LLMs are no different, and I can’t agree with you that open source models are not a threat to the big players.

This is just plain wrong. Again, there are zero open weight models that haven’t been developed by private companies. These companies, at the same time, offer superior closed source models because that’s their whole business model.

… we have an amazing amount of open source projects that people do simply because they want to. Those are the people we should support and the ones who freely train and fine tune open source models.

They are not.

There might be some people fine tuning models, but I can confidently assure you that there isn’t a single non profit entity out there that is spending tens of thousands of dollars in compute alone, just to give their model away for free. And that doesn’t even begin to account for data collection.

To your point about “software doesn’t become faster with time” mother fucker I remember windows 95, you’re delusional if you don’t think we’ve come an insane amount. I remember webpages taking minutes to load, interlacing vs non to help with image loading.

What are you even talking about. Websites in the 90s took longer to load because connections back then ran at 56Kbps tops, or ~5KBs, with latencies in the order of 500 to 1000 ms, when the average website would be like 10KB. Nowadays, an online newspaper weights 5 to 20 MB with average bandwidths of hundreds of megabits per second, with latencies of 50-100ms. Web development and its traversals are in such a particularly shitty state, browsing the modern Internet on less than 4GB of RAM is borderline impossible. In other words, software has become slower, and hardware is doing the heavy lifting now. And I can say this because I work in the field.

Phones require specialized hardware and designing, to run and produce, LLMs only require normal consumer grade hardware and the desire to learn how to make it work.

This is such a massive mischaracterization.

First of all, it’s easier to put a phone together with off the self parts, than it is to build a meaningfully useful LLM even with $50,000 worth of hardware at one’s disposal. Second, running a LLM was never the issue. Being able to produce and run a meaningfully useful LLM that has no strings attached to private interests is.

Honestly, I think you are out of your depth. Being a hobbyist is fine, but holy crap please inform yourself. None of this shit is easy or free or even cheap to build and run, and every foundation model is controlled by private interests.

I’m just showing that as technology progresses and scales it generally becomes cheaper and peoples access increases, again were literally on the internet now and have phones in our pockets that can do it, whereas 40 years ago PCs were much more expensive and internet was slow as hell.

Even if LLMs were free to download and use, who is going to subsidize training and fine tuning, when it takes hundreds of millions of dollars? Also, LLMs are software, not hardware. If there’s anything that we know about software is that it doesn’t become faster with time, quite the opposite.

The thing I don’t understand is that people believe the BS when all this is out there in the clear. Massive corporations open source models that pose no risk to their bottom line, then they spend millions of dollars to market their newest and latest, rinse and repeat, all fuelled by debt. Thus, self hosting will never catch up, and when the money dries up, there will be zero incentive to make more advanced models more affordable. In fact, since most of the time model improvements scale following training and hardware expenditure, they will become more expensive.

We shouldn’t trust big tech, I’m on Lemmy so that should be a bit of a given lol.

Is it though?

Like here you are, telling me that an example of “technology progress” is that “were literally on the internet now and have phones in our pockets that can do it, whereas 40 years ago PCs were much more expensive and internet was slow as hell”, when the phone market is effectively controlled by two companies, Apple and Google. Now imagine the same landscape with LLMs.

Specs are great for short term discussions about requisites and implementation.

But there’s this old adage, “Confluence is where knowledge goes to die”. I don’t think I’ve ever worked at a company where this wasn’t true.

If you write a spec, there’s a non zero chance that nobody will update it in a year, because it has no effect to the bottom line, and engineers have to be willing to look them up every time they make changes to code, which is never the case.

You are betting on massive corporations having a change of heart and putting all their resources at the disposition of the public, for essentially free. Otherwise, AI will never be affordable in the sense that everyone could have free access to models that matter.

And I know that you said that self hosting is a possibility. But let’s be real here: public weight models are available because they pose no risk to the bottom line of the companies training them. There are zero competitive models trained by a non profit. But even if that wasn’t true, the current DRAM shortage is proof that these companies will never allow anyone to match them. Same goes for electricity and water.

Honestly, after all these years of witnessing big tech shitting all over us, I cannot understand where all these hopes come from. Would be endearing if it wasn’t so reckless.

Read the book. If half of what’s in it is true, these people should be put in jail for the longest time.

Thanks. I’m done.

Of course you are.

Good example of what LLMs are great for, making people suffer from persistent Dunning Kruger.

The corollary to this is that code will generally become of lower quality, as more seniors burn out from taking on purely reviewer roles.

I find myself frequently giving up on writing specs or skills for LLMs because even the most expensive and advanced models cannot produce production quality code. They can sometimes produce correct code, when multiple passes are done and the most egregious mistakes are ironed out, but at that point I’ve already burned $200 worth of tokens.

