Damn, I hope this is not goodbye to uncensored models 😢
Anybody got a good list of models we should download right now before they start disappearing?
Pick up as many uncensored models as you can store ASAP. Those will absolutely be the first to go.
I would be less concerned about it being the end of uncensored models than I would be Nvidia finding ways to make sure there are no free models hosted there that work well on anything but Nvidia hardware for the foreseeable future.
Ok, how come these public utility databases are owned to be sold like that eg gitgub, twitter without any regulator pushback? Oh yea…
The Spark isn’t a bad computer, and they haven’t completely gotten rid of local GPUs, so Nvidia does somewhat care about local llms, but power corrupts and this is just more of it.
let the fuckening begin!
Fuck
The core llama.cpp maintainers also work at HF and will now work for Nvidia I guess. Llama.cpp is a pretty significant part of the local LLM stack, especially since other tools like Ollama and LMStudio are just GUIs built on top.
Local LLMs have gotten to the point where they are a serious threat to Anthropic and OpenAI, and Nvidia has a lot of skin in the game. If Nvidia wanted to do some serious damage to local LLMs, they are now in a position to do so.
I’m also imagining they may try to squeeze out support for other GPU vendors. I’m using an AMD 7900 XTX to run Qwen 3.8 27B that I downloaded from HF to run on llama.cpp, which currently works like a dream. The 7900 XTX is the only sanely priced 24GB GPU left in 2026 (under $1k vs. $2k, $3k, $4k for Nvidia 24-32GB cards). Combined with OpenCode or Pi, a setup like this basically eliminates the need to use Anthropic or OpenAI products in the same way Jellyfin eliminates the need to use streaming services.
I’m sure Nvidia and their buddies don’t like one thing I’ve said in this comment and may very well be plotting to put a stop to it, so the community may need to step up our game and get our eggs out of the big tech basket.
The core llama.cpp maintainers also work at HF and will now work for Nvidia I guess. Llama.cpp is a pretty significant part of the local LLM stack, especially since other tools like Ollama and LMStudio are just GUIs built on top.
I guess that explains why features like quantized KV caches are lagging behind. Maintainers are purposely dragging their feet.
the community may need to step up our game and get our eggs out of the big tech basket.
The community has chosen to not fight at all, which is worse. Anti-AI sentiment is at an all-time high.
Publicly. Privately, these hypocrites still whisper in ChatGPT’s ear when they get lazy enough. Or use some feature in Photoshop or some other software that they didn’t even understand was AI-driven.
guess that explains why features like quantized KV caches are lagging behind
Yeah, I was using the TheTom fork for a while and not sure why TQ KV cache hasn’t merged yet.
Anti-AI sentiment is at an all-time high
I suppose the tech bros have understandably soured a lot of people on LLMs with all of the negative societal costs due to their greedy and irresponsible bejavior. On the other hand, the concepts of the perceptron and artificial neural networks from the 40s and 50s are finally coming to fruition and we have these things now that are legitimately artificially intelligent that we can run on our gaming PCs. They are overhyped and used in ways they make no sense, but they’re also useful as long as their limitations are kept in mind. From a technology perspective, they’re cool as hell.
Oh no! -forks code- anyway…
do we have experts to work on it full time, paid?
Well the good news is they can’t take away from you what you already have. It being an open source project, I’m assuming if they do anything to deliberately gut AMD performance, it’ll get forked.
Also
The 7900 XTX is the only sanely priced 24GB GPU left in 2026 (under $1k vs. $2k, $3k, $4k for Nvidia 24-32GB cards).
Not on sale anymore, at least not at any vendor in my country, I searched an aggregate pricing website. Amazon has a few used ones left of some models, but that’s probably a 2 or 3 digit figure across SKUs. Hold on to yours with an iron grip.
What kind of tok/s are you getting with it on Qwen 3.8 27B and how’s the output quality? I may consider getting one if I can find one used or import from abroad.
I get around 40 token/s and it frequently has become reliable enough to drop sonnet for me. So take that as you will
Intel B50 and B60 pros are at microcenter right now perfect for this.
