That's the part that kinda saddens me.

I mean at the risk of using an Ai-bro analogy, we still learn paper math even though we eventually use calculators, so we understand what the machine is doing, but...

... There's this nagging sense with LLMs like people will just stop coding themselves because it feels pointless, and we won't develop any new techniques or ways of thinking. There won't be any more ways for rookies to build experience.

I imagine the motivation to write apps and stuff starts to diminish when you feel like you're just doing it manually for its own sake, not because it's actually better. (Even if it would be better)

Like, there ya go, gobbled up all the organic data until ~2024 and then human-led effort and progress just...stops...? I can't imagine many programmers are thrilled to teach a machine how to outdo them.

I'm really curious what fresh CS grads are thinking right now.

Like the endgame goal is what, the LLMs get smart enough to just improve themselves and humans just...stop thinking and all become "idea guys"? All the coders just become middle managers over "agents"?

I think it's fun and novel to chat with my GPU, but the big picture implications of this disruption are... mostly depressing, especially watching all the jobs get worse and this tech not even being capable of freeing up any time in my domestic life.

It just makes us all feel under threat. All of the time. Like under the robber barons of the gilded age.

agreed, a tool that you have to actively fight to do something well is a garbage tool. it's like trying to cut some chicken meat with a knife that's magnetically drawn to living beings. You can make some nice diced chicken with it, but it is way more likely that you'll slip up and cut off your finger than with an actual knife.

regardless of the original intention, the purpose of generative "AI" is to become a manager that's out of touch with the actual creation process, so at best no one will know how something works, at worst no one will be able to fix it if it breaks.

Conquered is a funny way of saying stolen.

Coerced purchasing, or purchasing from someone else who stole, or purchasing under intentionally false understanding, is still stealing.

This argument is conflating the bad boss with the misattribution of the tool. You could do that for any tool throughout history. But I still fail to see how you go from "bad boss is bad because they make us misattribute tool" to "good boss is bad because he's not wallowing with me about how bad boss is bad".

As anyone who has known about him for more than a couple of years knows, Linus has a pretty solid track record of calling out stupidity. It's just virtually always stupidity that affects the kernel. If he had to stop and tell bosses at other companies how to operate, Linux never would have made it out of the 90s.

Furthermore - I suspect he doesn't even care if you understand it. He cares if it works reliably. If you accomplish the task of solving hard problems by taking shrooms and dancing naked on your keyboard, I don't see why he would care.

Does Linus support steaming artwork and a fascist surveillance state? Because using and supporting and entrenching the technology use is exactly that.

This is really more a case of possibility #2. Specifically, why are you rolling your own in the first place? The whole reason the output of the AI is "probably flawless", is because this operation already has a million implementations. What's the value in conjuring up another one and adding it to the code that needs to be maintained? Yeah, binary trees aren't terribly complicated, but neither is struct Vector3 { float x, y, z; }; or any of its operations, until you realise you'll be running those operations an awfully lot, so it would be better if it had the right alignment and SIMD instructions were used in its operations, and of course, if you have Vector3 you'll probably also need a Matrix4x3 and a Quaternion etc., so in the end it's just better to use an existing solution.

This is what I mean with turning good devs into mediocre ones: the worst solution becomes the easiest solution, so that's what people will lean towards.

Of course, there can also be situations like you have a very specific use-case and using something off-the-shelf would be sub-optimal, or you're working with old code and making their own was the decision the previous person or your dumbass past-self made, so that's just what you have to work with. But the AI is going to struggle more with those cases too.

Lastly, I also mentioned "energy". If you're trying to put in words things that aren't easily conveyed in English (or any other natural language) all the time, you burn through your cognitive endurance like crazy. This is why you don't want to be in a highly technical meeting in the morning, so being in one all day with a moron you can't really come to common understanding with (beyond the size of its context window) is even worse.

He doesn't get it... Have the big corpo-shilling libertarians rotted his brain?

Like, imagine if using "AI" consisted of whatever big corpo SaaS you subscribe to sending your prompt off to someone to work on then send the result back to you. Isn't that an idiotic dependency to have? And why the fuck even go through the middleman? Oh they have the "best workers?" Fuck that. Nvidia makes the best cards, but you don't see me buying from them.

Do people like Linus not realise how silly it is to offload your work to some middleman service that just steals work from people who do not consent to sharing their labour? Who are desperately throwing insane, GDP-level boatloads of cash at it this unknown technology in the hopes that they'll somehow get people hooked on their particular chatbot?

The intentions of these companies are not good. Their entire valuation is not based on being a "force multiplier" or whatever bullshit, but on being able to threaten or fire workers if they don't slot themselves into the newest enunciation of hype-based capital.

