Announcement We are happy to announce that ggml.ai (the founding team of llama.cpp) are joining Hugging Face in order to keep future AI truly open. Georgi and team are joining HF with the goal of s...
Read the original article
Comments
I consider HuggingFace more "Open AI" than OpenAI - one of the few quiet heroes (along with Chinese OSS) helping bring on-premise AI to the masses.
I'm old enough to remember when traffic was expensive, so I've no idea how they've managed to offer free hosting for so many models. Hopefully it's backed by a sustainable business model, as the ecosystem would be meaningfully worse without them.
We still need good value hardware to run Kimi/GLM in-house, but at least we've got the weights and distribution sorted.
By data-ottawa 2026-02-2014:465 reply Can we toss in the work unsloth does too as an unsung hero?
They provide excellent documentation and they’re often very quick to get high quality quants up in major formats. They’re a very trustworthy brand.
Yeah, they're the good guys. I suspect the open source work is mostly advertisements for them to sell consulting and services to enterprises. Otherwise, the work they do doesn't make sense to offer for free.
By danielhanchen 2026-02-2023:031 reply Haha for now our primary goal is to expand the market for local AI and educate people on how to do RL, fine-tuning and running quants :)
By WanderPanda 2026-02-213:473 reply Amazing work and people should really appreciate that the opportunity costs of your work are immense (given the hype).
On another note: I'm a bit paranoid about quantization. I know people are not good at discerning model quality at these levels of "intelligence" anymore, I don't think a vibe check really catches the nuances. How hard would it be to systematically evaluate the different quantizations? E.g. on the Aider benchmark that you used in the past?
I was recently trying Qwen 3 Coder Next and there are benchmark numbers in your article but they seem to be for the official checkpoint, not the quantized ones. But it is not even really clear (and chatbots confuse them for benchmarks of the quantized versions btw.)
I think systematic/automated benchmarks would really bring the whole effort to the next level. Basically something like the bar chart from the Dynamic Quantization 2.0 article but always updated with all kinds of recent models.
By danielhanchen 2026-02-2112:08 Thanks! Yes we actually did think about that - it can get quite expensive sadly - perplexity benchmarks over short context lengths with small datasets are doable, but it's not an accurate measure sadly. We're actually investigating currently what would be the best efficient course of action on evaluating quants - will keep you posted!
> How hard would it be to systematically evaluate the different quantizations? E.g. on the Aider benchmark that you used in the past?
Very hard. $$$
The benchmarks are not cheap to run. It'll cost a lot to run them for each quant of each model.
By danielhanchen 2026-02-2112:081 reply Yes sadly very expensive :( Maybe a select few quants could happen - we're still figuring out what is the most economical and most efficient way to benchmark!
By illusive4080 2026-02-2112:551 reply Roughly how much does it cost to run one of the popular benchmarks? Are we talking $1,000, $10,000, or $100k?
By danielhanchen 2026-02-229:58 Oh it's more time that's the issue - each benchmark takes 1-3 hours ish to run on 8 GPUs, so running on all quants per model release can be quite painful.
Assume AWS spot say $20/hr B200 for 8 GPUs, then $20 ish per quant, so assuming benchmark is on BF16, 8bit, 6, 5, 4, 3, 2 bits then 7 ish tests so $140 per model ish to $420 ish/hr. Time wise 7 hours to 1 day ish.
We could run them after a model release which might work as well.
This is also on 1 benchmark.
By danielhanchen 2026-02-2112:09 Working on it! :)
By arcanemachiner 2026-02-2017:58 I hope that is exactly what is happening. It benefits them, and it benefits us.
not that unsung! we've given them our biggest workshop spot every single year we've been able to and will do until they are tired of us https://www.youtube.com/@aiDotEngineer/search?query=unsloth
By danielhanchen 2026-02-2112:09 Appreciate it immensely haha :) Never tired - always excited and pumped for this year!
By danielhanchen 2026-02-2023:01 Oh thank you - appreciate it :)
By danielhanchen 2026-02-2023:01 Thank you!
It's insane how much traffic HF must be pushing out of the door. I routinely download models that are hundreds of gigabytes in size from them. A fantastic service to the sovererign AI community.
My fear is that these large "AI" companies will lobby to have these open source options removed or banned, growing concern. I'm not sure how else to explain how much I enjoy using what HF provides, I religiously browse their site for new and exciting models to try.
