this post was submitted on 06 Jan 2025
745 points (97.7% liked)

memes

10826 readers
4822 users here now

Community rules

1. Be civilNo trolling, bigotry or other insulting / annoying behaviour

2. No politicsThis is non-politics community. For political memes please go to [email protected]

3. No recent repostsCheck for reposts when posting a meme, you can only repost after 1 month

4. No botsNo bots without the express approval of the mods or the admins

5. No Spam/AdsNo advertisements or spam. This is an instance rule and the only way to live.

Sister communities

founded 2 years ago
MODERATORS
 
top 50 comments
sorted by: hot top controversial new old
[–] [email protected] 6 points 1 day ago (1 children)

How much if a discount are you expecting to start gaming on a 30k card with no video output?

[–] [email protected] 2 points 1 day ago (1 children)

Not for gaming, for running AI open source models and other AI shenanigans. My 4080 Super has been filling my gaming needs and will for years to come, but it's not enough for my AI interests lol

The most I can get out of this 4080 is running a ~7B param model, but I want to run cooler shit like that new open source DeepSeek v3 that dropped the other day.

[–] [email protected] 6 points 23 hours ago (1 children)

So you're waiting for the AI bubble to burst because you can't wait to run all the cool new AI models?

[–] [email protected] 2 points 23 hours ago* (last edited 23 hours ago) (1 children)

Yea, the underlying tech is what interests me and I have a few potential use cases. Use cases that I would never entrust a random company with. For example, the concept of MS recall is cool, I'd never trust Microshits implementation though. But an open source local version that I'm in control of all the security implementations? Hell yea lol

[–] [email protected] 1 points 22 hours ago

That's the problem. If the use case is super cool, and 99% of people have no knowledge (or motivation) to set it up for themselves, the online services will keep existing, and the bubble won't really burst.

Even if some single companies fail (and they will, there some truly horrendous ideas getting funding), the big players will buy the GPUs wholesale before they hit ebay.

[–] [email protected] 12 points 1 day ago

The problem is that the enterprise level cards can't really perform at the consumer market level nor are they designed for it. Many don't even have video outputs.

[–] [email protected] 3 points 23 hours ago

Lol no, I mean it would be a bubble if it didn't provide anything useful, or transformative, but that's far from the truth.

Like it or not, even LLMs have been found to help in health treatments, mental support , workplace efficiency and so on

AI is here to stay, it's basically the next industrial revolution

[–] [email protected] 5 points 1 day ago

Nuclear energy could also be a weird side effect of the bubble.

[–] [email protected] 17 points 1 day ago (2 children)

I believe it is likely that there will be a burst at some point, just as with the dot-com burst.

But I think many people wrongly think that it will be the end of or a major setback for AI.

I see no reason why in twenty years AI won't be as prevalent as "dot-com's" are now.

[–] [email protected] 1 points 1 day ago

I agree, history always repeats itself. But perhaps the timing is different, it could be 20 years, 10 years, or 50 years who knows

[–] [email protected] 1 points 1 day ago* (last edited 1 day ago)

Some current directions in AI, such as LLMs, seem to be dead-ends in the sense that those approaches cannot be incrementally improved much further to, for example, eliminate hallucinations or simply be capable of using logic along with those probability engines in such a way as to, at minimum, exclude the logically impossible from the results.

The dot-com stuff on the other hand was the very first bubble from the very first wave into a whole new technological direction that had just been unlocked and gave access to an entire technological branch of new ways of doing things - it the result of the very first wave of investment around the technology domain of worldwide digital communications and all the other tech branches that became possible due to it.

Basically the Internet was like openning a door to a various new areas of Tech (curiously that wasn't even all that amazingly complex as Tech goes, kinda like a basic wheel isn't exactly complicated but look at all that became possible with its invention), whilst the current AI wave (which is mainly the latest wave of work in the branch of Neural Networks, which is over 3 decades old) is more like a handful of massivelly complicated solutions which are the product of decades of work in a specific direction, some of which work in such a way that they can't be significantly further improved and hence can't be made to get past certain problems it has (the most obvious example being LLM hallucinations).

So whilst I do think that in 20 years there will be some prevalence of AI tech companies in some domains were the AI solutions of this wave of development on it do work well enough (say, entity detection on images), I don't think that will be anywhere comparable to what happened in the 20 years following the start of a new Tech Age which triggered by the Internet.

Mind you, 2 decades is a lot of time in Tech terms, so maybe somebody will come up with a whole different approach to AI in the meanwhile that breaks through the inherent limitations of the current one, just don't count on it.

