François Chollet, an influential AI researcher, is launching a new startup that aims to build frontier AI systems with novel designs.
The startup, Ndea, will consist of an AI research and science lab, and seek to “develop and operationalize” AGI. AGI, which stands for “artificial general intelligence,” typically refers to AI that can perform any task a human can. It’s a goalpost for many AI companies, including OpenAI.
“We’re betting on a different path to build AI capable of true invention, adaptation, and innovation,” Chollet wrote in a series of posts on X. “We believe we have a small but real chance of achieving a breakthrough — creating AI that can learn at least as efficiently as people, and that can keep improving over time with no bottlenecks in sight.”
Ndea plans to use a technique called program synthesis, in tandem with other technical approaches, to unlock AGI. Chollet believes that program synthesis, which allows AI to generalize to problems it hasn’t seen before from only a few examples, can help to overcome the most intractable problems in AI research.
Program synthesis is traditionally computing-intensive. But Chollet believes this limitation can be overcome — and that overcoming it will help accelerate scientific progress.
“[W]e are not alone in recognizing the potential of program synthesis — it’s a technique every frontier AI lab is now starting to explore,” reads a blog post on Ndea’s website. “We are at the crest of a pivotal moment in scientific history and the world deserves every possible direct, unique attempt to build AGI.”
Ndea, which Chollet is co-founding with Zapier co-founder and head of AI Mike Knoop, hasn’t disclosed whether it has raised any capital from outside investors. But the company is currently hiring for remote research positions, suggesting that there’s at least some financial backing.
Knoop says that he’s stepping away from his day-to-day work at Zapier to make Ndea his focus, but that he’ll remain a member of Zapier’s board.
“We’re assembling the world’s top program synthesis team,” Knoop said in a post on X. “Our first focus is on deep learning-guided program synthesis to create AGI that can invent, adapt, and innovate … But even more exciting is the chance to metaphorically time-travel into the future: learn, invent, and discover things that would not organically happen for decades or even centuries.”
Chollet, who recently launched a nonprofit with Knoop to develop benchmarks for AGI, is the latest high-profile AI researcher to leave big tech to found an independent AI lab. Perhaps best known as the creator of Keras, a high-level, open source API that can be used to create AI models and tackle machine learning tasks, Chollet announced last November that he would depart Google after close to a decade there.
Ilya Sutskever, OpenAI’s former chief scientist and co-founder, last year founded Safe Superintelligence, an AI lab that has raised over $1 billion in capital from investors including Sequoia and Andreessen Horowitz. Elsewhere, longtime Google AI researcher and Stanford academic Fei-Fei Li is heading World Labs, a company developing AI systems that can generate video game-like 3D simulations.