ModelScope is the Chinese equivalent of Hugging Face and a good back up. All the open models are Chinese anyways
By thot_experiment 2026-02-2022:263 reply Not true! Mistral is really really good, but I agree that there isn't a single decent open model from the USA.
Mistral is cool and I wish them success but it consistently ranks extremely low on benchmarks while still being expensive. Chinese models like DeepSeek might rank almost as low as Mistral but they are significantly cheaper. And Kimi is the best of both worlds with incredible benchmark results while still being incredibly cheap
I know things change rapidly so I'm not counting them out quite yet but I don't see them as a serious contender currently
By thot_experiment 2026-02-213:14 Sure, benchmarks are fake and I use Mistral over equivalently sized models most of the time because it's better in real life. It runs plenty fast for me, I don't pay for inference.
By BoredomIsFun 2026-02-219:401 reply > it consistently ranks extremely low on benchmarks
As general purpose chatbots small Mistral models are better than comparably sized Chiniese models, as they have better SimpleQA scores and general knowledge of Western culture.
By seanmcdirmid 2026-02-2115:181 reply It’s really hard to beat qwen coder, especially for role play where the instruction following is really useful. I don’t think their corpus is lacking in western knowledge, although I wonder if Chinese users get even better results from it?
By BoredomIsFun 2026-02-2115:341 reply > It’s really hard to beat qwen coder, for role play
I am not sure if you actually tried that. Mistrals are widely asccepted go-to models for roleplay and creative writing. No Qwens are good at prose, except for their latest big Qwen 3.5.
> I don’t think their corpus is lacking in western knowledge,
It absolutely does, especially pop culture knowledge.
By seanmcdirmid 2026-02-2116:371 reply Instruct and coder just follow instructions so well though. I guess I’ve just never been able to make mistral work well, I guess.
By BoredomIsFun 2026-02-2117:441 reply Qwen3 30B A3B and that big 400+ B Coder were absolutely terrible at editing fiction. I would tell them what to change in the prose and they'd just regurgitate text with no changes.
By seanmcdirmid 2026-02-221:031 reply Did you try asking Gemini what model to use and how to configure/set it up? It has worked wonders for me, ironically (since I’m using a big model to setup smaller local models).
By BoredomIsFun 2026-02-227:041 reply > Did you try asking Gemini what model to use and how to configure/set it up?
That would besuboptimal, as Gemini has too old knowledge cutoff. I am long past the need for such an advice anyway, as I've been using local models since mid 2024.
By seanmcdirmid 2026-02-2211:341 reply Gemini will search the web for most things (at least if you are using it via the web search interface), it isn’t limited to the knowledge it was trained on. Actually, I’m a bit mortified that not everyone knows this. If you ask Gemini (from the search interface) about a current event that happened yesterday, they will use search to pull in context and work with that. Also about model that was released yesterday, it can do that.
It’s only a very low level model access where search isn’t used. Local models also need to be configured to use search, and I haven't had a use case to do that yet.
Gemini seems to call this “grounding with google search”. If you have Gemini installed in your enterprise, it will also search internal data sources for context.
By BoredomIsFun 2026-02-2219:221 reply > Gemini will search the web for most things (at least if you are using it via the web search interface), it isn’t limited to the knowledge it was trained on.
If decides to do so, and even then baked in knowledge would influence the result.
In any case I do not need Gemini or any other LLMs to figure out setting for my llama.cpp, thank you very much.
By seanmcdirmid 2026-02-2220:26 It has always searched the web for me, and it can give me pretty good guidance about a model released in the last week. All models ATM are trying to reduce dependence on internal knowledge mostly through RAG. Anyways, this part of LLMs has gotten much better in the last 6 months.
If you are able to figure out the right settings for a model Thats was released last week, then great for you! But it sounds like you just don’t trust LLMs to use current knowledge, and have some misconception about how they satisfy recent knowledge requests.
Why are you talking price when we are talking local AI?
That doesn't make any sense to me. Am I missing something?
By dirasieb 2026-02-218:00 15 missed calls from your local power company
By seanmcdirmid 2026-02-2115:18 Apple silicon is crazy efficient as well as being comparable to GPUs in performance for max and ultra chips.
