Chief Scientist: Yann LeCun, Tapestry Kickoff, Paris
📅 May 20, 2026FunctionalTV14 MIN68 VIEWS15 SEGMENTS · 2 SPEAKERS
Yann LeCun discusses Tapestry, an open source AI initiative aimed at giving countries, regions, institutions, and communities a way to collaboratively train frontier models without pooling private data.
In this conversation with Alexy Khrabrov in Paris, LeCun explains why AI sovereignty will require open source collaboration, federated training, diverse local data, and broad participation from governments, universities, companies, and individual contributors. He also clarifies his position on LLMs: they are clearly useful for coding, math, and information access, but he does not see them as a...
What Yann LeCun said
Written from the verified transcript and checked against it. Every figure links to the moment it was said.
Yann LeCun, former Meta chief AI scientist and now at Amina's, described the Tapestry project, an open-source platform for collaboratively training a frontier AI model. He said the idea originated at Meta but was not pursued. Tapestry aims to let countries and regions retain data sovereignty while exchanging only parameter vectors to train a consensus model. LeCun noted strong demand from countries like Switzerland, Germany, France, UAE, Vietnam, India, Japan, and Korea for AI sovereignty, and argued they should pool resources rather than build separate LLMs. He said Meta is becoming less open, creating a role for Tapestry. He emphasized the mission of preserving democracy and cultural diversity, and highlighted technical challenges in distributed training, possibly on commodity hardware. He acknowledged the techniques are not yet proven at scale and welcomed PhD students to contribute.
Key takeaways
Tapestry is an open-source platform for collaboratively training a frontier AI model without exchanging data, only parameter vectors.
LeCun left Meta in early 2026 and is now at Amina's, but Tapestry is a separate non-commercial project.
Countries like Switzerland, Germany, France, UAE, Vietnam, India, Japan, and Korea are building their own LLMs, which LeCun calls a waste of resources.
LeCun said LLMs are useful but not a path to human-level intelligence or AGI.
The mission of Tapestry is to protect democracy and cultural diversity by ensuring diverse AI assistants.
Numbers and commitments
Figure
What it refers to
Type
At
12 years ago
When LeCun wrote a white paper on distributed training
All right. Hello everybody. I'm Alexy from the founding committee and head of AI Alliance and now head of committee at Lexel. And here we are at FPT in Paris, France with Yann LeCun who is a chief scientific advisor to the new Tapestry project of the AI Alliance. Please tell us what the Tapestry is, how it came about, and how you became basically scientific lead.
So Tapestry project is an old idea I had that I was originally trying to push within Meta, but it didn't go very far. It's the idea that there's going to be a future where every one of us is going to get our entire information diet through AI assistants. And even though the AI assistants are produced by a handful of companies on the west coast of the US or in China, we're in trouble because we're not going to have enough diversity of sources of information. So what I'm proposing is an open source platform which would constitute a foundation model on top of which to train AI systems, and those AI systems would be highly diverse. That would provide people with a high diversity of AI assistants with their language, their culture, their value system, their centers of interest, their political biases, whatever they want. And so the best way to do this is to have a coalition of multiple countries or regions or contributors collaborating in some sort of federation to train a frontier model. And they don't have to exchange data. They can retain sovereignty on their data. The way they exchange are parameter vectors. And the system is trained collaboratively. Everybody will have access to it. And each country or region may use their own local data and their local languages, the local library material, everything. So, I think it's pretty much the only way that we're going to get access to a wide diversity of AI systems in the future.
Right. And so, basically, you've been the head of AI at Meta and you led research at Meta and Llama was a big success, right? And so, essentially, it's not enough for the world, right? Because we don't know exactly the data on which it was trained, right? So, we've been here for a two-day workshop kicking off Tapestry. So, we've actually heard representatives of multiple countries, from multiple labs, and so there are kind of some of the architectural ideas. How would you summarize kind of the current state of the idea? How should the world go about building these models, scientific and open source AI community?
Right. So, first of all, I need to make it clear that I'm not at Meta anymore. I left Meta early 2026. I'm involved in a new company called Amina's, but this effort is separate. It's really a collaborative, non-commercial open project. So, what I've seen in the past is a very strong desire from a lot of countries around the world, countries that are neither the US nor China, to have some level of sovereignty in AI. And I see a lot of efforts in countries like Switzerland, Germany, France, UAE, Vietnam, India, Japan, Korea, efforts to build their own LLM. And it's a bit of a waste of resources, right? All of those countries should get together and train a common LLM. I see some countries also that in the past have built on top of Llama. But it's not clear what role Meta is going to play in the open source world in the future. Because they're kind of becoming much less open. So I think there is an important role to play for an organization like this one. Where this is the kickoff meeting, we hope to rally a lot of talented young people. We want this to be a bottom-up project where people who think they can contribute technically can contribute technically without having to go through major bureaucracy or anything. But I think we're going to get a lot of support from governments around the world. Either money or political support or all kinds of resources in general.
Right. Because there is so much of a demand. Right. You mentioned that, so AI Alliance held several meetings at the AI Action Summits in Paris and New Delhi. And so you mentioned that basically those governments turned to you for advice. The heads of states were interested in this specifically.
Right, so governments are understanding that this is important. And so there is demand. There is demand around the world and so this initiative is actually put together to answer the actual demand that people around the world have. That's right. There is a huge demand and a lot of countries realize that their only path to AI sovereignty is through open source and collaboration. Right. So that gives a big boost for projects like this one. But I think what's important also is, and pardon me to get a little technical, but what we want is to build a software infrastructure so that various contributors can train their model locally on their own data, preserve the sovereignty on their data, they don't have to exchange their data, the only information that circulates between the contributors are parameter vectors for models. And we would train those models in a distributed fashion but in such a way that they eventually arrive at a consensus model that is as good as if it had been trained on the totality of all the data accessible to everyone. And that's an important aspect because that's probably a way to get those open models to get better than the proprietary models. They will have access to more data, more diverse data as well, using regional data that the commercial entities don't have access to. So this is a technical architectural challenge for the model builders, so this is not something that yet exists. No, it's still, we still have to figure out. We know the techniques will exist, we know that they kind of work, they are prototypes. It's feasible, we know it's feasible, but the details have to be worked out.
Yes. You know what I like about this, that you are known to proclaim the opinions that all LLMs, that new scientists, new researchers should not go in them. And your company is building a world model, but here you are advising a very practical today endeavor. And you actually said that these things are useful, there is a place for them, there is a demand for them, and you actually propose a way to make them better and make them useful for the world. So it's probably, it could be a PhD for somebody who will figure out how to do this.
8 more exchanges in this transcript
Sign in free to read the rest of this interview. No card required.
