Hugging Face's Clem Delangue on Open Source AI and the LLM Bubble | MTS Live

Summary

Podcast Summary: The a16z Show — Hugging Face's Clem Delangue on Open Source AI and the LLM Bubble | MTS Live

Overview In this episode of The a16z Show, hosts Theo Jaffe and Sophia Puccini sit down with Clem Delangue, CEO of Hugging Face, for a candid discussion on open source AI, the landscape’s growing “LLM bubble,” debates over AI safety and regulation, global dynamics (particularly between the US and China), and how open source is shaping AI’s future both in software and robotics. The conversation offers insights from the cutting edge of AI infrastructure and addresses the philosophical and practical stakes in the open AI debate.

---

Open Source AI: Benefits, Risks, and the Global Landscape

US vs. China in Open Source AI

• Historical Leadership & Shifting Trends (02:30)
  • Clem Delangue: “Historically the US was super, super strong with open source… the T in ChatGPT is actually coming from Transformer, which was open source from Google. Unfortunately… things tended to close down in the US and… in contrast, China is the strongest open source contributor today.”
  • Notable open source contributions from China: startups and academia in the US often use models, libraries, and datasets developed by Chinese teams (DeepSeq, Quen, Kimi).

Safety and Releasing AI Models

• Balancing Progress and Risk (05:12, 07:48)
  • Delangue argues that fears of open source AI capabilities (like with large language models) are consistently overblown. Referring to the release of powerful models (“Mythos” as an example), he notes how initial fears subside as communities adapt.
  • “Proof point is that, you know, mythos… was crazy dangerous and now it’s starting to be deployed kind of like everywhere… probably in a few weeks or a few months, everyone is going to be using Mythos and not kind of like destroy the world as a result.”
• Analogy on Regulation (08:09)
  • “The idea of restricting a technology like AI based on risks is just like… you would say… let’s tie down everybody’s hands, right, because it’s too dangerous. But in reality, you don’t want to do that because your hands are so useful… The way you want to control it is untie everyone, give the freedom to everyone, and then regulate or fight the bad actors.”

Defender vs. Attacker Dynamics

• Open sourcing models helps not just potential “attackers” but also “defenders” (security experts, good actors) (06:45).

---

The “LLM Bubble” and Industry Sustainability

Is There an AI/LLM Bubble? (03:44)

• Delangue’s Assessment:
  • Delangue distinguishes between AI as a whole (not in a bubble) versus a possible “LLM bubble.”
  • “If there’s one specific domain of AI where there’s maybe a risk of over investing, it’s large language models distributed behind APIs. You see the building of crazy data centers… a lot of revenue growth with uncertain margins and long-term sustainability.”
• The conversation suggests a need to watch for sustainability challenges and the risk of hype outpacing real economic or social value.

---

International Regulation & Policy

US-China AI Regulation Discussions (08:59)

• Clem’s view on high-level international talks:
  • “I hope both countries will be able to agree on fostering more transparency, more openness to… help more people access this technology.”
  • Notes the importance of international collaboration and mentions personalities like Nvidia’s Jensen Huang contributing to dialogue.

---

Robotics: The Next Frontier for Open Source AI

Hugging Face’s “Le Robot” Initiative (10:09)

• Robotics Unlocking New AI Use Cases:
  • Hugging Face has shipped nearly 10,000 Ricci Nini robots worldwide (“probably one of the most widely distributed robots of the year at this point”).
  • Open platform, extensive app store (300+ apps), and use cases from helping kids interact with AI beyond the laptop, to practical household tasks like cooking.
  • Clem emphasizes that robotics gives people “empowering” and “new use cases,” highlighting the potential social impact.
• US vs. China Robotics Leadership:
  • “It’s very likely that Chinese are going to dominate robotics, or at least they’re already dominating… it’s important that we build more in the US… I hope to see a lot more in the coming months on the topic.”

