You must log in or # to comment.
At least anecdotally, Andreas over at 82MHz.net tried running a AI model locally on his laptop and it took over 10 minutes for just one prompt.
OK just the 4th sentence clearly shows this person has no clue what they’re talking about.
Yep, clueless. I stopped reading at that point. For the audience, large language models come in all sizes and you can run some small but useful ones fairly quickly even without a GPU. They keep getting more capable for the size as well. Remember the uproar about Deepseek R1? Well, progress hasn’t stopped.
It’s not even that. It’s like trying to run an AAA game on a 10 year old laptop and complaining the game is garbage because your frame rates are too low.
AI is the evolution of tools. Like any other modern tool, either learn it and use it or be left behind.
Is Instagram a modern tool too? Just - felt segregated due to not using it and such shit in my teens, and now it’s apparently something from the past, and I still haven’t used it.
And a great many tools have a brief period of excitement before people realize they aren’t actually all that useful. (“The Segway will change the way everyone travels!”) There are aspects of limited AI that are quite useful. There are other aspects that are counter-productive at the current level of capability. Marketing hype is pushing anything with AI in the name, but it will all settle out eventually. When it does, a lot of people will have wasted a lot of time, and caused some real damage, by relying on the parts that are not yet practical.
marketing hype is pushing anything with AI in the name, but it will all settle out eventually
Agreed. “use it or be left behind” itself sounds like a phrase straight out of a marketing pitch from every single AI-centric" company that pushes their “revolutionary” product. It’s a phrase that i hear daily from c-suite executives that know very little of what they’re talking about. AI (specifically generative) has its usecases, but it’s nowhere near where the marketing says it is. And when it finally does get there, i think people are going to be surprised when they don’t find themselves in the utopia that they’ve been promised.
Absolutely agreed all around.
For me, in my job, AI has been a fantastic tool. It feels like I was using a plain screw driver and now I have a cordless, light power drill. It really is doing “grunt” work for me so I can focus on the more complex tasks.
I find it funny that in the year 2000 while attending philosophy at University of Copenhagen I predicted strong AI around 2035. This was based on calculations of computational power, and estimates of software development trailing a bit.
At the time I had already been interested in AI development and matters of consciousness for many years. And I was a decent programmer. I already made self modifying code back in 1982. So I made this prediction at a time where AI wasn’t a very popular topic, and in the middle of a decades long futile desert walk without much progress.And for 15 about years, very little continued to happen. It was pretty obvious the approach behind for instance Deep Blue wasn’t the way forward. But that seemed to be the norm for a long time.
But it looks to me that the understanding of how to build a strong AI is much much closer now, as I expected. We might actually be halfway there!
I think we are pretty close to having the computational power needed now in AI specific datacenter clusters, but the software isn’t quite there yet.I’m honestly not that interested in the current level of AI, although LLM can yield very impressive results at times, it’s also flawed, and I see it as somewhat transitional.
For instance partially self driving cars are kind of irrelevant IMO. But truly self driving cars will make all the difference regarding how useful it is, and be a cool achievement for current level of AI evolution when achieved.So current level AI can be useful, but when we achieve strong AI it will make all the difference!
Edit PS:
Obviously my prediction relied on the assumption that brains and consciousness are natural phenomena, that don’t require a god. An assumption I personally consider a fact.I find it funny that in the year 2000 while attending philosophy at University of Copenhagen I predicted strong AI around 2035.
That seems to be aging well. But what is the definition of “strong AI”?
Self aware consciousness on a human level. So it’s still far from a sure thing, because we haven’t figured consciousness out yet.
But I’m still very happy with my prediction, because AI is now at a way more useful and versatile level than ever, the use is already very widespread, and the research and investments have exploded the past decade. And AI can do things already that used to be impossible, for instance in image and movie generation and manipulation.But I think the code will be broken soon, because self awareness is a thing of many degrees. For instance a dog is IMO obviously self aware, but it isn’t universally recognized, because it doesn’t have the same degree of selv awareness humans have.
This is a problem that dates back to 17th century and Descartes, who claimed for instance horses and dogs were mere automatons, and therefore couldn’t feel pain.
This of course completely in line with the Christian doctrine that animals don’t have souls.
But to me it seems self awareness like emotions don’t have to start at human level, it can start at a simpler level, that then can be developed further.PS:
It’s true animals don’t have souls, in the sense of something magical provided by a god, because nobody has. Souls are not necessary to explain self awareness or consciousness or emotions.
It was pretty obvious the approach behind for instance Deep Blue wasn’t the way forward.
