Welcome to LocalLLaMA! Here we discuss running and developing machine learning models at home. Lets explore cutting edge open source neural network technology together.
Get support from the community! Ask questions, share prompts, discuss benchmarks, get hyped at the latest and greatest model releases! Enjoy talking about our awesome hobby.
As ambassadors of the self-hosting machine learning community, we strive to support each other and share our enthusiasm in a positive constructive way.
Rules:
Rule 1 - No harassment or personal character attacks of community members. I.E no namecalling, no generalizing entire groups of people that make up our community, no baseless personal insults.
Rule 2 - No comparing artificial intelligence/machine learning models to cryptocurrency. I.E no comparing the usefulness of models to that of NFTs, no comparing the resource usage required to train a model is anything close to maintaining a blockchain/ mining for crypto, no implying its just a fad/bubble that will leave people with nothing of value when it burst.
Rule 3 - No comparing artificial intelligence/machine learning to simple text prediction algorithms. I.E statements such as "llms are basically just simple text predictions like what your phone keyboard autocorrect uses, and they're still using the same algorithms since <over 10 years ago>.
Rule 4 - No implying that models are devoid of purpose or potential for enriching peoples lives.
I wonder how much of that money comes from selling cards to LLM-providers, and then using the money from those cards to invest in LLM-providers so they can buy more cards. And then having them use those cards as collateral for loans to get more cards.
All for consumable products that consistently fail to provide a cent worth of profit.
Q3 FY 2026 ended on 10-2025.
wut.
The dates, just like anything else within our economy, aren't real.
yeah did my head in as well, it's like they made up their own financial quarters
Kind of crazy that the thing that people generally consider to be better than anyone else at is just their side hustle now.
On the bright side, there is actual number behind the nvidia stock hype, but n the down side, will the need for GPU stay so high for long enough for nvidia to pay the massive investment to keep up with production? Or will they end up with empty factories in 2 years?
It looks like the progression of cancer.
It remains to be seen whether we'll have surgical intervention, or a dead patient in a few years.
At least AMD/ATI and Nvidia came up when gaming was the core reason to buy a dedicated graphics card. They have the dev pipeline to at least still make good drivers for gaming. Chart makes me think - that's why at least the proprietary drivers for Mali, PowerVR, Adreno, etc are all so mediocre when it comes to games. AMD great on Linux. Nvidia great just proprietary. Intel it was well regarded until they made Arc cards and people started comparing them to AMD/Nvidia and the Linux/Windows performance gap for Arc cards was very noticable. Qualcomm hyped up Linux support before the X Elite. Still mediocre.
The chart makes me think, the only hope for good drivers for gaming from non-AMD/Linux will be the open source Adreno driver. The 8 Elite and 8 Elite Gen 5 should be getting initial support early next year.
https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38450
The rest, which company is going to put in money to do a driver fix for some video game with buggy rendering from like 2010 when they can be focusing on fixing any bugs brought up having to do with pytorch and whatever other stuff people use
You would think with all the money from data centers their support for enterprise GPU use would be great. But as someone who's tried to install it, I can confirm it's really basic and the documentation is sparse.
My main issue is that, the life cycle of their entreprise grade GPU are way too short (and the gaming one are worse) . My job involve some obsolescence management and while the GPU obsolescence isn't the most complicated one, while we finished system validation on a GPU generation it's already time for purchasing to send a last buy order
Also curious of what the other segment really is
Also curious of what the other segment really is
Nintendo Switches, Nvidia Shield, and G-Force Now?
The icons beside Other (Eye, Car, Factory) have me guessing surveillance, automotive, and industrial.
Recently there are news reports saying companies like Meta are overstating their gpu life cycle length to show paper profits, what's a typical life cycle length for you? I'm hoping we'll get a massive glut of discounted gpus when its time for datacenters to upgrade.
My problem isn't much GPU dying, or getting not powerful enough, but that as we work in a "certified environment", everyone sourcing can't procure anymore a GPU model, we can't just buy the next generation and assume it works, but have to run an extensive list of test, then update documentations including list of supported GPU per system version(especially for service) .
Basically, it's the kind of stuff where you can loose a lot of time (and which isn't always high on priority list) so feel like as soon as we finally approved the usage of a GPU version, we have our suplier telling us it's not available anymore. I would love to have a GPU manufacturer offering 10-15 years of market availability (Like we get for FPGA) .