---

Why Hugging Face Won as the “GitHub of AI”

Differences Between Hosting Code vs. AI Artifacts (12:44)

• Hosting and sharing AI models and data is fundamentally different from sharing code.
  • “For us, the volume of files is much, much larger than what GitHub is doing… last week we added 2 petabytes of data… the equivalent of 500,000 2-hour movies.”
• Technical and infrastructural differences made Hugging Face a better fit for the needs of the AI open source community.
• GitHub focused elsewhere, primarily on code and coding assistants, whereas Hugging Face tailored infrastructure to AI’s unique demands.

---

Notable Quotes & Memorable Moments

• Balancing Safety and Progress (08:09)
  • “The way you want to control it is untie everyone, give the freedom to everyone, and then regulate or fight the bad actors.” — Clem Delangue
• On Open Source AI’s Global Shift (02:30)
  • “Today, if you ask most startups, most academia in the US that are using open source, they're usually using Chinese open source models.” — Clem Delangue
• The LLM Bubble (03:44)
  • “If there’s a bubble it’s probably an LLM bubble, but we’ll see what happens in the next few months.” — Clem Delangue
• Robotics Applications (10:41)
  • “When you see the Ricci Nini on a kitchen table, looking around and helping you cook, you realize that it’s enabling, empowering, creating new use cases that are just not possible with a laptop or phone.” — Clem Delangue

---

Key Timestamps

• 00:00 — Opening analogy on tech regulation (“tie down everybody’s hands”)
• 02:30 — US vs. China in open source AI
• 03:44 — The “LLM bubble” discussion
• 05:12 — Should labs restrict open source for safety?
• 07:48 — Regulation analogy expanded
• 08:59 — International agreements and hopes for US-China detente
• 10:09 — Robotics as the next open source frontier
• 12:44 — Why Hugging Face succeeded as “GitHub for AI”

---

Conclusion This episode dives deep into the ideological and practical challenges of open source AI, with Clem Delangue offering a strong defense of openness as a competitive and societal catalyst. The conversation touches on the geopolitical dynamics of AI, the risks and realities of “LLM bubble” talk, and the potential for AI-driven robotics to change how we interact with technology. Hugging Face’s unique community success is examined in light of technical and cultural needs that large generalist platforms couldn’t meet. The episode provides an up-to-date, global perspective on where open source AI is heading and why it matters.

Transcript

[Clement DeLonge]

The idea of restricting a technology like AI based on risks is just like, for example, you would say, okay, some people can punch other people, so let's tie down everybody's hands, right? Because it's too dangerous. Some people can punch, right, but in reality you don't want to do that because your hands are so useful. The way you want to control it is untie everyone and then regulate or fight the bad actors. So for example, if hacking that creates cybersecurity risks, it's illegal, right? So you have to fight it, but not by preventing everyone from getting these capabilities. Otherwise you slow down progress, you create massive gaps in terms of controls, in terms of capabilities, and you create actually additional risks.

[Podcast Host (A16Z)]

This episode originally aired on mts. Open source software built much of the modern Internet. Linux, Apache, Kubernetes and even the transformer architecture behind ChatGPT all spread because researchers and developers could study, modify and improve them in public. But AI is increasingly moving in the opposite direction, with the most powerful models distributed behind closed APIs controlled by a small number of companies. At the same time, China has emerged as one of the biggest contributors to open source AI, while debates around safety, regulation and access are becoming more politically charged. And now those same tensions are extending into robotics, where AI is beginning to move off the screen and into the physical world. Theo Jaffe and Sophia Puccini speak with Clem DeLonge, CEO at Hugging Face.

[Theo Jaffe]

We are live here on mts with Clement DeLong, who is the CEO of Hugging Face, which has been really an incredible resource for anyone who's interested in large language models and especially open weight large language models. I've been a Hugging Face user for a while now, so it's great to have you here. Clem, thanks so much for coming on mts.

[Clement DeLonge]

Yeah, of course, thanks for having me.

[Theo Jaffe]

Absolutely.

[Sophia Puccini]

Okay, so you are a big proponent of open source. First of all, how do you predict and you believe that open source is like a very important, you know, thing for innovation and competition? So can you compare and contrast sort of like the open source environments in the US and China to start?

