site banner
PaperclipPerfector  2d ago (text post)   2561 thread views

Small-Scale Question Sunday for September 6, 2026

Do you have a dumb question that you're kind of embarrassed to ask in the main thread? Is there something you're just not sure about?

This is your opportunity to ask questions. No question too simple or too silly.

Culture war topics are accepted, and proposals for a better intro post are appreciated.

2
Top Bottom Old Controversial Comments
Jump in the discussion.

No email address required.

What exactly is going on with Hugging Face? Seems they have been bought by Nvidia?

I dunno, something feels strange but I don't know what. First the AI attacks and now this, and reading up on their background on Wikipedia just has the sound of something that was a bit dodgy. Maybe it's me. But I don't see how you go from "we're creating a chatbot aimed at teenagers" to "robotics! AI! Nvidia!" and managing to get linked with/mentioned alongside AWS, Meta, etc. along the way.

I'm way too ignorant to have an opinion, so that's why I'm asking the knowledgeable is there anything there, or have they successfully managed to generate hype and ride the wave all the way to a very fat buy-out (before the inevitable bubble bursts/crash/AI takes over the world and paperclips us all)?

They host the vast majority of the world's model weights. If you are doing work on AI, you almost certainly download the initial weights from HF because it's extremely fast and free (takes special infrastructure). Nvidia is taking control of a big part of the AI ecosystem.

There's a big question of how you monetize this stuff without enshittifying it -- right now, HuggingFace's basically just selling storage and download, a space with historically not great margins, and the obvious market option of selling model inference runs into content moderation processing problems. Probably a commoditize your complement strategy.

But yeah, if you want a hand on open weights for a lot of AI-focused stuff, HuggingFace is the standard, right now, whether you're looking at LLMs, audio processing, or a lot of vision input work. ImageGen has gravitated to other areas, but even then a lot of people also upload to HF.

That said, what's the moat for HF, at least in this usage? I mean, if Amazon or Google (or even any other second-tier provider) wanted to compete with them on model storage, what would be the obstacles they'd need to overcome? Storage and delivery of massive byte arrays - even at a very large scale - seems like a problem that a lot of companies can successfully solve. There's an incumbent advantage, but it can be overcome relatively quick.

NVidia's moat is Cuda and having a large number of models in use around helps support that moat since any competitors will have to implement enough of Cuda to support most of them instead of higher level layer that only works for major closed AI providers. It's not so much that HF has much moat (which they only do by being the default in most software) but that ensuring HF's existence NVidia helps support their own moat.

Oh, I know what Nvidia's moat is, that's not in question. HF though seems to not have much. I mean, I am sure Nvidia guys have good case for it, but I doubt it gives Nvidia a lot of control over AI ecosystem. Anytime any major player would feel restrained by whatever HF is doing, they can implement a clone very easily. And most software would trivially adapt (as easy as changing an URL prefix, likely).

At the moment having HF in your code means you can try a new model almost just by swapping out the string in your model load function.

I don't think it's a moat in the technical sense. You do need the infra and I suspect it's not that easy, but not that hard either. I think it's just a coordination point, like owning Twitter or Github. If you're subtle, you can put a thumb on the scale here and there: mildly encourage A, mildly discourage B. NVIDIA got their position largely by making early AI research far easier on CUDA libraries than on any other GPU, so forward-thinking acquisition of coordination points isn't out of character for them.

They do offer an inference platform with API, I believe they also standardized testing APIs, training, etc... While not impossible to beat, the moat is that tooling is already written to hook up to their APIs.

Ironically, AI being used to rewrite tools that hook up to HF is likely the biggest threat.

What is this place?

This website is a place for people who want to move past shady thinking and test their ideas in a court of people who don't all share the same biases. Our goal is to optimize for light, not heat; this is a group effort, and all commentators are asked to do their part.

The weekly Culture War threads host the most controversial topics and are the most visible aspect of The Motte. However, many other topics are appropriate here. We encourage people to post anything related to science, politics, or philosophy; if in doubt, post!

Check out The Vault for an archive of old quality posts. You are encouraged to crosspost these elsewhere.


Why are you called The Motte?

A motte is a stone keep on a raised earthwork common in early medieval fortifications. More pertinently, it's an element in a rhetorical move called a "Motte-and-Bailey", originally identified by philosopher Nicholas Shackel. It describes the tendency in discourse for people to move from a controversial but high value claim to a defensible but less exciting one upon any resistance to the former. He likens this to the medieval fortification, where a desirable land (the bailey) is abandoned when in danger for the more easily defended motte. In Shackel's words, "The Motte represents the defensible but undesired propositions to which one retreats when hard pressed."

On The Motte, always attempt to remain inside your defensible territory, even if you are not being pressed.


New post guidelines

If you're posting something that isn't related to the culture war, we encourage you to post a thread for it. A submission statement is highly appreciated, but isn't necessary for text posts or links to largely-text posts such as blogs or news articles; if we're unsure of the value of your post, we might remove it until you add a submission statement. A submission statement is required for non-text sources (videos, podcasts, images).

Culture war posts go in the culture war thread; all links must either include a submission statement or significant commentary. Bare links without those will be removed.

If in doubt, please post it!


Rules


Recommended Posts And Communities

Recommended Realtime Chats