Anything you do to willingly boost or massage these corporate services is enabling actually vile companies to continue puffing air into their stupid pointless bubble until it bursts and the next generation has to eat tuna helper sans tuna just to be able to afford to live.

This comment disguising itself as nuanced by saying the word nuanced a bunch is really nuanced.

killing the previous owner has been a long established way of getting land for thousands of years

Yes that’s one of the mechanisms to steal the land

theft requires a victim, the dead do a bad job of filling that role

That is true but does not pertain to what I wrote.

To be more precise, LLMs' present valuation in the market is not based on their being a skill so much as the end of skill.

So I assume you agree that if someone murdered you, you would prefer they go free and live a happy life.

After all what would they be punished for. There is no victim.

You can know a hole is not straight. You can see it. Besides, if your house is built by someone that doesn't know how a drill works, well, it's going to crash down over you. Then why are we accepting works of people that can't and doesn't know how to properly use the tool?

Now, both of your examples have a person having to use the tool and know how to use it... IA is having the drill doing a hole by itself and just say "good enough". The house will crash down.

My point stand against both your examples. Both requires a human doing the work and the human to know how to use it to be useful. If the difference between a surgeon and a butcher is precision and knowledge, why are we allowing the butchers in on everything?

"IA" at best is a tool and at worst is a substitute. A "tool" that doesn't ask anything from you is not a tool.

I stand by it's not a tool. It's a substitute. But even by labeling as a tool, why should accept works from people that doesn't know how to use it and when? By following your examples we can't trust anyone claiming "I did this" with an "IA" until we or a professional inspect it, and since you can't inspect the absurd volume of things "IA" spits, the only solution left is ignore it all. Either you trust blindly on the guy that says he knows how to use the drill and risk your house falling down, or only accept works of a professional, and if you only take the job of professionals then what's the use for "IA" then?

Conversations on this subject will continue to stagnate and miss the important points as long as people continue to act like the discussion of ethics is the same as the discussion of efficacy. People who disagree with AI on a principle of its impacts being negative will often call it “a lying machine which is regularly wrong and offers no utility”.

May I suggest that people with criticisms of the ethics and impact of AI take a charitable position towards the utility offered by AI. Accept something along the lines of a positive account given by an experienced software engineer - it’s useful, it presents risks but also it presents new ways to get certain work done well in much less time etc.

Even if you do not believe this, the point is it is a much more ambiguous argument than the ethical argument or the concerns over the impact of AI on the species, environment or society. You weaken those arguments by putting them along side poorly justified, somewhat falsifiable claims that AI just doesn’t work.

Is this the Linus Tech Tips guy? Isn’t he like a known asshole these days or something? Or am I thinking of another Linus?

I mean, it would be better if you could write it in assembler.

if there's no punishment for murder in war why should there be punishment for killings in peace

Why not be transparent though?

LLM is like a compiler? Then please replace the C sources with English and let it compile!

this is Linus Torvalds, the guy who originally wrote Linux and is still in charge of kernel development

Proof: Rollercoaster tycoon

So why did he include that part? No particular reason? Was he just making an unrelated remark about some of his friends? Accidentally included that bit but it was intended for a different email?

The simplest explanation is "similar to all the other times he's made back handed remarks or direct statements criticizing the intellect of who he was speaking to, he was criticizing the intellect of who he was speaking to".

This is an odd hill to die on. Linus is famously callous and rude in the Linux mailing list so to declare that this time he was just making a completely off-handed and unrelated remark about his bois is a difficult position to defend.

All I'm saying is that determinism has a specific meaning in computer science. In an LLM, the rng value is determined by the seed. Given the same inputs, an LLM always produces the same output. It does not matter whether the user understands the code or operation at all. It is purely dependent on the program itself.

A non-deterministic program, is one where the output is determined by factors outside the code itself. This usually requires non-deterministic programming language constructs. For example, if you convert a set to a list in python, the order of the elements is non-deterministic. Out of the control of the program. Same inputs, different output.

Not really true.

LLM valuation in the market is aimed at several things, making unskilled people appear skilled is the smallest of them.

The most valuable position in the market right now is as a substitute of search enginesm

The second most valuable position in the market is as a tool to speed up work for already skilled people.

And I would argue that the third most valuable position in the market is a bot tool to substitute human iteration.

Unskilled people doing slop might be the most visible face of AI as it's very annoying but it's a minority of it's market, just a very loud minority.

For local AI image generation, often the input parameters are indeed recorded into the output image. You can then use the metadata of the image to get the parameters and re-generate the exact same image. This is the definition of determinism in computer science.

A lot of blatant and obvious reasons that really shouldn’t require an explanation.