By cpburns2009 2026-02-2114:121 reply If you have the hardware to run expensive models, is the cost of electricity much of a factor? According to Google, the average price in the Silicon Valley Area is $0.448 per kWh. An RTX 5090 costs about $4,000 and has a peak power consumption of 1000 W. Maxing out that GPU for a whole year would cost $3,925 at that rate. It's not particularly more expensive than that hardware itself.
By culi 2026-02-2118:15 At that point it'd be cheaper to get an expensive subscription to a cloud platform AI product. I understand the case for local LLMs but it seems silly to worry about pricing for cloud-based offerings but not worry about pricing for locally run models. Especially since running it locally can often be more expensive
By thot_experiment 2026-02-2120:21 for almost the entire year, yes.
By ac29 2026-02-2218:13 Arcee is working on that, see a blog post about their newest in progress model here: https://www.arcee.ai/blog/trinity-large
Its still not fully post trained and its a non-reasoning model, but its worth keeping an eye on if you dont want to use the Chinese models that currently are the best open-weight options.
By CamperBob2 2026-02-214:33 To be fair there are lots of worse models than OpenAI's GPT-OSS-120b. It's not a standout when positioned next to the latest releases from China, but prior to the current wave it was considered one of the stronger local models you can reasonably run.
By throwaway27448 2026-02-2022:521 reply They can try. I don't think they'll be able to get the toothpaste back in the tube. The data will just move our of the country.
By seanmcdirmid 2026-02-2115:14 Many of the models on hugging face are already Chinese. It’s kind of obvious that local AI is going to flourish more in China than the USA due to hardware constraints.
By dotancohen 2026-02-211:06 How do you choose which models to try for which workflows? Do you have objective tests that you run, or do you just get a feel for them while using them in your daily workflow?
By toofy 2026-02-2112:15 it’s only a matter of time. we have all seen first hand how … wrong … these companies behave, almost on a regular basis.
there’s a small tinfoil hat part of me that suspects part of their obscene investments and cornering the hardware market is driven by an conscious attempt to stop open source local from taking off. they want it all, the money, the control, and to be the only source of information to us.
By Onavo 2026-02-2019:31 Bandwidth is not that expensive. The Big 3 clouds just want to milk customers via egress. Look at Hetzner or CloudFlare R2 if you want to get get an idea of commodity bandwidth costs.
Yup, I have downloaded probably a terabyte in the last week, especially with the Step 3.5 model being released and Minimax quants. I wonder what my ISP thinks. I hope they don't cut me off. They gave me a fast lane, they better let me use it, lol
By fc417fc802 2026-02-2020:011 reply Even fairly restrictive data caps are in the range of 6 Tb per month. P2P at a mere 100 Mb works out to 1 TiB per 24 hours.
Hypothetically my ISP will sell me unmetered 10 Gb service but I wonder if they would actually make good on their word ...
By 3eb7988a1663 2026-02-211:401 reply I have a 1.2TB cap before you start getting charged extra, so you might need to recalibrate your restrictive level.
By fc417fc802 2026-02-212:083 reply Is that with a WISP by chance? Or in a developing country? Or are there really wired providers with such low caps in the western world in this day and age?
ATT once told me if I don't pay for their TV service then my home gigabit fiber would have a 1TB cap. They had an agreement with the apartment building so I had no other choice of provider.
By fc417fc802 2026-02-215:411 reply Buy our off brand netflix or else we'll make it so you can't watch netflix. How is that legal?
By Zetaphor 2026-02-2117:45 The law is written by the highest bidder, and the telecom lobbyists are very generous
By nagaiaida 2026-02-213:58 well it's my wired cap a stone's throw from buildings with google cloud logos on the side in a major us city, so...
By zargon 2026-02-214:27 Comcast.
> We still need good value hardware to run Kimi/GLM in-house
If you stream weights in from SSD storage and freely use swap to extend your KV cache it will be really slow (multiple seconds per token!) but run on basically anything. And that's still really good for stuff that can be computed overnight, perhaps even by batching many requests simultaneously. It gets progressively better as you add more compute, of course.
> it will be really slow (multiple seconds per token!)
This is fun for proving that it can be done, but that's 100X slower than hosted models and 1000X slower than GPT-Codex-Spark.
That's like going from real time conversation to e-mailing someone who only checks their inbox twice a day if you're lucky.
