No Priors Ep. 5 | With Huggingface’s Clem Delangue

No Priors Ep. 5 | With Huggingface’s Clem Delangue

No PriorsMay 19, 202337m

Elad Gil (host), Clem Delangue (guest), Sarah Guo (host), Elad Gil (host)

Clem Delangue’s personal journey into machine learning and early career at eBayHugging Face’s origin as a conversational ‘AI Tamagotchi’ and its strategic pivotBuilding and scaling an open-source ML community and model hubOpen-source vs proprietary AI models, specialization, and industry adoptionML model modalities, sizes, and emerging application domainsInfrastructure costs, decentralization, consent in datasets, and ethical AIHugging Face’s business model, GitHub analogy, and future directions in AI

In this episode of No Priors, featuring Elad Gil and Clem Delangue, No Priors Ep. 5 | With Huggingface’s Clem Delangue explores hugging Face’s Clem Delangue on Open-Source AI, Community, and Infrastructure Clem Delangue traces his path from top eBay seller to machine learning entrepreneur, explaining how a chance encounter at a trade show led him into computer vision and ultimately co-founding Hugging Face.

Hugging Face’s Clem Delangue on Open-Source AI, Community, and Infrastructure

Clem Delangue traces his path from top eBay seller to machine learning entrepreneur, explaining how a chance encounter at a trade show led him into computer vision and ultimately co-founding Hugging Face.

He recounts Hugging Face’s pivot from a playful ‘AI Tamagotchi’ chatbot to becoming the leading open-source hub for ML models, driven by organic traction around their Transformers library and a strong community-first culture.

Delangue discusses the balance between exploration and exploitation inside startups, the strategic importance of open-source in countering AI power concentration and bias, and the role Hugging Face plays as infrastructure and distribution layer for ML.

Looking ahead, he highlights infrastructure efficiency, decentralized training, consent-aware datasets, and full-stack ML companies—especially in biology and chemistry—as the most exciting frontiers.

Key Takeaways

Leave room for structured experimentation alongside core execution.

Delangue argues startups should consistently devote ~30–40% of effort to exploration, both before and after product-market fit, to avoid getting stuck in local optima and to surface new directions like Hugging Face’s pivot to open source.

Get the full analysis with uListen AI

Make community engagement everyone’s job, not a separate function.

Instead of hiring dedicated community managers, Hugging Face expects all employees—including researchers—to interact with users, tweet, and support open source, which creates authenticity and deeper feedback loops.

Get the full analysis with uListen AI

Follow organic pull from open-source usage to guide product strategy.

The company doubled down on Transformers and the model hub only after seeing explosive, organic adoption by researchers and enterprises, using that traction to justify a full strategic shift and to raise their Series A.

Get the full analysis with uListen AI

Expect a multi-model ecosystem, not a single dominant AI model.

With over 250,000 models from ~15,000 companies on Hugging Face, Delangue notes that specialized models often outperform giant general models on cost, latency, and accuracy for specific tasks, ensuring a diverse model landscape.

Get the full analysis with uListen AI

Invest early in infrastructure and cost-aware ML operations.

He emphasizes that the field has under-focused on the real cost and speed of running models and calls out infrastructure optimization as a major enabler, hinting that platforms like Hugging Face should become gateways to compute.

Get the full analysis with uListen AI

Build in the open to address power concentration and bias.

Projects like BigScience (BLOOM) and BigCode show how open collaboration and transparent training can democratize capabilities and allow affected communities to participate in surfacing and mitigating model biases.

Get the full analysis with uListen AI

Future opportunity lies in full‑stack ML-native companies, especially in bio/chem.

Delangue is particularly bullish on companies that both invent and productize ML (e. ...

Get the full analysis with uListen AI

Notable Quotes

It's funny how a single small encounter like that can completely change your trajectory.

Clem Delangue

We were like, 'Okay, we're gonna build some sort of an AI Tamagotchi.'

Clem Delangue

We make sure to spend at least like 30 or 40% of the company's efforts on explorating new things.

Clem Delangue

If there wouldn't have been as much open source as there's been in the past five years, we would be decades away from where we are now.

Clem Delangue

I don't really believe in this scenario of one model, one company to rule them all.