[Clement DeLonge]

Yeah. So I mean, historically the US was super, super strong with open source, right? That's kind of like what, what led to, to the current AI revolution, right? Like the, the T in ChatGPT is actually coming from Transformer, which was open source from, from Google. Unfortunately, for the past past few years this trend has, has changed and, and things tended to kind of like close down in, in the US and kind of like Frontier Labs more kind of like sharing their models behind closed source APIs, the China so the complete opposite movement, they're the strongest open source contributors today. If you ask most startups, most academia in the US that are using open source, they're usually using Chinese open source models. Right. You've probably heard of Deep Seq, of Quen, of Kimi. There are kind of like a bunch of companies and organizations in China contributing massively to the field of open source.

[Sophia Puccini]

Great. So you recently said we're in an LLM bubble. What makes you think that?

[Clement DeLonge]

Well, I was asked if we were in an AI bubble and I said we're probably not in AI as kind of like a general field bubble. But I feel like if there's one specific domain of AI where there's so much investment that there's maybe a risk of over investing, it's large language models distributed behind APIs. Right. Like you see the building of crazy data centers for it and obviously you see a lot of revenue growth but with kind of like uncertain margins and certain kind of like long term sustainability and moat for it. So if there's a bubble it's probably an LLM, but we'll see what happens in the next few months.

[Theo Jaffe]

Well, you're a big proponent of open source, you know, as we all know. But do you think that labs should ever restrict releasing their models in an open source way for safety reasons? Like yeah, in 2022, 23, it was way too early for that. The models at the time were toys, but now we have stuff like Claude Mythos which supposedly can really assist people with cyber attacks. We have models that are increasing pretty dramatically in bio capability which could be even scarier. So do you think companies should still be releasing their models open source?

[Clement DeLonge]

So the interesting thing is that we've had these conversations and this kind of like talking point for a while in AI when we were early at the game face, I think six, seven years ago at the time it was GPT2 and there was already a lot of people saying that it was too dangerous to release in open source. At the time it was six, seven years ago when basically it was nothing more than just, just an auto complete. I think we've seen progressively that these were quite overblown and I think they're also overblown today. Right. And proof point is that, you know, mitos, I think when it was announced was it like three weeks ago, a month ago it was crazy dangerous and now it's, it's starting to be deployed kind of like everywhere. Right. I think they just gave access to the first international organization it's in South Korea, I think, yesterday or something like that. And probably in a few weeks or in a few months, everyone is going to be using metos and not kind of like destroy the world as a result. So I think with the current models, it's safe to release behind APIs, it's safe to release in open source, and it's actually the safest way because it gives everyone the capabilities to not only build the systems, but also build the protection systems. So if we talk, for example, for cybersecurity, the biggest risk is that a few players have capabilities that other people don't have. And so the attackers could have capabilities that the defenders have. Whereas kind of like if you make it more open, actually, it's usually easier for the defenders to react and kind of make the whole system safer. So that's kind of like what we see with each releases, where there are always kind of like overgrown concerns before and then progressively just we all adapt and the benefits kind of like outweighs the risks.

[Sophia Puccini]

Yeah, it feels like we'll still be dealing with this problem in like 50 years where somebody releases like some sort of like open source robotics, you know, robot or program or something, and then everyone is like, no, you shouldn't have done that. It's so risky. And then we'll just adapt again.

[Clement DeLonge]

It's kind of like the story of technology, you know, Like, I mean, the idea of like restricting a technology like AI based on risks is just like, for example, you would, you would say, okay, some people can punch other people, so let's tie down everybody's hands. Right, because it's too dangerous. Some people can punch.

[Theo Jaffe]

Right.

[Clement DeLonge]

But in reality, you don't want to do that because your hands are so useful. They're creating so many good things in the world. You need your hands. The way you want to control it is untie everyone, give the freedom to everyone, and then regulate or fight the bad actors. Right? So, for example, if you know hacking, that creates cybersecurity risks, I mean, it's illegal, right? So you have to make it illegal, you have to fight it, but not by preventing everyone from getting these capabilities, because otherwise you slow down progress, you create massive gaps in terms of controls, in terms of capabilities, and you create actually additional risks.

[Theo Jaffe]

Well, right now on the topic of regulation, President Trump is in China where he will be meeting with Xi Jinping over the next couple of days, and they're going to be discussing, among other things, AI regulation and international AI agreements. So what do you hope to get out of this in Terms of open source.