By zozbot234 2026-02-2112:10 You'd need real rack-scale/datacenter infrastructure to properly match the hosted models that are keeping everything in fast VRAM at all times, and then you only get reasonable utilization on that by serving requests from many users. The ~100X slower tier is totally okay for experimentation and non-conversational use cases (including some that are more agentic-like!), and you'd reach ~10X (quite usable for conversation) by running something like a good homelab.
At a certain point the energy starts to cost more than renting some GPUs.
Yeah, that is hard to argue with because I just go to OpenRouter and play around with a lot of models before I decide which ones I like. But there's something special about running it locally in your basement
By dotancohen 2026-02-211:09 I'd love to hear more about this. How do you decide that you like a model? For which use cases?
By fc417fc802 2026-02-2020:04 Aren't decent GPU boxes in excess of $5 per hour? At $0.20 per kWhr (which is on the high side in the US) running a 1 kW workstation 24/7 would work out to the same price as 1 hour of GPU time.
The issue you'll actually run into is that most residential housing isn't wired for more than ~2kW per room.
Why doesn't HF support BitTorrent? I know about hf-torrent and hf_transfer, but those aren't nearly as accessible as a link in the web UI.
By embedding-shape 2026-02-2016:025 reply > Why doesn't HF support BitTorrent?
Harder to track downloads then. Only when clients hit the tracker would they be able to get download states, and forget about private repositories or the "gated" ones that Meta/Facebook does for their "open" models.
Still, if vanity metrics wasn't so important, it'd be a great option. I've even thought of creating my own torrent mirror of HF to provide as a public service, as eventually access to models will be restricted, and it would be nice to be prepared for that moment a bit better.
I thought of the tracking and gate questions, too, when I vibed up an HF torrent service a few nights ago. (Super annoying BTW to have to download the files just to hash the parts, especially when webseeds exist.) Model owners could disable or gate torrents the same way they gate the models, and HF could still measure traffic by .torrent downloads and magnet clicks.
It's a bit like any legalization question -- the black market exists anyway, so a regulatory framework could bring at least some of it into the sunlight.
By embedding-shape 2026-02-2016:192 reply > Model owners could disable or gate torrents the same way they gate the models, and HF could still measure traffic by .torrent downloads and magnet clicks.
But that'll only stop a small part, anyone could share the infohash and if you're using the dht/magnet without .torrent files or clicks on a website, no one can count those downloads unless they too scrape the dht for peers who are reporting they've completed the download.
By fc417fc802 2026-02-2020:11 > unless they too scrape the dht for peers who are reporting they've completed the download.
Which can be falsified. Head over to your favorite tracker and sort by completed downloads to see what I mean.
By sowbug 2026-02-2016:25 Right, but that's already happening today. That's the black-market point.
By Barbing 2026-02-2119:20 That would be a very nice service. I think folks might rely on it for a number of reasons, including that we'll want to see how biases changed over time. What got sloppier, shillier...
Wouldn’t it still provide massive benefits if they could convince/coerce their most popular downloaded models to move to torrenting?
By intrasight 2026-02-2111:48 Benefit to you, but great downside to the three letter agencies that inject their goods into these models.
By homarp 2026-02-2018:00 how are all the private trackers tracking ratios?
By taminka 2026-02-2017:52 most of the traffic is probably from open weights, just seed those, host private ones as is
I still don't know why they are not running on torrent. Its the perfect use case.
By heliumtera 2026-02-2016:04 How can you be the man in the middle in a truly P2P environment?
By freedomben 2026-02-2016:051 reply That would shut out most people working for big corp, which is probably a huge percentage of the user base. It's dumb, but that's just the way corp IT is (no torrenting allowed).
It's a sensible option, even when not everyone can really use it. Linux distros are routinely transfered via torrent, so why not other massive, open-licensed data?
By freedomben 2026-02-2016:162 reply Oh as an option, yeah I agree it makes a ton of sense. I just would expect a very, very small percentage of people to use the torrent over the direct download. With Linux distros, the vast majority of downloads still come from standard web servers. When I download distro images I opt for torrents, but very few people do the same
By Const-me 2026-02-2021:03 > very small percentage of people to use the torrent over the direct download
BitTorrent protocol is IMO better for downloading large files. When I want to download something which exceeds couple GB, and I see two links direct download and BitTorrent, I always click on the torrent.
On paper, HTTP supports range requests to resume partial downloads. IME, it seems modern web browsers neglected to implement it properly. They won’t resume after browser is reopened, or the computer is restarted. Command-line HTTP clients like wget are more reliable, however many web servers these days require some session cookies or one-time query string tokens, and it’s hard to pass that stuff from browser to command-line.
