Clem Delangue: The Ultimate Guide to Investing in AI; Elon's Threat to Sue OpenAI | E1013

Clem Delangue: The Ultimate Guide to Investing in AI; Elon's Threat to Sue OpenAI | E1013

The Twenty Minute VCMay 12, 202358m

Clément Delangue (guest), Harry Stebbings (host), Narrator, Narrator

Origin and evolution of Hugging Face from AI companion app to AI platformOpen-source vs proprietary AI models and the “one model vs many models” debateCentralized AI hubs (Silicon Valley) vs globally distributed AI talentBusiness model, monetization, and network effects at Hugging FaceLegal and regulatory issues in AI (training data, content access, regulation timing)Startup funding dynamics, investor selection, and the true role of VCsChallenges of building AI-first companies: hiring, compute costs, and moats

In this episode of The Twenty Minute VC, featuring Clément Delangue and Harry Stebbings, Clem Delangue: The Ultimate Guide to Investing in AI; Elon's Threat to Sue OpenAI | E1013 explores clem Delangue Explains Open AI Ecosystems, Startup Moats, And Fundraising Reality Clem Delangue, CEO of Hugging Face, discusses the evolution of his company from a consumer ‘Tamagotchi AI’ app to a leading open-source AI platform, and why openness has driven most AI progress. He contrasts two visions of the AI future: one ultra-centralized model (e.g., OpenAI) versus a world of many specialized, often open-source models owned and tuned by individual companies. Delangue argues that long-term advantage comes from building in-house AI capabilities rather than relying solely on black-box APIs, even if the latter are easier at first. He also shares candid views on startup fundraising, the role of VCs, regulation, and why founders must focus on enjoying the act of building rather than expecting things to get easier at later stages.

Clem Delangue Explains Open AI Ecosystems, Startup Moats, And Fundraising Reality

Clem Delangue, CEO of Hugging Face, discusses the evolution of his company from a consumer ‘Tamagotchi AI’ app to a leading open-source AI platform, and why openness has driven most AI progress. He contrasts two visions of the AI future: one ultra-centralized model (e.g., OpenAI) versus a world of many specialized, often open-source models owned and tuned by individual companies. Delangue argues that long-term advantage comes from building in-house AI capabilities rather than relying solely on black-box APIs, even if the latter are easier at first. He also shares candid views on startup fundraising, the role of VCs, regulation, and why founders must focus on enjoying the act of building rather than expecting things to get easier at later stages.

Key Takeaways

Long-term differentiation in AI requires owning your models, not just calling APIs.

Using a single third-party model via API is like using Wix for your website—great for speed, but weak for deep optimization, cost control, and defensible product differentiation over time.

Get the full analysis with uListen AI

All serious companies will eventually run their own specialized AI models.

Delangue predicts that, analogous to custom codebases, every meaningful company will maintain models tuned to its data, constraints, and use cases rather than relying solely on a generic ‘one-size-fits-all’ model.

Get the full analysis with uListen AI

Open science and open source are the real engine of recent AI progress.

Breakthroughs like transformers, BERT, and diffusion models came from openly shared research and code; without this openness, Delangue estimates we’d be decades behind current capabilities.

Get the full analysis with uListen AI

Founders should optimize for where they are happiest, not just for the Valley.

He rejects the dogma that AI founders must move to Silicon Valley, arguing that great companies can be built from anywhere as long as founders are energized and can visit hubs when needed.

Get the full analysis with uListen AI

AI-first startups face their biggest constraint in talent, not only in compute.

There are very few people globally who’ve actually built and trained state-of-the-art architectures; competition for these hybrid scientist–engineer profiles is intense and expensive.

Get the full analysis with uListen AI

Usage is ‘delayed revenue’ for network-effect AI platforms.

Hugging Face prioritizes adoption and community usage over near-term revenue maximization, assuming that being the default platform where AI gets built will naturally unlock large monetization later.

Get the full analysis with uListen AI

Company building never truly gets easier; the difficulty just changes form.

Delangue emphasizes that each stage has its own challenges, so founders should build a company and working life they enjoy now, instead of suffering today in the hope that a later funding round will make everything simple.

Get the full analysis with uListen AI

Notable Quotes

Each stage has a lot of challenges. It doesn’t get easier, it just gets different.

Clem Delangue

Most of the progress we’re seeing in AI is based on open science and open source.

Clem Delangue

Using an AI API is like using Wix for your website—it feels great at the beginning, but you can’t really optimize or differentiate with it.

Clem Delangue

All companies will have their own AI models—basically their own GPT-4.

Clem Delangue

My assumption is that usage is delayed revenue, especially in a domain like AI.