That’s a weird example to pick. What exactly about Deep Blue do you think wasn’t the way forward?
Deep blue was mostly based on raw computational power, with very little ability to actually judge whether a draw was “good” without calculating the possibilities following it.
As I understand it, it only worked on Chess as a “mathematical” problem, and was incapable of judging strategic positions, except if it had “seen” it before, and already calculated the possible outcomes.
In short, there was very little intelligence, it was based only on memory and massive calculation power. Which indeed are aspects of intelligence, but only on a very low level.
Regarding energy/water use:
ChatGPT uses 3 Wh. This is enough energy to: […] Play a gaming console for 1 minute.
If you want to prompt ChatGPT 40 times, you can just stop your shower 1 second early. If you normally take a 5 minute shower, set a timer for 299 seconds instead, and you’ll have saved enough water to justify 40 ChatGPT prompts.
(Source: https://andymasley.substack.com/p/a-cheat-sheet-for-conversations-about)
These endless “AI bad” articles are annoying. It’s just click bait at this point.
Energy use: false. His example was someone using a 13 year old laptop to get a result and then extrapolating energy use from that. Running ai locally is the same energy as playing a 3d AAA game for the same time. No one screams about the energy footprint of playing games.
AAA game development energy use ( thousands of developers all with watt burning gpus spending years creating assets) dwarfs AI model building energy use.
Copyright, yes it’s a problem and should be fixed. But stealing is part of capitalism. Google search itself is based on stealing content and then selling ads to find that content. The entire “oh we might send some clicks your way that you might be able to compensated for” is backwards.
His last reason was new and completely absurd: he doesn’t like AI because he doesn’t like Musk. Given the public hatred between OpenAI and Musk it’s bizarre. Yes Musk has his own AI. But Musk also has electric cars, and space travel. Does the author hate all EV’s too? If course not, that argument was added by the author as a troll to get engagement.
I never thought I’d see the web fight for copyright.
For me it seems like all AI issues boil down to “I don’t like stions” which is fine but its kinda delusional to pretend it’s something else be it silly energy use complaints or hypocritical copyright nonsense.
Hi, I’m the writer of the article.
To be clear I am not trying to attack anyone who uses AI, just explain why I don’t use it myself.
Energy use: false
I don’t dispute that AI energy is/might be comparable to other things like making a AAA game (or other things like traveling). I also don’t want to say that ‘AI is bad’. However if I used AI more, I would still play the same amount of video games, thus increasing the total energy use. If I was to use AI it would probably replace lower energy activities like writing or searching the internet.
Copyright, yes it’s a problem and should be fixed. But stealing is part of capitalism. Google search itself is based on stealing content and then selling ads to find that content.
I agree with you that the copyright angle is a bad way to attack AI, however AI does seem like it ‘gives back’ to creatives even less than other things like search as well as actively competing with them in a way that search doesn’t. This isn’t my main objection so I don’t really want to focus on it.
His last reason was new and completely absurd
I considered leaving out the “I just don’t like it” reason but I wanted to be completely transparent that my decision isn’t objective. This is only one reason out of many - if it was just this problem then I would be quicker to ignore it. I get your point about EV’s - I don’t hate them despite the fact that Musk is/was an advocate for them. If I was to use a AI it would be something like Jan.ai which @Flagstaff@programming.dev mentioned.
Do you agree with me on my other main point on reliability?
However if I used AI more, I would still play the same amount of video games, thus increasing the total energy use.
Then that’s like writing about the evils of cars while driving a giant SUV for fun.
Do you agree with me on my other main point on reliability?
The Google AI forced on me in searches has seemed correct because every sentence has a footnote with a link to source that I usually click. The OpenAI code generation I used a year ago was brilliant. It wrote working VBScript for me which was a language I had no desire to learn. The microcontroller code for another project was also fantastic because it gave me an outline to start working with.
Running ai locally is the same energy as playing a 3d AAA game for the same time
I wonder if they’re factoring in the energy usage to train the model. That’s what consumes most of the power.
I addressed that in my second paragraph.
In another thread someone brought it up so I did some quick math to see if it was true:
Gta5 cost $300 million. 4000 developers each with the latest GPU burning hundreds of watts per employee to create the assets. A rough estimate of 750watt pc, 4,000 developers, 8 hour a day, 300 days a year, 5 years = 36 giga watt-hours. That’s the energy to power 3.6 million homes for a year and I’m not even including the HVAC costs of the office space. For 1 game.
AI training energy use is small in comparison. ChatGPT 4 cost $80m to train.
it’s okay, Rockstar is only launching one game every 15 years now /s
If gta5 was the only game sold, it wouldn’t be an energy issue!