Clem Delangue

Questions Answered in This Episode

How can an early-stage startup practically implement a 30–40% exploration budget without jeopardizing runway?

Clem Delangue traces his path from top eBay seller to machine learning entrepreneur, explaining how a chance encounter at a trade show led him into computer vision and ultimately co-founding Hugging Face.

Get the full analysis with uListen AI

In a world of powerful proprietary models, what concrete governance or policy mechanisms could strengthen the open-source AI ecosystem?

He recounts Hugging Face’s pivot from a playful ‘AI Tamagotchi’ chatbot to becoming the leading open-source hub for ML models, driven by organic traction around their Transformers library and a strong community-first culture.

Get the full analysis with uListen AI

What are the most critical technical and social challenges to making consent-based, opt-in datasets the norm in ML training?

Delangue discusses the balance between exploration and exploitation inside startups, the strategic importance of open-source in countering AI power concentration and bias, and the role Hugging Face plays as infrastructure and distribution layer for ML.

Get the full analysis with uListen AI

How might decentralized or online learning architectures realistically work at scale compared to today’s large, infrequent training runs?

Looking ahead, he highlights infrastructure efficiency, decentralized training, consent-aware datasets, and full-stack ML companies—especially in biology and chemistry—as the most exciting frontiers.

Get the full analysis with uListen AI

For enterprises just starting with ML, how should they decide between adopting large general-purpose models and building smaller, specialized ones?

Get the full analysis with uListen AI

Transcript Preview

Elad Gil

(instrumental music) . Clem, welcome to the podcast.

Clem Delangue

Thanks for having me.

Elad Gil

Oh, thanks so much for joining us. So, um, we were hoping to start with your background, which I think is really interesting. You grew up in France, where you ran an electronics shop on eBay, and were just so prolific that you ended up earning an internship opportunity with eBay. How'd you go from that to image recognition and eventually to Hugging Face?

Clem Delangue

Yeah, it's actually quite a funny story, um, because I was one of the biggest French seller on, on eBay when I was working with them. I was kind of like the user-facing, uh, team member, and so they were sending me to all these, like, trade shows in France, uh, which, which were, like, the worst experiences ever, because at the time, PayPal belonged to eBay, and so we had a shared booth. And so all the PayPal users would come to the booth and basically shout at me, because PayPal was keeping their money or blocking their accounts o- or things like that. It was basically kind of like the worst days ever. Uh, but during one of, one of these days, I kind of like, uh, uh, bumped into a guy with, like, big round glasses, like looking very, very nerdy, and, um, I, I remember pretty vividly, he told me, like, "Oh, you guys are eBay. Uh, you acquired not so long ago a barcode scanning company, uh, called, uh, red, Red Laser, to recognize objects and be able to, you know, show listings to, to people. Uh, but it sucks. You guys, you need to know that's pretty soon with machine learn-" I mean, he wasn't calling it machine learning at the time, but "With, with these new algorithms, you won't even need the barcodes anymore. You'll just recognize the object itself." And at the time, I was like, "Who is this crazy guy?" Wasn't really, uh, paying attention too much. But at night, I kind of like did my research, um, and, and realized that he was a pretty legit guy coming out of a legit, uh, engineering, uh, school in France, uh, with his small startups, uh, which raised a, uh, a little bit of money. Um, and, and one thing after the other, uh, I ended up leaving eBay, uh, to join this startup doing, doing machine learning for computer vision, and that's, uh, kind of like I, I, I made the move to, to machine learning. It was almost 15 years ago now, uh, but I, I, I don't regret it at all. Uh, it's, it's funny how, like, a single small encounter like that can completely change your trajectory.

Elad Gil

That's really cool, yeah. I think some other time I'll tell you a bit a- a- about a meeting I had at a trade show that completely took my life in a different direction too, so I think it's kind of odd how sometimes those things happen early in peoples' careers. Um, could you tell us a bit more about the early iteration of Hugging Face, how you decided to start it, the early days as a talking emoji and sort of where, where it went from there?

Install uListen to search the full transcript and get AI-powered insights

Get Full Transcript

Get more from every podcast

AI summaries, searchable transcripts, and fact-checking. Free forever.

Add to Chrome