[Clement DeLonge]

Yeah, I mean, I'm excited to see conversations about open source AI. Probably there's going to be some conversations about distillation, about collaborations between two countries. I hope both countries will be able to agree on fostering more transparency, more openness to kind of like help more people access, access this technology. I'm glad that Jensen hopped into, into, into the plane and joined these conversations because I think he has a lot of the right perspectives on, on the, this topic to kind of like basically create more collaboration between, between countries and kind of like share progress.

[Sophia Puccini]

Yeah. I'm curious about your robotics push. So you guys launched Le Robot in 2024 and you've talked about how robotics is the next frontier unlocked by AI and all of this stuff. How do you sort of see this playing out and what is the role of open source?

[Clement DeLonge]

Yes, I have two little robots behind me, two Riccinini. We've shipped almost 10,000 of them all over the world. So it's probably one of the most widely distributed robots of the year at this point. I think what's really cool with robotics is that enables kind of like very new use cases and better use cases for AI. So for example, for the Ricci Nini, you have an app store. Anyone can build apps. So there's been over 300 apps that have been created for it already. And when you see it in action, for example, with kids, empowering kids to interact with AI in a different way than looking at a laptop or looking at a phone, you realize that it's very empowering. When you see the Richie Mini on a kitchen table, looking around and helping you cook, you realize that it's enabling, empowering, creating new use cases that are just not possible just with a laptop and a phone. Right. That's why OpenAI and Sam Altman, for example, have talked a lot about their excitement about bringing new devices to market. There's an important China US component there because it's very likely that Chinese are going to dominate robotics, or at least they're already dominating. And so on this topic too, it's really important that we build more in the US on this topic and we obviously have a lot of strength for it. With the strength of the startup EcoSystem in the US, the strength of the frontier models. I hope to see a lot more in the coming months of the topic.

[Theo Jaffe]

Hugging Face has been compared to GitHub a lot, the GitHub of AI. But why wasn't GitHub the GitHub of AI? It seems like they've kind of fumbled a lot of things in the AI realm. So why do you think Hugging Face became sort of the go to place for model developers to deploy models and not GitHub?

[Clement DeLonge]

Yeah, I mean, I don't blame them. They have a lot of on their plates. Right. Like I think with the coding assistant they've kind of like been dealing with their own set of issues. The reality is that hosting and sharing AI artifacts is quite different than hosting code. So even if people have been calling us the GitHub of AI, I think it's two very different things. For example, for us, the volume of files of data that we're dealing with is much, much larger than what the GitHub is doing. For example, just last week we added 2 petabytes of data to the platform just last week. It's kind of like a matter of comparison. It's the equivalent of 500,000 2 hours movies that have been uploaded to Hugging Face just last week. So you have a lot of structural differences. And we managed to build our infrastructure capabilities in a way that makes it just better for people that are building in AI to use Hugging Face to host their models, their data sets, both publicly but also privately. We have a lot of private usage now, so that's kind of some of the reasons why we managed to do it. Whereas GitHub focused on other things.

[Theo Jaffe]

Totally. Well, that's pretty, pretty cool. We love Hugging Face and we really appreciate your early support of MTs and our drops. So it was great to have you on today. Clem, thanks so much for coming on MTS.

[Podcast Host (A16Z)]

Thanks for listening to this episode of the A16Z podcast. If you liked this episode, be sure to like, comment, subscribe, leave us a rating or review and share it with your friends and family. For more episodes, go to YouTube, Apple Podcasts and Spotify. Follow us on X1.6Z and subscribe to our substack@A16Z substack.com thanks again for listening and I'll see you in the next episode. This information is for educational purposes only and is not a recommendation to buy, hold or sell any investment or financial product. This podcast has been produced by a third party and may include paid promotional advertisements, other company references, and individuals unaffiliated with A16 zone. Such advertisements, companies and individuals are not endorsed by AH Capital Management, LLC, A16Z or any of its affiliates. Information is from sources deemed reliable on the date of publication, but A16Z does not guarantee its accuracy.