I just want a portable self hosted LLM for specific tasks like programming or language learning.
Welcome to the top of the sigmoid curve.
If you were wondering what 1999 felt like WRT to the internet, well, here we are. The Matrix was still fresh in everyone’s mind and a lot of online tech innovation kinda plateaued, followed by some “market adjustments.”
I think it’s more likely a compound sigmoid (don’t Google that). LLMs are composed of distinct technologies working together. As we’ve reached the inflection point of the scaling for one, we’ve pivoted implementations to get back on track. Notably, context windows are no longer an issue. But the most recent pivot came just this week, allowing for a huge jump in performance. There are more promising stepping stones coming into view. Is the exponential curve just a series of sigmoids stacked too close together? In any case, the article’s correct - just adding more compute to the same exact implementation hasn’t enabled scaling exponentially.
Short on the AI stocks before it crash!
Oh no!
Anyway…
I’ve been hearing about the imminent crash for the last two years. New money keeps getting injected into the system. The bubble can’t deflate while both the public and private sector have an unlimited lung capacity to keep puffing into it. FFS, bitcoin is on a tear right now, just because Trump won the election.
This bullshit isn’t going away. Its only going to get forced down our throats harder and harder, until we swallow or choke on it.
With the right level of Government support, bubbles can seemingly go on for literal decades. Case in point, Australian housing since the late 90s has been on an uninterrupted tear (yes, even in ‘08 and ‘20).
I think I’ve heard about enough of experts predicting the future lately.
Thank fuck. Can we have cheaper graphics cards again please?
I’m sure a RTX 4090 is very impressive, but it’s not £1800 impressive.
Just wait for the 5090 prices…
I just don’t get whey they’re so desperate to cripple the low end cards.
Like I’m sure the low RAM and speed is fine at 1080p, but my brother in Christ it is 2024. 4K displays have been standard for a decade. I’m not sure when PC gamers went from “behold thine might from thou potato boxes” to “I guess I’ll play at 1080p with upscaling if I can have a nice reflection”.
4k displays are not at all standard and certainly not for a decade. 1440p is. And it hasn’t been that long since the market share of 1440p overtook that of 1080p according to the Steam Hardware survey IIRC.
Maybe not monitors, but certainly they are standard for TVs (which are now just monitors with Android TV and a tuner built in).
That doesn’t really matter if people on PC don’t game on it, does it?
These are the primary display resolutions from the Steam Hardware Survey.
I think it’s just an upselling strategy, although I agree I don’t think it makes much sense. Budget gamers really should look to AMD these days, but unfortunately Nvidia’s brand power is ridiculous.
An the issue for PC gamers is that Nvidia has spent the last few years convincing devs to shovel DLSS into everything, rather than a generic upscaling solution that other vendors could just drop their own algorithms into, meaning there’s a ton of games that won’t upscale nicely on anything else.
Sorry, crypto is back in season.
I swapped to AMD this generation and it’s still expensive.
A well researched pre-owned is the way to go. I bought a 6900xt a couple years ago for a deal.
I used to buy broken video cards on ebay for ~$25-50. The ones that run, but shut off have clogged heat sinks. No tools or parts required. Just blow out the dust. Obviously more risky, but sometimes you can hit gold.
If you can buy a ten and one works, you’ve saved money. Two work and you’re making money. The only question is whether the tenth card really will work or not.
I used to get EVGA bstock which was reasonable but they got out of the business 😞
It’s so funny how all this is only a problem within a capitalist frame of reference.
What they call “AI” is only “intelligent” within a capitalist frame of reference, too.
Fingers crossed.
Marcus is right, incremental improvements in AIs like ChatGPT will not lead to AGI and were never on that course to begin with. What LLMs do is fundamentally not “intelligence”, they just imitate human response based on existing human-generated content. This can produce usable results, but not because the LLM has any understanding of the question. Since the current AI surge is based almost entirely on LLMs, the delusion that the industry will soon achieve AGI is doomed to fall apart - but not until a lot of smart speculators have gotten in and out and made a pile of money.
Yay
The hype should go the other way. Instead of bigger and bigger models that do more and more - have smaller models that are just as effective. Get them onto personal computers; get them onto phones; get them onto Arduino minis that cost $20 - and then have those models be as good as the big LLMs and Image gen programs.
Other than with language models, this has already happened: Take a look at apps such as Merlin Bird ID (identifies birds fairly well by sound and somewhat okay visually), WhoBird (identifies birds by sound, ) Seek (visually identifies plants, fungi, insects, and animals). All of them work offline. IMO these are much better uses of ML than spammer-friendly text generation.
those are all classification problems, which is a fundamentally different kind of problem with less open-ended solutions, so it’s not surprising that they are easier to train and deploy.
Platnet and iNaturalist are pretty good for plant identification as well, I use them all the time to find out what’s volunteering in my garden. Just looked them up and it turns out iNaturalist is by Seek.
This has already started to happen. The new llama3.2 model is only 3.7GB and it WAAAAY faster than anything else. It can thow a wall of text at you in just a couple of seconds. You’re still not running it on $20 hardware, but you no longer need a 3090 to have something useful.