I live in Montenegro, CDN connectivity is not great here. Only a few of them like steam and GOG saturate my 300 megabit/sec download link. Others are much slower, e.g. windows updates download at about 100 megabit/sec. BitTorrent protocol almost always delivers the 300 megabit/sec bandwidth.
By zrm 2026-02-2016:49 With Linux distros they typically put the web link right on the main page and have a torrent available if you go look for it, because they want you to try their distro more than they want to save some bandwidth.
Suppose HF did the opposite because the bandwidth saved is more and they're not as concerned you might download a different model from someone else.
By thot_experiment 2026-02-2022:30 I have terabytes of linux isos I got via torrents, many such cases!
It's hard to overstate the impact Georgi Gerganov and llama.cpp have had on the local model space. He pretty much kicked off the revolution in March 2023, making LLaMA work on consumer laptops.
Here's that README from March 10th 2023 https://github.com/ggml-org/llama.cpp/blob/775328064e69db1eb...
> The main goal is to run the model using 4-bit quantization on a MacBook. [...] This was hacked in an evening - I have no idea if it works correctly.
Hugging Face have been a great open source steward of Transformers, I'm optimistic the same will be true for GGML.
I wrote a bit about this here: https://simonwillison.net/2026/Feb/20/ggmlai-joins-hugging-f...
By carbocation 2026-02-2017:591 reply Because many of us think simonw has discerning taste on this topic and like to read what he has to say about it, so we upvote his comments.
i don't doubt this. i just find it questionable that one particular poster always gets in the spotlight when AI is the topic - while other conversations in my opinion offer more interesting angles.
Upvote the conversations that you find to be more interesting. If enough people do the same, they too will make it to the top.
By coldtea 2026-02-211:04 Parent implies there might be some "boosting" involved, in which case, "upvote the conversations that you find to be more interesting" wont change anything...
Not saying this is the case, but it's what the comment implies, so "just upvote your faves" doesn't really address it.
By colesantiago 2026-02-2019:36 Agreed,
I would like to see others, being promoted to the top rather than Simon’s constant shilling for backlinks to his blog every time an AI topic is on the front page.
At a guess that's because my comment attracted more up-votes than the other top-level comments in the thread.
I generally try to include something in a comment that's not information already under discussion - in this case that was the link and quote from the original README.
By seanhunter 2026-02-2019:461 reply It’s at the top because of upvotes. They don’t have an “if simonw: boost” branch in the code.
the code is not public, so we can't know. i think it's much more nuanced and certain users' comments might get a preferential treatment, based on factors other than the upvote count - which itself is hidden from us.
By ComplexSystems 2026-02-2020:19 > the code is not public, so we can't know.
I feel like you're making this statement in bad faith, rather than honestly believing the developers of the forum software here have built in a clause to pin simonw's comments to the top.
By satvikpendem 2026-02-2020:491 reply > certain users' comments might get a preferential treatment
This does not happen. It hasn't even happened when pg made the forum in the first place.
By dcrazy 2026-02-2021:01 I thought dang explicitly said it does happen? It certainly happens for stories.
By ontouchstart 2026-02-2019:06 Attention feeds attention.
Attention is ALL You Need.
HN goes through phases. I remember when patio11 was the star of the hour on here. At another time it was that security guy (can't remember his name).
And for those who think it's just organic with all of the upvotes, HN absolutely does have a +/- comment bias for users, and it does automatically feature certain people and suppress others.
> And for those who think it's just organic with all of the upvotes, HN absolutely does have a bias for authors, and it does automatically feature certain people and suppress others.
Exactly.
There are configurable settings for each account, which might be automatically or manually set—I'm not sure–, that control the initial position of a comment in threads, and how long it stays there. There might be a reward system, where comments from high-karma accounts are prioritized over others, and accounts with "strikes", e.g. direct warnings from moderators, are penalized.
The difference in upvotes that account ultimately receives, and thus the impact on the discussion, is quite stark. The more visible a comment is, i.e. the more at the top it is, the more upvotes it can collect, which in turn makes it stay at the top, and so on.
It's safe to assume that certain accounts, such as those of YC staff, mods, or alumni, or tech celebrities like simonw, are given the highest priority.