Replaying the seed will only get you the same result using a particular engine and a particular random number generator

And the seed and rng are all part of the LLM. If you use the same inputs you get the same output. That's the definition of determinism.

Determinism does not care whether or not the user/programmer understands the program or not. The word you're probably looking for is "unpredictable" or "less predictable".

The most valuable position in the market right now is as a substitute of search enginesm

The second most valuable position in the market is as a tool to speed up work for already skilled people.

I perceive them as having equal value. In each case the technology is forced on users in a top-down way and results in worse outcomes.

I don't expect you'll agree; that's not the same as what I wrote being "Not really true."

This mindset is the entire problem. "Oh we don't need to worry about ethics or morals, just efficiency" is exactly the issue with every pro-AI argument and every pro-capatalism argument. Its the issue with our governments, with investing, with corporations.

I say fuck anyone who argues ethics is not important when talking about efficacy. Fuck that mindset and fuck AI.

I partially agree. If forced by stupid management results would clearly be worse.

Also people using LLM as search engines have a lot to do with internet entshitification. Surfing the web have become so obnoxious that asking an LLM the answer to something is easier.

This have two obvious conclusions from my point of view. First is that the best scenario would be a not entshitified internet. Second is that when LLMs, specially the big ones owned by big companies, get their dose of entshitification, users will be back on square one of not having a good way of extracting information out of this massive sea of information that we have built and which should be much easier to navigate.

I think the point is that it actually hurts an ethics-based anti-AI argument to also claim that it has no utility, because the no-utility argument is weak and therefore makes the rest of the stance (i.e., the ethics-based argument) look weak by association.

Exactly, if you say that AI is ethically bad and doesn't work, to someone who uses it and clearly sees it works well for them, it just invalidates your argument.

If you just stick to the ethics your argument can't be dismissed in that way.

I couldn't agree more. There is a long list of valid reasons to hate on AI and the industry but AI = slop barely makes it onto the list anymore.

Isn’t that an idiotic dependency to have?

You are correct, but also there are many of these big corpo SaaS's so it's not like if one goes away they would be stuck. Locally run AI models are also increasingly capable and I firmly believe that in 5 years open source local models will be the primary source of AI coding assistance either way.

You could say the same about any of the IDE's, commercial or not. Yes an IDE is a dependency that could away, but also there is plenty of alternatives.

Do people like Linus not realise how silly it is to offload your work to some middleman service that just steals work from people who do not consent to sharing their labour?

I don't see how it's "silly". It's unethical, nothing silly about it. I think that someone like Linus, who freely shares code he writes with all and thinks everyone should do so, has little care for the woes of proprietary codebases being slurpled up by the coding models. After all he dedicated his life to fighting commercial codebases.

Anything you do to willingly boost or massage these corporate services is enabling actually vile companies to continue puffing air into their stupid pointless bubble until it bursts and the next generation has to eat tuna helper sans tuna just to be able to afford to live.

Agree.

Piracy is good, and AI is piracy 😏

I think there's a broader question here from an evolutionary perspective. Humans are fundamentally tool users, and technology has been central to our development as a species. It's possible that this evolutionary niche we've occupied will eventually reach its limits and contribute to our own downfall, but I'm not convinced AI is uniquely responsible. It's part of a much longer trajectory of technological progress.

We've evolved to solve problems and use tools to make our lives easier. That doesn't mean technological progress is inevitable, but these tendencies, combined with competition, make it difficult to imagine us ever really stopping. Even if we stopped today, who's to say another generation of humans 10,000 years from now wouldn't continue down the same path?

I don't think this is exclusively about capitalism. Capitalism certainly incentivizes growth and reducing labor costs, but competition, greed, and the pursuit of technological advantages have existed under other economic systems too. That doesn't diminish the legitimate concerns about how AI is being used against workers.

I also wonder whether our short lifespans make us particularly bad at considering the long-term consequences of our innovations. Who is going to sacrifice the opportunities of their own generation to prevent a possible catastrophe centuries from now, especially when there's no guarantee it would stop technological progress?

I have a degree in psychology and particularly enjoyed evolutionary psychology, so I think it's worth asking whether the deeper problem is our own nature, and whether our capacity for collective restraint can keep pace with our ability to innovate.

I know this raises more questions than it answers, but we've faced civilization-changing technologies before, from the printing press and electricity to nuclear weapons. AI won't be the last. Perhaps the more important question is how we learn to manage these transformations when our tendency to create them seems so deeply rooted in who we are.

The other way of helping with that, is give Apple, Adobe and Windows a huge advantage, so people can't easily move to free / open software.