Well, you see, that’s the really hard part of LLMs. Getting good results is a direct function of the size of the model. The bigger the model, the more effective it can be at its task. However, there’s something called compute efficient frontier (technical but neatly explained video about it). Basically you can’t make a model more effective at their computations beyond said linear boundary for any given size. The only way to make a model better, is to make it larger (what most mega corps have been doing) or radically change the algorithms and method underlying the model. But the latter has been proving to be extraordinarily hard. Mostly because to understand what is going on inside the model you need to think in rather abstract and esoteric mathematical principles that bend your mind backwards. You can compress an already trained model to run on smaller hardware. But to train them, you still need the humongously large datasets and power hungry processing. This is compounded by the fact that larger and larger models are ever more expensive while providing rapidly diminishing returns. Oh, and we are quickly running out of quality usable data, so shoveling more data after a certain point starts to actually provide worse results unless you dedicate thousands of hours of human labor producing, collecting and cleaning the new data. That’s all even before you have to address data poisoning, where previously LLM generated data is fed back to train a model but it is very hard to prevent it from devolving into incoherence after a couple of generations.
this is learning completely the wrong lesson. it has been well-known for a long time and very well demonstrated that smaller models trained on better-curated data can outperform larger ones trained using brute force “scaling”. this idea that “bigger is better” needs to die, quickly, or else we’re headed towards not only an AI winter but an even worse climate catastrophe as the energy requirements of AI inference on huge models obliterate progress on decarbonization overall.
That would be innovation, which I’m convinced no company can do anymore.
It feels like I learn that one of our modern innovations was already thought up and written down into a book in the 1950s, and just wasn’t possible at that time due to some limitation in memory, precision, or some other metric. All we did was do 5 decades of marginal improvement to get to it, while not innovating much at all.
Are you talking about something specific?
This is why you’re seeing news articles from Sam Altman saying that AGI will blow past us without any societal impact. He’s trying to lessen the blow of the bubble bursting for AI/ML.
Huh?
The smartphone improvements hit a rubber wall a few years ago (disregarding folding screens, that compose a small market share, improvement rate slowed down drastically), and the industry is doing fine. It’s not growing like it use to, but that just means people are keeping their smartphones for longer periods of time, not that people stopped using them.
Even if AI were to completely freeze right now, people will continue using it.
Why are people reacting like AI is going to get dropped?
It’s absurdly unprofitable. OpenAI has billions of dollars in debt. It absolutely burns through energy and requires a lot of expensive hardware. People aren’t willing to pay enough to make it break even, let alone profit
People differentiate AI (the technology) from AI (the product being peddled by big corporations) without making clear that nuance (Or they mean just LLMs, or they aren’t even aware the technology has a grassroots adoption outside of those big corporations). It will take time, and the bubble bursting might very well be a good thing for the technology into the future. If something is only know for it’s capitalistic exploits it’ll continue to be seen unfavorably even when it’s proven it’s value to those who care to look at it with an open mind. I read it mostly as those people rejoicing over those big corporations getting shafted for their greedy practices.
People are dumping billions of dollars into it, mostly power, but it cannot turn profit.
So the companies who, for example, revived a nuclear power facility in order to feed their machine with ever diminishing returns of quality output are going to shut everything down at massive losses and countless hours of human work and lifespan thrown down the drain.
This will have an economic impact quite large as many newly created jobs go up in smoke and businesses who structured around the assumption of continued availability of high end AI need to reorganize or go out of business.
Search up the Dot Com Bubble.
Because in some eyes, infinite rapid growth is the only measure of success.
People pay real money for smartphones.
People pay real Money for AIaaS as well…
Because novelty is all it has. As soon as it stops improving in a way that makes people say “oh that’s neat”, it has to stand on the practical merits of its capabilities, which is, well, not much.
I’m so baffled by this take. “Create a terraform module that implements two S3 buckets with cross-region bidirectional replication. Include standard module files like linting rules and enable precommit.” Could I write that? Yes. But does this provide an outstanding stub to start from? Also yes.
And beyond programming, it is otherwise having positive impact on science and medicine too. I mean, anybody who doesn’t see any merit has their head in the sand. That of course must be balanced with not falling for the hype, but the merits are very real.
There’s a pretty big difference between chatGPT and the science/medicine AIs.
And keep in mind that for LLMs and other chatbots, it’s not that they aren’t useful at all but that they aren’t useful enough to justify their costs. Microsoft is struggling to get significant uptake for Copilot addons in Microsoft 365, and this is when AI companies are still in their “sell below cost and light VC money on fire to survive long enough to gain market share” phase. What happens when the VC money dries up and AI companies have to double their prices (or more) in order to make enough revenue to cover their costs?
I understand that it makes less sense to spend in model size if it isn’t giving back performance, but why would so much money be spent on larger LLMs then?
Nothing to argue with there. I agree. Many companies will go out of business. Fortunately we’ll still have the llama3’s and mistral’s laying around that I can run locally. On the other hand cost justification is a difficult equation with many variables, so maybe it is or will be in some cases worth the cost. I’m just saying there is some merit.