OP said “people like Musk” not just Musk. He’s just the easiest example to use.
There’s a huge difference between an outright Nazi like Musk and an average techbro.
Given how much of Tech bro ideology is informed by types like Nick Land and Curtis Yarvin, no there isn’t.
They are more alike than they are different to normal people.
That’s my point : no.
No, there isn’t. Musk is a tech bro cranked up to 11.
Copyright, yes it’s a problem and should be fixed.
The quick fix: stick to open-source like Jan.ai.
Long-term solution: make profiting AI companies pay for UBI. How to actually calculate that, though, is anyone’s guess…
Don’t make “profiteering AI companies” pay for UBI. Make all companies pay for UBI. Just tax their income and turn it around into UBI payments.
One of the major benefits of UBI is how simple it is. The simpler the system is the harder it is to game it. If you put a bunch of caveats on which companies pay more or pay less based on various factors, then there’ll be tons of faffing about to dodge those taxes.
make profiting AI companies pay for UBI
As I said, many companies steal content and repackage it for sale. Google did it long before AI. AI is only the most recent offender. Courts have been splitting hairs for decades over music similarities and that’s ignoring that entire genres are based on copying the work of influential artists.
I agree on the part that Musk sucks, OpenAI also sucks.
And yup, open source (if you can really call them that, I’d say they’re more like openly available) locally hosted LLMs are cool and have gotten pretty efficient nowadays.
My 5 year old M1 MacBook Pro runs models like Qwen3:14b at decent speeds and it’s quite capable (although I only ever use it for bullshitting lol).
I agree, there are still good reasons not to use commercial AI products though.
https://www.mintpressnews.com/trump-killed-minerva-stargate-make-secret-more-dangerous/289313
A new AI/informational war arms race? Whatever, because…
I just don’t like it
Copyright, yes it’s a problem and should be fixed.
No, this is just playing into another of the common anti-AI fallacies.
Training an AI does not do anything that copyright is even involved with, let alone prohibited by. Copyright is solely concerned with the copying of specific expressions of ideas, not about the ideas themselves. When an AI trains on data it isn’t copying the data, the model doesn’t “contain” the training data in any meaningful sense. And the output of the AI is even further removed.
People who insist that AI training is violating copyright are advocating for ideas and styles to be covered by copyright. Or rather by some other entirely new type of IP protection, since as I said this is nothing at all like what copyright already deals with. This would be an utterly terrible thing for culture and free expression in general if it were to come to pass.
I get where this impulse comes from. Modern society has instilled a general sense that everything has to be “owned” by someone, even completely abstract things. Everyone thinks that they’re owed payment for everything that they can possibly demand payment for, even if it’s something that just yesterday they were doing purely for fun and releasing to the world without a care. There’s this base impulse of “mine! Therefore I must control it!” Ironically, it’s what leads to the capitalist hellscape so many people are decrying at the same time they demand more.
When an AI trains on data it isn’t copying the data, the model doesn’t “contain” the training data in any meaningful sense.
And what’s your evidence for this claim? It seems to be false given the times people have tricked LLMs into spitting out verbatim or near-verbatim copies of training data. See this article as one of many examples out there.
People who insist that AI training is violating copyright are advocating for ideas and styles to be covered by copyright.
Again, what’s the evidence for this? Why do you think that of all the observable patterns, the AI will specifically copy “ideas” and “styles” but never copyrighted works of art? The examples from the above article contradict this as well. AIs don’t seem to be able to distinguish between abstract ideas like “plumbers fix pipes” and specific copyright-protected works of art. They’ll happily reproduce either one.
That article is over a year old. The NYT case against OpenAI turned out to be quite flimsy, their evidence was heavily massaged. What they did was pick an article of theirs that was widely copied across the Internet (and thus likely to be “overfit”, a flaw in training that AI trainers actively avoid nowadays) and then they’d give ChatGPT the first 90% of the article and tell it to complete the rest. They tried over and over again until eventually something that closely resembled the remaining 10% came out, at which point they took a snapshot and went “aha, copyright violated!”
They had to spend a lot of effort to get that flimsy case. It likely wouldn’t work on a modern AI, training techniques are much better now. Overfitting is better avoided and synthetic data is used.
Why do you think that of all the observable patterns, the AI will specifically copy “ideas” and “styles” but never copyrighted works of art?
Because it’s literally physically impossible. The classic example is Stable Diffusion 1.5, which had a model size of around 4GB and was trained on over 5 billion images (the LAION5B dataset). If it was actually storing the images it was being trained on then it would be compressing them to under 1 byte of data.