I've noticed this on my own account. Before being warned for an IMO bullshit reason, my comments started to appear near the middle, and quickly float down to the bottom, whereas before they would usually be at the top for a few minutes. The quality of what I say hasn't changed, though the account's standing, and certainly the community itself, has.
I don't mind, nor particularly care about an arbitrary number. This is a proprietary platform run by a VC firm. It would be silly to expect that they've cracked the code of online discourse, or that their goal is to keep it balanced. The discussions here are better on average than elsewhere because of the community, although that also has been declining over the years.
I still find it jarring that most people would vote on a comment depending on if they agree with it or not, instead of engaging with it intellectually, which often pushes interesting comments to the bottom. This is an unsolved problem here, as much as it is on other platforms.
By Eisenstein 2026-02-2116:202 reply There is a saying that if everyone you encounter seems to be unreasonable, maybe it isn't the other people that are being unreasonable.
This isn't to say that social media is fair, or that people vote properly or that any ranking system based on agreement by readers is a good one. However, generally when you are getting negativity communicated to you and you are seeing consistently poor results around actions you take, it is going to be useful to examine the possibility that there is a difference in how you perceive what you are doing vs how others do. In that case spending time trying to figure out ways in which you are being wronged so that you can continue in the same manner is going to be time wasted.
You seem to be assuming that everything is organic and above board on here. That it's all just user/community stimuli, and if someone flies high well clearly it's great content, from which we can infer the reverse as well.
We don't have the source for HN, nor do we have the obvious bias metadata that the moderators have put in place, but simply paying attention betrays that manipulation mechanisms exist and are heavily utilized.
For instance I clearly have a "bad guy" flag on my account, and frequently see my highly rated comments sorted below literally greyed out comments. Comments older than mine, so it isn't just the normal "well newer comments get a boost", it's just that there is a comment "DEI" in place where some people get a freebie boost and some people get a freebie detriment. It's why often mediocre content and comments by the core group is always floating high.
And let me make it very clear that I do not care. I don't harbour any delusions about some tight community or the like, and HN is not important in my life or my ego. I also know that it's basically a propaganda network for YC (I mean...it's right in the URL), and good for them. It's their site and they can do anything they want with it.
I only commented because some people really think this place is a meritocracy+democracy. That isn't how it works, even if they really want people to think that.
By Eisenstein 2026-02-225:29 No one is under the assumption that any social media space is going to be meritocratic or democratic. The assumption is that some percentage of users are manipulating it and the backend and admins are doing the same. It is an attention economy. I don't think anyone is naive about this. My comment was merely a take on the 'the video game controller is broken' excuse that everyone has when they need to cover for their ego. Sometimes the controller is broken, but it almost never is.
How are you getting persecution complex from what I said? If anything, your comment might be feeding that delusion. :)
My point is that HN definitely has certain weights associated with accounts, which control the karma, visibility, and ultimately discussion of certain topics.
This problem doesn't affect only negativity or downvotes, but upvotes as well. The most upvoted comments are not necessarily of the highest quality, or contribute the most to the discussion. They just happen to be the most visible, and to generally align with the feeling of the hive mind.
I know this because some of my own comments have been at the top, without being anything special, while others I think are, barely get any attention. I certainly examine my thinking whenever it strongly aligns with the hive mind, as this community does not particularly align with my values.
I also tend to seek out comments near the bottom of threads, and have dead comments enabled, precisely to counteract this flawed system. I often find quality opinions there, so I suggest everyone do the same as well.
An essential feature of a healthy and interesting discussion forum is to accomodate different viewpoints. That starts by not burying those that disagree with the majority, or boosting those that agree. AFAIK no online system has gotten this right yet.
By rymc 2026-02-2019:36 the security you mean is probably tptacek (https://news.ycombinator.com/user?id=tptacek)
By throwaway2027 2026-02-2019:51 Time flies and simonw his AI feedback isn't always received favorably, sometimes he pushes it too much.
By satvikpendem 2026-02-2020:48 They aren't pinned, people just vote on them, and more so because simonw is a recognizable name with lots of posts and comments.
By magicalhippo 2026-02-2113:41 New comments get a boost, and as such are frequently near the top just due to that. Frequent upvotes also boosts. There might be other factors.
However these things are dynamic and change over time. As I read the discussion just now, the GP comment was the ~5th top-level comment.
By francispauli 2026-02-2020:23 thanks for reminding me i need to follow his blog weekly again
I'm regularly amazed that HuggingFace is able to make money. It does so much good for the world.