Edit: just wanted to add that I was there when Darpanet morphed into the Internet and the dot-net bubble that came out of it. At the time everybody and their dog was playing around with making a websites, people were trying new stuff on top of those websites, inventing new comms protocols, wiriting programs that talked to other programs over the network, creating entirely new business models anchored on making a website a storefront - the Internet was Freedom. This AI wave doesn't feel at all like that - sure plenty of people are deploying models created by others and trying them out, but very few are creating new models and a lot of that Tech comes pre-monetised and locked down by large companies who are trying to get money out of anything people do with it - the whole things is not at all like the "we've open this whole new domain, you guys figure out what to do with it" that was the birth of the Internet.

[–] [email protected] 5 points 1 day ago (2 children)

You don't want a used GPU that's been running overclocked for years on end bro.

[–] [email protected] 5 points 1 day ago

Multiple outlets including LTT and Gamers Nexus have debunked this.

The only thing you may have to do if you notice unusual performance is reapply thermal paste to the GPU, and that’s only because most thermal paste will dry out after years of sitting around or being used

[–] [email protected] 2 points 1 day ago

The price of gpus would go down as there would be less demand.

[–] [email protected] 3 points 1 day ago

Aw, I was hoping the whole thing would rhyme after the first line.

[–] [email protected] 46 points 1 day ago (1 children)

GPU prices are gonna get cheaper, annnnyyyy day now folks, any day now

[–] [email protected] 11 points 1 day ago (2 children)

That's what I'm saying since 2020, people don't have patience any more.

load more comments (2 replies)
[–] [email protected] 66 points 2 days ago (6 children)
[–] [email protected] 62 points 2 days ago

Just deserts.

[–] [email protected] 19 points 1 day ago

We aint found shit

load more comments (4 replies)
[–] [email protected] 11 points 1 day ago

There may be a dip in prices for a bit, but since covid, more companies have realized they can get away with manufacturing fewer units and selling them for a higher price.

[–] [email protected] 31 points 2 days ago

Same goes for the death of windows 10. I want me some cheap Linux boxes.

[–] [email protected] 16 points 1 day ago (2 children)

Unfortunately, this time around the majority of AI build up are GPUs that are likely difficult to accomodate in a random build.

If you want a GPU for graphics, well, many of them don't even have video ports.

If your use case doesn't need those, well, you might not be able to reasonably power and cool the sorts of chips that are being bought up.

The latest wrinkle is that a lot of that overbuying is likely to go towards Grace Blackwell, which is a standalone unit. Ironically despite being a product built around a GPU but needing a video port, their video port is driven by a non-nvidia chip.

load more comments (2 replies)
[–] [email protected] 15 points 1 day ago (1 children)

Crypto miners watcing AI bubble.

[–] [email protected] 12 points 1 day ago

Silverlight, what are you doing here? Go on, get outta here!

[–] [email protected] 9 points 1 day ago (2 children)

I just want the ai hate bubble to burst, the hype bubble probably needs to go as well but honestly I care less about that

[–] [email protected] 21 points 1 day ago (1 children)

I want the hype bubble to burst because I want them to go back to making AI for useful stuff like cancer screening and stop trying to cram it into my fridge or figure out how to avoid paying their workers, and the hate bubble isn't going to stop until that does.

load more comments (1 replies)
load more comments (1 replies)
[–] [email protected] 9 points 2 days ago* (last edited 2 days ago) (7 children)

The AI bubble burst is gonna wreck the market and send us into another recession. Especially if the crypto bubble bursts at a similar time.

[–] [email protected] 9 points 1 day ago

Those in charge brought it on themselves. They can’t see through their own hype and lies and cult-like worship. It’s the rest of us that will hurt the most though.

load more comments (6 replies)
[–] [email protected] 9 points 2 days ago* (last edited 2 days ago) (2 children)

Word of caution: these used GPUs will have spent a lot of time doing AI bullshit with as little downtime as possible. Make sure you get a good enough deal for the risk that it just dies in 2 months.

[–] [email protected] 1 points 23 hours ago

As little downtime as possible is actually better than having more downtime, as the major problems come from heating/cooling cycles

load more comments (1 replies)
[–] DudeImMacGyver 9 points 2 days ago

It's both for me

[–] [email protected] 4 points 1 day ago

I'll buy a dozen and build my own AI farm tyvm.

it would be amazing to have several models running and have them interact together for some real immersive DnD rp.

[–] [email protected] 0 points 1 day ago (2 children)

Why do people care about powerful GPUs? Just use whatever works.

[–] [email protected] 2 points 23 hours ago

Gaming aside, powerful GPUs are useful for a lot of work like 3D graphics, video editing, neural networks, and a lot of stuff i don't know about

[–] [email protected] 1 points 23 hours ago

Some things like vr also just barely work without a powerful GPU

I have a 3060 but with the sheer number of pixels that my reverb g2 has its really not enough

load more comments
view more: next ›