Agreed, I was referring more to future updates. Obviously we’re good with what is available now.
I can’t speak to pricing and availability outside the U.S., but it looks like the one I got went up $100:
https://www.newegg.com/asrock-radeon-rx7900xtx-24g-radeon-rx-7900-xtx-24gb-graphics-card-triple-fans/p/N82E16814930084.
I traded in my 3070 and my final price was in the 700s. Last I looked, used ones were going for $800 on eBay vs. $1200 for a used 3090.
I run 3.8 27B at q4 with q4 context up to 200k. Decode is generally in the 30s and pp starts in the 700s and drops to the 400s as context approaches 200k. I use mostly Sonnet 5 at work and I would rate this setup with the OpenCode desktop app as pretty comparable overall for coding at least. Let’s just say I have no reason to use any cloud models, not that I would do that voluntarily outside of being compelled to at work.
I got mine used, around 600 imperial credits. Look for ads that provide proof of working and benchmarks (like FurMark)
what if the LLAMA.CPP devs working at Nvidia improves CUDA support and keeps other vendor support.
It would be great but they seem to favour server computing as that has the greatest margins. I have a feeling being the biggest company on the planet at $5tn is not enough.
So they are just slapping random price tags on things now. It’s a database of AI questions. They have paid 13 billion dollars for a database, and everyone’s acting like that’s a perfectly rational sensible thing to do. No one in the financial industry has any sense anymore.
A company that’s only product is something that people either don’t want or actively hate has paid an eye-watering amount of money for a database to train their product on, this training will have no effect whatsoever on whether people want it.
I have this nice bridge with lots of training examples if anybody’s interested, 100 trillion dollars please
Yeah I don’t understand these valuations at all. I wish I did so that I could get a billion dollars for a website that doesn’t particularly own anything.
for a database to train their product on
Isn’t hugging face a database of already trained models?
Partially. But they also host datasets.
not everyone has a hate boner for AI
It’s an expensive toy that might become a viable product in a decade that doesn’t mean it’s a valuable company today. Irresponsible spending like this is exactly what led to the dot com bubble.
its a useful tool for me doing research.
Yeah because it was impossible to Google things in the past. Every time anybody comes up with a use for AI it’s basically just automating something that isn’t even that hard for you to do. If you want to automate simple tasks that’s absolutely fine but it’s not the second coming as people keep insisting.
Come back to me when an AI can actually add value to something, when it can do something that is not just automating a simple task, but it’s capable of doing things that humans are not. I keep being told that we’re only a few years away from the technological singularity, so wake me up when we actually get there.
What kind of research do llms actually help with?
In my experience, if the answer can’t be found in the top google results, the AI simply makes shit up.
The fact that it lies is so irritating, if it at least admitted that it didn’t know the answer it would at least be a useful search engine. The fact that it makes stuff up though means I always have to scroll past the AI summary, that I didn’t ask for, in order to actually go to the website to check the results myself, because I can’t trust it.
The worst part is that mostly lies for things that aren’t immediately obvious, so you ask it “what is the color of the sun” and it appears to work just fine, but the moment you use it for any real research, it will just fabricate things wholecloth.
thats completely false when you only ask the llm a very basic question like “what color is the sun?” you need to talk to the model like its smarter than a google search engine. “what color is the sun? pull the data from scientific articles and other sites that study the sun”
if you google the first question, you will have list of different pages, where if you use AI to search on your query, it will model an answer from thoese sources, with said sources linked, similar to a Wikipedia page.
There are thousands of papers on the color of the sun, its trivial to google a paper, if you need an LLM to do this, that’s on you. But if you ask it valve clearances for an indonesian motorbike in english, it will make it up, even though the service manual is available in Indonesian. If you ask it how to synthesize a chemical that nobody bothers to write about because its uninteresting or impossible, it will straight up lie.
i think that learning to prompt AI is somthing people lack as a skill and then call it bad and a liar like its some intelligence when it isn’t
Remember, the whole AI scheme exists to take the very concept of ownership from us. This takeover is hostile.