Clem Delangue

Questions Answered in This Episode

If you’re an early-stage startup, when is the right moment to transition from API-based AI to training or heavily fine-tuning your own models?

Clem Delangue, CEO of Hugging Face, discusses the evolution of his company from a consumer ‘Tamagotchi AI’ app to a leading open-source AI platform, and why openness has driven most AI progress. ...

Get the full analysis with uListen AI

How can smaller companies practically start building in-house AI capabilities without access to top-1% ML talent and massive compute budgets?

Get the full analysis with uListen AI

What governance or compensation models might fairly align content creators and AI model trainers as data licensing becomes more contentious?

Get the full analysis with uListen AI

In a world where ‘all companies have their own models,’ what new kinds of infrastructure, tooling, or standards will be most critical to avoid fragmentation?

Get the full analysis with uListen AI

How should founders balance the push for rapid AI adoption inside enterprises with the need to address bias, misuse, and regulatory uncertainty from day one?

Get the full analysis with uListen AI

Transcript Preview

Clément Delangue

When you, like, an early stage founder struggling and you're like, "Oh, but that's gonna be good. I'm gonna struggle. It's gonna be really hard. But in one year, in two years when I'm gonna be bigger, it's gonna be easier," I'm sorry, but it, it won't be easier. Uh, the truth is that each stage has a lot of challenges. And so I think something important is for entrepreneurs to realize that and enjoy what they're doing now, and try to find and build a company that they enjoy building. Not the joy of, you know, getting to a series D, getting a series C, getting to IPO, but really the joy of, of building, right?

Harry Stebbings

(instrumental music) Clem, I am very excited for this. As we just said, I've stalked the shit out of you from Lee Fix or Pat Grady, Ollie at Datadog, Dev at Mongo, Thibault Elzzier who told me about the very early days. So thank you so much for joining me today.

Clément Delangue

Thanks so much for having me. I'm excited about that.

Harry Stebbings

This will be great. So I wanna start with, uh, a little bit of context. Hugging Face, where did the name come from and what's the origin of the company founding? In a short two to three minutes.

Clément Delangue

Yeah. When we started Hugging Face, we joked with my co-founders, Julien and Thomas, that we wanted to be the first company to go public with an emoji rather than the three-letter ticker, you know? We felt like the three-letter ticker, like, on the NASDAQ and, and all that is, is kind of boring. Felt like it was, was time for a refresh and to finally have emojis up there on the boards. So we absolutely wanted an emoji as a name, and the choice is the Hugging Face emoji, right? The one, the one with hands like that, was our favorite emoji, so we were like, "Okay, let's do that." We thought maybe we would keep it for a few weeks, you know, like for a few months at, at most. And then the community started to put it everywhere, you know, like on social media, on their clothes, like, literally everywhere. So we were like, "Oh, maybe we're gonna keep it." And now it's become kind of, like, such, such a brand, so popular that unfortunately it's gonna be hard for us to, to change it.

Harry Stebbings

Listen, at least when you do go public, the ticker will be an emoji. So I mean, you know, uh, all you need to do is, you know, get to that stage. Uh, well, in terms of company founding, like, why did you decide this was the idea that you wanted to spend 10 years, 20 years of your life on?

Clément Delangue

We actually started with something completely different. Um, the reality is that the company was, was formed because of some sort of, you know, professional crush between me and my co-founders where we were like, "We absolutely wanna work together." Um, plus our excitement about AI, right? It was seven years ago, uh, so it was not of use as it is now. Um, not enough people were talking about it at the time, but we were super excited about it as kind of, like, a new paradigm to, to build technology, new opportunities and, and all of that. The first company, the first startup I worked for, like, 15 years ago was already doing AI. We w- weren't calling it AI at, at the time, so I had some sort of a glimpse of the, of the capabilities. And when we started Hugging Face, when we started the company, we were like, "Okay, uh, what can we work on that is going to be both scientifically challenging..." Because one of our co-founders, uh, Thomas, is, is a scientist, uh, and we, we all have a lot of interest in, in the science side of things. "... but at the same time entertaining?" So we actually started with some sort of a Tamagotchi AI or, like, AI friend, howev- however you call it, some sort of, like, a Siri, Alexa or, like, ChatGPT, uh, except just for, like, entertainment, not for, like, the boring, like, productivity aspect of it. Um, and we actually did that for, for almost three years. We raised our pre-seed and seed rounds on, on this idea. We got a product out with a couple of, uh, billion messages exchanged between users and, and this Tamagotchi AI. Uh, but as, you know, it sometimes, sometimes happens, um, when we started sharing a little bit of the underlying technology and the underlying platform that we were building to, to do that, we saw a lot of traction from the community, from the open source community, from companies using that. And so that basically made us pivot from this Tamagotchi AI to this AI platform that we are now.

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