To the author’s point, if I need to make my specs so fine grained that I could write the code instead, what’s the benefit in relying on a LLM?

Ugh, chicken? Give me some free steak. Chop chop!

Again, CV is not new. Computational biological simulation isn’t new either. More computational power and better algorithms have been a source of significant progress in healthcare for many decades now. If we go back twenty years, protein folding simulation was all the rage, but of course most people outside CompSci hadn’t heard of it.

I call the current AI “hype” because all these advancements have been going on for a while, but most people are catching up only now because they got hooked on marketing material they see on the news.

Anyway, I’m going to paste my message here once again.

Funny that you call mine “ideological” though, since you are the one making claims without any substance, e.g. “it’s only going to get better”. How could you even know? Not even researchers at the very edge do. There have been concerns about the future availability and quality of data. Plenty of researchers have come forward pointing that poisoning a LLM is exceedingly easy. Really, how do you know that “it’s going to get better”? Explain that to me. What do you know that everybody else doesn’t?

How do you even know that AI, as we know it, it’s going to be revolutionary in the near future? Most people only know of technology successes because of survivorship bias, but I’ve been through several revolutions that faded out. How is this one different? And why would you think you’re right, when not even expert researchers are sure?

Now, are you an AI researcher? What do you know about any of this, exactly?

You haven’t answered a single one of my questions.

I’m going to assume that you fell for the hype and know nothing of what you’re talking about.

The author of the article is misrepresenting several historical facts.

The pope didn’t try to “ban” printed books, but keep publications under tight catholic control under threat of excommunication. If we were to apply this to the current AI landscape, the “church” would be a number of massive corporations fighting to keep their stolen data “closed”.

Fust wasn’t “chased out” because scribes feared a loss of influence. They already were notaries and bureaucrats, they were doing just fine. The issue was publishing control under the church mandate, which again, correlates to what AI companies are doing right now.

And I suspect your position comes from not doing any due diligence on the matter.

Funny that you call mine “ideological” though, since you are the one making claims without any substance, e.g. “it’s only going to get better”. How could you even know? Not even researchers at the very edge do. There have been concerns about the future availability and quality of data. Plenty of researchers have come forward pointing that poisoning a LLM is exceedingly easy. Really, how do you know that “it’s going to get better”? Explain that to me. What do you know that everybody else doesn’t?

How do you even know that AI, as we know it, it’s going to be revolutionary in the near future? Most people only know of technology successes because of survivorship bias, but I’ve been through several revolutions that faded out. How is this one different? And why would you think you’re right, when not even expert researchers are sure?

You haven’t read it, have you.

The studies we reviewed show that the use of AI has improved the radiologists’ performances, treatment response, diagnostic accuracy, and decision-making in handling complex cases.

Hardly a game changer of the magnitude you think of. Moreover, CV is not generative. Pattern matching on X-rays has been common for a while, and has little to do with the current heavily marketed landscape of LLMs for everything.

I figured you wouldn’t be able to look past your own personal experience. I’m sorry to say that most people outside your bubble cannot afford either the subscription nor the hardware to run usable LLMs locally.

“Sharing code is bad now” because a handful of companies scraped it and not only they haven’t given anything back, they are reselling it in different shapes, and telling people that now all that data is proprietary. So, yes, stolen is an apt word for it.

Anyway, all this talk about “democratizing” knowledge is bullshit. Libraries democratized knowledge. The internet democratized knowledge. Anyone can learn how to code if they put the time and read a book and practice.

But delegated thinking is the opposite of acquiring knowledge, so what the hell are you people yapping about.

You missed the context. I don’t blame you.

Tell me how in hell are you going to audit every single variation of code generated by a LLM, that’s equivalent to a whole library. I’ll wait.

You are missing the most important questions here: who can afford it, and who owns it.

It’s easy to be pro LLM when $20 a month is not a big deal.

I know it’s code. You are missing the point.

Any library with a critical user mass is auditable, because a fraction of those users would take the time to do so, whereas all LLM generated variations of the same library cannot and will never be auditable.

So it’s a search tool. Where are all those AI generated cancer treatments, then?

Regardless, it’s a tool that very few can afford at the level it might be genuinely useful for original research.

Printing presses made knowledge more widely available for everyone.

LLMs do the exact opposite.

Libraries can be audited. LLM generated code cannot.

Edit: to clarify, it is impossible to audit all LLM generated code across a number of projects, that would replace a single library. It simply won’t happen, because there will always be a non trivial number of users who will copy and paste code without inspecting it. In contrast, widely used open source libraries may be audited by a small subset of their users, and the rest would benefit from that.