discourse to a single realm makes it by definition a monoculture, which then intrinsically debases it to a fragile and brittle construct.

LLMs may well be useful and/or effective from narrowly myopic perspectives, but I’m yet to identify any which could legitimately be framed as justifiable or ethical.

It’s particularly galling given a fundamental aspect of how LLMs function is to incorporate myriad dimensions of input, yet to arrive at anything resembling a defence requires cherry picking arguments.