It's a lot more difficult to be controlled by a monopoly when the OS is free, and anyone can roll their own

At the moment, money is basically going straight to closed source US companies for almost every desktop computer. Remember when Tim Cook gave Trump a gift publicly the other day? We KNOW Microsoft and Apple DEFINITELY aren't on our side... They're not even subtle about it anymore (weirdly though, it feels like Microsoft is more on our side these days than Apple)

Yes! Another problem is, not everyone is using AI in the sense that people think of.

AI is a large field, it studies many different methodologies and aspects of automation, prediction, generation and observation, yet today most people outside the field use it for LLMs and image generation, which throws the baby out with the bath water.

In addition to that, there are publicly available LLMs, some have datasets obtained entirely from consenting parties and public domain works. You don't have to pay billionaire technocratic fascist a single dime to use such marvelous tools computer scientists worked decades to bring forth.

A machine can never create art, it can never take responsibility and should never be allowed to make decisions, it's a tool.

Everybody understands what you're saying. Yes, the source of randomness in a computer is deterministic (unless you have a physical source of entropy). Every computer program is deterministic.

Nothing about the operation of LLMs relys on that determinism or views it as useful. LLMs are probabilistic in nature and require randomness (emulated or real). When people are saying it's not deterministic they are stating that there are multiple outputs possible for a single input (excluding the seed to the RNG because it's not necessary to use a deterministic source of randomness).

That is fundamentally different to a compiler, which is the overall point being made.

This current iteration of AI would not be possible without big tech. If you actually thought about it for more than five minutes you would realize you cannot reconcile those two and still see LLMs as "just a tool". Sit on that one for a minute. Utilitarianism rots your brain.

I find myself asking the same 3 questions when considering whether or not to use an AI tool.

Is it a straightforward task where I can save significant time with a tool?

Can I verify the results?

If the tool AI makes a mistake, or if I am not thorough enough in verification, is the risk of failure low?

I'm a firm believer is using the right tool for the right job. Most of the AI's people are using are not designed for the things they are using them for, and I don't see many practical uses for generative LLMs. I will acknowledge that there are purpose built AIs to be used as a tool, but even then most of them tend to do a below average job. We're also seeing a lot of algorithmic driven models being labeled as AI, similar to calling something a "smart" device.

Is Google Maps also an addiction? I'm entirely certain it's reduced or even eliminated peoples ability to navigate to places by memory, eye or even the stars.

Fucking Socrates was against writing, because people would become lazy and not remember stuff and not discuss stuff and not listen, because they could just read it.

It is a fucking tool, because every person gets to decide how it's used. You see lazy idiots use AI without knowledge or skill, and then assume the problem is AI, instead of the people. These people copy pasted Stack Overflow before, they didn't change.

Why remove a tool from everyone, just because some idiots exist?

Ok, how the fuck is that relevant to the argument?? LLMs write code, then deterministically it gets compiled, type checked, linted, etc. At that point you absolutely have deterministic code that is as valid as can be verified. EXACTLY like if a human wrote it. Unless you're suggesting that programmers never made mistakes that don't show up in tests....

Maybe you're stupid? Because the argument seems pretty fucking clear.
Is a compiler deterministic towards the INTENDED functionality? I input instructions in the front and out comes something that produces the EXACT functionality I INTENDED? Because that's the relevant comparison here.

Ah yes, because even a single fucking percent of people that use an interpreter or compiler understand how they work, let alone can create one, or validate their output.

Have you heard of open weight models, or are you just willfully ignorant?

Which is very different from compilers, especially for large languages with existing open source code. Those you just change willy nilly, as you want.

But I rarely want the output twice? At least in this domain of coding, the entire point is doing it once.

Secondly, I can definitely validate it in most ways that matter. Ironically, by compiling it for example, I make 100% certain the output was well defined code. Type checking, linting, tests, etc. all help ensure I got functionally valid code.

At that point I might not really care if the function name or a variable name is different, just like I don't care that the compiler did some kind of heuristic optimization or minified the names or something.

In both cases, I have a target functionality, and I need to validate it was reached. The way I got there doesn't matter (in most cases).

It's a bit like Docker vs. Nix. Sure, docker isn't really reproducible perfectly, but it fulfills it's goal for most people most of the time for a fraction of the time and effort taken.

I dislike LLMs/GenAI mostly due to ethical reasons (I'm aware the code quality from LLMs has increased over time), but I've already accepted the fact it's gonna be used a lot. In most cases, I can avoid software using it in favor of alternatives, but that's not always possible, such as Linux itself. I'm not gonna risk staying on outdated kernels, and I don't have the time and knowledge to maintain a fork of a whole kernel.

As for my own code, I refuse to let AI touch it because I like thinking about how to code something, and I get really happy when it works.