How solid is its business model? Is it long-term viable? Will they ever "sell out"?
By microsoftedging 2026-02-2015:571 reply FT had a solid piece a few weeks back: "Why AI start-up Hugging Face turned down a $500mn Nvidia deal"
https://giftarticle.ft.com/giftarticle/actions/redeem/9b4eca...
sounds very interesting, but even though it says giftarticle.ft, I got blocked by a paywall.
By nerevarthelame 2026-02-2016:321 reply To summarize, they rejected Nvidia's offer because they didn't want one outsized investor who could sway decisions. And "the company was also able to turn down Nvidia due to its stable finances. Hugging Face operates a 'freemium' business model. Three per cent of customers, usually large corporations, pay for additional features such as more storage space and the ability to set up private repositories."
By bee_rider 2026-02-2016:48 Freemium seems to be working pretty well for them—what’s the alternative website, after all. They seem to command their niche.
By culi 2026-02-2019:20 find the Bypass Paywalls Clean extension. Never worry about a paywall again
They have paid hosting - https://huggingface.co/enterprise and paid accounts. Also consulting services. Seems like a pretty good foundation to me.
By julien_c 2026-02-2016:12 and a lot of traction on paid (private in particular) storage these days; sneak peek at new landing page: https://huggingface.co/storage
Their business model is essentially the same as GitHub. Host lots of stuff for free and build a community around it, sell the upscaled/private version to businesses. They are already profitable.
By HanClinto 2026-02-2017:39 This is what Sourceforge did too, and they still had the DevShare adware thing didn't they?
GitHub is great -- huge fan. To some degree they "sold out" to Microsoft and things could have gone more south, but thankfully Microsoft has ruled them with a very kind hand, and overall I'm extremely happy with the way they've handled it.
I guess I always retain a bit of skepticism with such things, and the long-term viability and goodness of such things never feels totally sure.
By heliumtera 2026-02-2016:091 reply >Will they ever "sell out"?
Oh no, never. Don't worry, the usual investors are very well known for fighting for user autonomy (AMD, Nvidia, Intel,IBM, Qualcomm)
They are all very pro consumers and all backers are certainly here for your enjoyment only
These are all big hardware firms, which makes a lot of sense as a classic 'commoditize the complement' play. Not exactly pro-consumer, but not quite anti-consumer either!
By smallerize 2026-02-2019:22 heliumtera is being sarcastic.
By RussianCow 2026-02-2020:432 reply > By law they are required to do whatever they can to maximize profits.
I know it's a nit-pick, but I hate that this always gets brought up when it's not actually true. Public corporations face pressure from investors to maximize returns, sure, but there is no law stating that they have to maximize profits at all costs. Public companies can (and often do) act against the interest of immediate profits for some other gain. The only real leverage that investors have is the board's ability to fire executives, but that assumes that they have the necessary votes to do so. As a counter-example, Mark Zuckerberg still controls the majority of voting power at Meta, so he can effectively do whatever he wants with the company without major consequence (assuming you don't consider stock price fluctuations "major").
But I say this not to take away from your broader point, which I agree with: the short-term profit-maximizing culture is indeed the default when it comes to publicly traded corporations. It just isn't something inherent in being publicly traded, and in the inverse, private companies often have the same kind of culture, so that's not a silver bullet either.
By chucksmash 2026-02-2112:30 It's a worthwhile point to make because if people believe that misconception then it lets companies wash their hands of flagrantly bad behavior. "Gosh, we should really get around to changing the law that makes them act that way."
By 5o1ecist 2026-02-210:35 [flagged]
By HanClinto 2026-02-2017:56 Great points.
Valve is one of my top favorite companies right now. Love the work they're doing, and their products are amazing.
Can hardly wait for the Steam Frame.
By I_am_tiberius 2026-02-2014:242 reply I once tried hugging face because I wanted I worked through some tutorial. They wanted my credit card details during the registration as far as I remember. After a month they invoiced me some amount of money and I had no idea what it was. To be honest, I don't understand what exactly they do and what services I was paying for, but I cancelled my account and never touched it again. For me that was a totally intransparent process.
By shafyy 2026-02-2014:27 Their pricing seems pretty transparent: https://huggingface.co/pricing
By in-silico 2026-02-220:34 Sounds like a personal skill issue