Why do we need a centralized hub for AI models anyway? I’ve never figured this out. We flock to these big “friendly” fucking companies with obviously unsustainable business models trying to own and sell things that should be shared in a decentralized, democratized mesh anyway. AI models should all be distributed magnet links, not hosted files. Why do we do this to ourselves?
The way things are going at Huggingface now, I imagine we probably will have to start building the infrastructure we need to handle AI models as distributed links sooner rather than later. Let the enshittification begin, we’ll move on to a different tactic for sharing AI models while cheerfully they squeeze cash out of people and businesses too lazy to adapt. Everything working as it should, I guess.
The way things are going at Huggingface now, I imagine we probably will have to start building the infrastructure we need to handle AI models as distributed links sooner rather than later.
Then “we” better get started. Mega corpos have fully embraced LLMs, to the point that they’ve distrusted human input too much, which is still a vital component of successful adoption. Eventually, more and more companies will wise up and figure out how to find the right balance.
Where are we at? Oh, right… we’re too busy arguing about how all AI is bad and sticking our fucking heads in the sand until the Big Bad AI Problem goes away.
We have to fix our attitudes if we have any hope of surviving this mess and not ending up as a Cyberpunk-wannabe dystopia by the time 2077 hits. Use the fucking weapons given to us!
For real, this is the perfect application for torrent. If I were building a registry that had to have sustainable cost for an open source community it would use torrents as a file transfer backend. If I were building something that I was planning to sell to a monopoly firm that has monopoly money… HTTP and cloud storage all thw way! 😄
There are worse things than copyrights not mattering anymore.
Forget about rights as a whole for us plebs.
To the one Downvoter I’m curious, why are you such a little bitch?
This comment got so much misintepreted lmao
Now there are 6 of them.
Probably cause it’s a dumb comment, you people see conspiracies everywhere.
It was only a matter of time
They should put up a back up torrent of all the currently available models before transferring ownership. Thus opening the doors for future websites.
NVIDIA is probably trying to shutdown the LLM-at-home market (which means you don’t need huge clouds of graphics cards or even an Internet connection). Sad day for people who like to control their data.
Nvidia also sells super expensive graphics cards. They probably want you to buy a couple of 5090s or a Spark.
They know they can’t shut down the entire self hosted LLM market but they can direct it towards using Nvidia GPUs by making sure llama.cpp devs don’t spend much time on ROCm support, etc. And they probably know that ClosedAI and Anthropic’s days on the frontier are limited.
I heard somewhere they are buying it because their own development platform sucks. So they are going to replace it with hugging face.
Which doesn’t preclude them from locking all the good stuff behind a subscription, of course.
Maybe a dev fee to post
Before any of these news I would use modelscope occasionally, it is the same thing as huggingface but Chinese and with a massive collection of mcp servers. Although they do not have the same amount of merged and fine tuned models, they are still a pretty good service
We just can’t have nice things can we.
We are in a war with the rich. Always have been. We will never have nice things unless you take it from them.
HuggingBay to the rescue 🏴☠️
You wouldn’t pirate a model
oh that’s actually a thing, neat
I fail to understand the use of the term open source when the resource itself can be bought by oligarchs. Like organic vegetables, or clean coal.
It was open source till the ones who has the keys (which shouldn’t be a thing for FOSS in the first place) got an offer that made their greed take over. It’s a thing that keeps happening and it needs to be a very strong lesson that anything open source needs to be handled in a way where no one can do this.
I’ve downloaded a terabyte worth of models and didn’t pay a cent. storage and bandwidth cost money
I just put together a budget starter PC for local inference. I guess this is my cue to set up and get downloading.
ELI5 huggingface?
Its a big box of toys but replace the toys with downloadable AI models and the box is a website
ngl the first image on their website is pretty much all the explanation you need
I am a LLM, what is this?
hosts all the ai models and is the primary resource for downloading them for pretty much all kinds, like depthmaps, masking too, also llms, inage generators, etc.