The merits are real. I do understand the deep mistrust people have for tech companies, but there’s far too much throwing out of the baby with the bath water.
As a solo developer, LLMs are a game-changer. They’ve allowed me to make amazing progress on some of my own projects that I’ve been stuck on for ages.
But it’s not just technical subjects that benefit from LLMs. ChatGPT has been a great travel guide for me. I uploaded a pic of some architecture in Berlin and it went into the history of it, I asked it about some damage to an old church in Spain - turned out to be from the Spanish civil war, where revolutionaries had been mowed down by Franco’s firing squads.
Just today, I was getting help from an LLM for an email to a Portuguese removals company. I sent my message in English with a Portuguese translation, but the guy just replied back with a single sentence in broken English:
“Yes a can , need tho mow m3 you need delivery after e gif the price”
The first bit is pretty obviously “Yes I can” but I couldn’t really be sure what he was trying to say with the rest of it. So I asked ChatGPT who responded:
It seems he’s saying he can handle the delivery but needs to know the total volume (in cubic meters) of your items before he can provide a price. Here’s how I’d interpret it:
“Yes, I can [do the delivery]. I need to know the [volume] in m³ for delivery, and then I’ll give you the price.”
Thanks to LLMs, I’m able to accomplish so many things that would have previously taken multiple internet searches and way more effort.
Okay now justify the cost it took to create the tool.
Hope?
It’s been 5 minutes since the new thing did a new thing. Is it the end?
As I use copilot to write software, I have a hard time seeing how it’ll get better than it already is. The fundamental problem of all machine learning is that the training data has to be good enough to solve the problem. So the problems I run into make sense, like:
- Copilot can’t read my mind and figure out what I’m trying to do.
- I’m working on an uncommon problem where the typical solutions don’t work
- Copilot is unable to tell when it doesn’t “know” the answer, because of course it’s just simulating communication and doesn’t really know anything.
2 and 3 could be alleviated, but probably not solved completely with more and better data or engineering changes - but obviously AI developers started by training the models on the most useful data and strategies that they think work best. 1 seems fundamentally unsolvable.
I think there could be some more advances in finding more and better use cases, but I’m a pessimist when it comes to any serious advances in the underlying technology.
Not copilot, but I run into a fourth problem:
4. The LLM gets hung up on insisting that a newer feature of the language I’m using is wrong and keeps focusing on “fixing” it, even though it has access to the newest correct specifications where the feature is explicitly defined and explained.I’ve also run into this when trying to program in Rust. It just says that the newest features don’t exist and keeps rolling back to an unsupported library.
Oh god yes, ran into this asking for a shell.nix file with a handful of tricky dependencies. It kept trying to do this insanely complicated temporary pull and build from git instead of just a 6 line file asking for the right packages.
“This code is giving me a return value of X instead of Y”
“Ah the reason you’re having trouble is because you initialized this list with brackets instead of
new().”“How would a syntax error give me an incorrect return”
“You’re right, thanks for correcting me!”
“Ok so like… The problem though.”
Yeah, once you have to question its answer, it’s all over. It got stuck and gave you the next best answer in it’s weights which was absolutely wrong.
You can always restart the convo, re-insert the code and say what’s wrong in a slightly different way and hope the random noise generator leads it down a better path :)
I’m doing some stuff with translation now, and I’m finding you can restart the session, run the same prompt and get better or worse versions of a translation. After a few runs, you can take all the output and ask it to rank each translation on correctness and critique them. I’m still not completely happy with the output, but it does seem that sometime if you MUST get AI to answer the question, there can be value in making it answer it across more than one session.
So you use other people’s open source code without crediting the authors or respecting their license conditions? Good for you, parasite.
I completely understand where you’re coming from, and I absolutely agree with you, genAI is copyright infringement on a weapons-grade scale. With that said, though, in my opinion, I don’t know if calling people parasites like this will really convince people, or change anything. I don’t want to tone police you, if you want to tell people to get fucked, then go ahead, but I think being a bit more sympathetic to your fellow programmers and actually trying to help them see things from our perspective might actually change some minds. Just something to think about. I don’t have all the answers, feel free to ignore me. Much love!
Very frequently, yes. As well as closed source code and intellectual property of all kinds. Anyone who tells you otherwise is a liar.
Ah, I guess I’ll have to question why I am lying to myself then. Don’t be a douchebag. Don’t use open source without respecting copyrights & licenses. The authors are already providing their work for free. Don’t shit on that legacy.
Programmers don’t have the luxury of using inferior toolsets.
That statement is as dumb as it is non-sensical.
Ahh right, so when I use copilot to autocomplete the creation of more tests in exactly the same style of the tests I manually created with my own conscious thought, you’re saying that it’s really just copying what someone else wrote? If you really believe that, then you clearly don’t understand how LLMs work.
I know both LLM mechanisms better than you, it would appear, and my point is not so weak that I would have to fabricate a strawman that I then claim is what you said, to proceed to argue the strawman.
Using LLMs trained on other people’s source code is parasitic behaviour and violates copyrights and licenses.
