Will Everyone Become an AI Builder? Clem Delangue on Hugging Face, Agents, Local AI & Robotics

**Clément Delangue** (0:00)
The numbers of people who are going to be able to become AI builders is going to explode. If you would have a world where only a few companies could do software, it would be quite scary. You want open source to create more jobs? how am I gonna stay relevant?

**Ksenia Se** (0:22)
Thank you, Clem, for agreeing for this interview. I'm a big fan of Hugging Face and what you guys are doing for open source community. It's been amazing to know you for many years, but meet you for the first time in person.

**Clément Delangue** (0:33)
Yes, thanks for having me.

**Ksenia Se** (0:35)
Let's start with your recent post about ML Intern, how you were playing with it on Hugging Face. What is the most surprising and funny lesson that you learned how agents work with real motion-during tasks now?

**Clément Delangue** (0:50)
Yeah, so what's interesting is that I think the default coding agents are pretty bad at building AI. And you saw that, I think it's when Andres Carpati released, I think, auto-research, or I don't remember if it was auto-research or something before. He said, oh, I barely use any agents to build it, just because either it's like out of distribution or like it just doesn't work yet, really, to build AI. But with a couple of like tweaks to the harnesses, to the models, connection to tools like the Hugging Face Hub, you can actually make a lot of progress.
So we were surprised that now ML Intern is managing to fine-tune some small models, to create some datasets, to convert models to different formats.
Today, the team got it to pass the interview tests that they had for researchers. In half hour, it aces the tests. So we've been excited about it. If agents can lower the barrier to entry to build AI, it's going to be very valuable for the world because it's going to enable more people to do open source models, to do open datasets, to maybe play with local models, which historically have been a little bit hard to do, but I think now it's getting easier.

**Ksenia Se** (2:23)
How do you see this development in the coming months? Where the exploration is?

**Clément Delangue** (2:28)
I think the numbers of people who are going to be able to become AI builders is going to explode. It's going to go from maybe a few hundred thousands or low millions of people who have the skills to do this kind of work, to maybe tens of millions, fifties of millions, maybe a hundred million at some point. Maybe at some point, every software engineer will be able to optimize models themselves, train models themselves, and that obviously would be amazing because it would mean that they're not going to only rely on close source APIs and on third-party vendors who can dictate their conditions in a way, right? Increase their prices whenever they want, deprecate models whenever they want, change them behind the scenes that you're not even sure why the quality has gone down on your workloads, and so it gives back some controls to the builders, which is nice.

**Ksenia Se** (3:30)
A couple of months ago, I was doing a little interview with Steve Yage, and he said, oh, and non-technical people will definitely come to this coding world.
How do you feel about that? Are we ready?

**Clément Delangue** (3:44)
The beauty of AI is that a lot of it is driven by data sets and text in general, right? So compared to software engineering, where you had to kind of like learn programming language, I think AI has a potential to have a much wider base of users, of people who can contribute to it. So I hope it happens. I think it would be good too, because the more diversity of builders that you have, I think the wider the perspectives, and I think for the field, what is good is that it's going to drive it towards more actual challenges and things that are important for people.
I feel like if more people could build AI, maybe we would have a little bit less video AI slope, and maybe a little bit more of biology, chemistry, medicine, climate change, AI. Some things that a couple of Silicon Valley guys don't care so much about, but that other people care about. It brings more perspective and so pretty more problem solved.

**Ksenia Se** (4:59)
You think that more creation with AI will essentially get to some quality point, when people stop creating slope, because it's a lot of slope right now, and when they stop creating slope and start actually solving problems?

**Clément Delangue** (5:14)
I think there's a lot of other things in slope to build, and the more you enable and empower people to build, the more they're going to build other stuff than slope. Also, I think empowering more people to become AI builders will change the perspective of the public on AI. Obviously, right now, we have a terrible perception of AI by the public. If you look at the studies, it's crazy. Either people are crazy scared or they hate AI, or they don't want to hear about it.