AIs don’t seem to be able to distinguish between abstract ideas like “plumbers fix pipes” and specific copyright-protected works of art.
This is simply incorrect.
The NYT was just one example. The Mario examples didn’t require any such techniques. Not that it matters. Whether it’s easy or hard to reproduce such an example, it is definitive proof that the information can in fact be encoded in some way inside of the model, contradicting your claim that it is not.
If it was actually storing the images it was being trained on then it would be compressing them to under 1 byte of data.
Storing a copy of the entire dataset is not a prerequisite to reproducing copyright-protected elements of someone’s work. Mario’s likeness itself is a protected work of art even if you don’t exactly reproduce any (let alone every) image that contained him in the training data. The possibility of fitting the entirety of the dataset inside a model is completely irrelevant to the discussion.
This is simply incorrect.
Yet evidence supports it, while you have presented none to support your claims.
Learning what a character looks like is not a copyright violation. I’m not a great artist but I could probably draw a picture that’s recognizably Mario, does that mean my brain is a violation of copyright somehow?
Yet evidence supports it, while you have presented none to support your claims.
I presented some, you actually referenced what I presented in the very comment where you’re saying I presented none.
You can actually support your case very simply and easily. Just find the case law where AI training has been ruled a copyright violation. It’s been a couple of years now (as evidenced by the age of that news article you dug up), yet all the lawsuits are languishing or defunct.
Learning what a character looks like is not a copyright violation
And nobody claimed it was. But you’re claiming that this knowledge cannot possibly be used to make a work that infringes on the original. This analogy about whether brains are copyright violations make no sense and is not equivalent to your initial claim.
Just find the case law where AI training has been ruled a copyright violation.
But that’s not what I claimed is happening. It’s also not the opposite of what you claimed. You claimed that AI training is not even in the domain of copyright, which is different from something that is possibly in that domain, but is ruled to not be infringing. Also, this all started by you responding to another user saying the copyright situation “should be fixed”. As in they (and I) don’t agree that the current situation is fair. A current court ruling cannot prove that things should change. That makes no sense.
Honestly, none of your responses have actually supported your initial position. You’re constantly moving to something else that sounds vaguely similar but is neither equivalent to what you said nor a direct response to my objections.
When an AI trains on data it isn’t copying the data, the model doesn’t “contain” the training data in any meaningful sense.
I’d say it can be a problem because there have been examples of getting AIs to spit out entire copyrighted passages. Furthermore, some works can have additional restrictions on their use. I couldn’t for example train an AI on Linux source code, have it spit out the exact source code, then slap my own proprietary commercial license on it to bypass GPL.
I’d say it can be a problem because there have been examples of getting AIs to spit out entire copyrighted passages.
Examples that have turned out to either be a result of great effort to force the output to be a copy, a result of poor training techniques that result in overfitting, or both combined.
If this is really such a straightforward case of copyright violation, surely there are court cases where it’s been ruled to be so? People keep arguing legality without ever referencing case law, just news articles.
Furthermore, some works can have additional restrictions on their use. I couldn’t for example train an AI on Linux source code, have it spit out the exact source code, then slap my own proprietary commercial license on it to bypass GPL.
That’s literally still just copyright. There’s no “additional restrictions” at play here.
GPL is a license that uses copyright law as enforcement.
Yes, that’s what I said. There are no “additional restrictions” from having a GPL license on something. The GPL license works by giving rights that weren’t already present under the default copyright. You can reject the GPL on an open sourced piece of software if you want to, but then you lose the additional rights that the GPL gives you.
If a larger youtuber steals the script and content of a video from a smaller youtuber, as far as i know, it wouldnt be illegal. It would hurt the smaller youtuber and benefit the larger one. It would make people mad if they found out about it, but there wouldnt be people who propose changing copyright law to include ideas
I am using youtubers as the example because this happened and a lot of people got angry and its similar to the AI situation
People can complain that something unethical is legal without proposing new copyright laws without flawsStealing script and content is a copyright violation. It could be a lawsuit but is usually ignored because of legal costs.
By stealing, i meant rewriting it, following the same points but using unique words. Is that illegal or are you referring to completely copying the text?
I thought you meant actual copying like when Linus tech tips used some Gamer’s Nexus script word for word.
Sure. But that’s not what’s happening when an AI is trained. It’s not “stealing” the script or content of the video, it’s analyzing them.
By analyzing, isnt it awarding points based on how well you can replicate it and redoing it in an attempt to obtain more points?
Very basically, yes. But the result is a model that doesn’t actually contain the training data, it’s too small for it to be physically possible.
