François Chollet

AI researcher, creator of the Keras library and ARC-AGI benchmark, and founder of Endia.

Mentions

Appeared in

Discussed in

Key positions and views

Scaling the current generation of deep learning models is an inefficient and insufficient path to AGI, as it primarily enhances pattern matching rather than true reasoning or fluid intelligence.

True general intelligence must be measured by a system's ability to efficiently learn and become competent at any novel task, a principle embodied in the ARC-AGI benchmark.

AGI is fundamentally an algorithmic problem that likely has an elegant, compact solution (e.g., <10,000 lines of code), rather than one requiring incomprehensible scale.

Program synthesis offers a more promising foundation for future AI systems than parametric deep learning, promising greater data efficiency and generalizability.

Current LLM-based technology can fully automate any problem domain that has formally verifiable solutions and a trusted reward signal, such as mathematics and coding.

Podcast consensus on Chollet

Points of consensus

▶François Chollet is the creator of the ARC-AGI benchmark, which is designed to measure a form of intelligence that goes beyond pattern matching.Feb–Apr 2026

▶He holds a strong belief that simply scaling up current deep learning models (LLMs) is an insufficient path to achieving Artificial General Intelligence (AGI).Apr 2026

▶His definition of intelligence, outlined in his 2019 paper "On the Measure of Intelligence," serves as a foundational concept for his work and related initiatives like the ARC Prize Foundation.Feb–Apr 2026

▶Chollet is actively developing an alternative to deep learning based on program synthesis through his research lab, Endia.Feb–Apr 2026

Points of debate

▶Chollet presents a conflicting view on the difficulty of AGI, simultaneously arguing it will be achieved in the 2030s while also claiming it's a simple problem that could have been solved in the 1980s with a small codebase.Apr 2026

▶He is both highly ambitious and pragmatically cautious about his own work, founding Endia to build a new AI substrate while estimating its chance of success at only 10% to 15%.

▶There is a tension between his view that current LLM-based technology can "fully automate" any domain with verifiable rewards and his broader position that the entire LLM stack is inefficient and will be obsolete in 50 years.Apr 2026

▶He believes it is too late to stop AI development, yet his work on the ARC-AGI benchmark is designed to steer the field away from its current, potentially limited, trajectory towards a more robust form of intelligence.Apr 2026

Key themes

▶Critique of the Scaling ParadigmApr 2026

Chollet consistently argues that increasing the scale of deep learning models is not a viable path to AGI. He supports this with evidence from his 2016 research on gradient descent and the poor performance of base LLMs on the ARC-AGI benchmark, even after a 50,000x increase in model scale.

Investors heavily allocated to companies whose primary competitive advantage is scale of compute and data for training LLMs face a significant thesis risk if Chollet's view that a different, more efficient paradigm is required for AGI proves correct.

▶Measuring and Defining 'True' IntelligenceApr 2026

A core part of Chollet's work is defining and measuring a more general, human-like intelligence. He operationalizes this through the ARC-AGI benchmark series, which evolves from static reasoning puzzles (V1) to interactive, game-like environments that test "agentic intelligence" and skill acquisition (V3).

The ARC-AGI benchmark acts as a leading indicator for the future direction of AI research; its increasing adoption could shift focus and funding from pure scaling towards models that demonstrate efficient learning and adaptation in novel environments.

▶Pioneering Program Synthesis as an AlternativeApr 2026

Through his lab, Endia, Chollet is developing a new machine learning substrate based on program synthesis. This approach, which includes methods like "symbolic descent," aims to create models that are more data-efficient and generalizable than current parametric deep learning systems.

Endia represents a high-risk, high-reward venture that could fundamentally disrupt the AI hardware and software stack, as its success would favor different computational models than the matrix multiplication-heavy architecture that currently dominates the industry.

▶AGI as an Elegant, Algorithmic ProblemApr 2026

Chollet frames the pursuit of AGI not as a problem of brute-force scale, but as a search for the right algorithm. He speculates that the final AGI solution will be a surprisingly small codebase (under 10,000 lines) and that the core concepts could have been implemented with 1980s-era computing resources.

This perspective suggests that the winner in the AGI race may not be the entity with the most GPUs, but rather a small, focused team that discovers the correct algorithmic breakthrough, potentially leveling the playing field for smaller research labs and startups.

Source episodes

Sentiment over time

Not enough data for timeline

Changes over time

2016

While at Google Brain, Chollet's research indicated that gradient descent was insufficient for discovering generalizable algorithms for reasoning, instead defaulting to pattern matching.

2019

Chollet published his paper "On the Measure of Intelligence" and created the first version of the ARC-AGI benchmark, containing 800 tasks he personally designed to test fluid intelligence.

Circa 2020-2023

Observed that base LLMs, including early models like GPT-3, scored zero or near-zero on ARC-AGI V1, reinforcing his view that scaling pre-training alone was not improving fluid intelligence.

Present

Founded the research lab Endia to develop a new AI substrate based on program synthesis. Simultaneously, he is advancing the ARC-AGI benchmark to V3 to measure "agentic intelligence" in interactive, game-like environments.

Future

Plans to release future versions of the ARC-AGI benchmark, with V4 focusing on continual learning and V5 on invention, while predicting AGI will be achieved in the early 2030s.

Suggested prompts

How does François Chollet's definition of AGI as efficient skill acquisition in novel tasks directly influence the design of the ARC-AGI benchmark series? &nearr;What are the primary technical and commercial risks associated with Endia's program synthesis approach compared to the dominant deep learning paradigm? &nearr;If Chollet is correct that AGI is a relatively simple algorithmic problem, why has the AI field been so heavily dominated by capital-intensive scaling approaches? &nearr;Analyze the potential impact on the AI industry if the ARC-AGI V3 benchmark for "agentic intelligence" becomes the new standard for evaluating progress toward AGI. &nearr;

Key concepts

ARC-AGI Benchmark 2 ep Artificial General Intelligence (AGI) 2 ep Scaling Limitations 1 ep Endia 1 ep Program Synthesis 1 ep Deep Learning 1 ep Agentic Intelligence 1 ep Symbolic AI 1 ep Verifiable Rewards 1 ep

Notable quotes

“I do believe that you know when you create a gi retrospectively it will turn out that it's a code base that's less than 10 000 lines of code and that if you had if you had known about it back in the in the 1980s you could have done a gi back then using the computer resources available”

François Chollet · François Chollet: Why Scaling Alone Isn’t Enough for AGI

“I personally don't think that machine learning or AI in 50 years is still going to be built on this stack.”

François Chollet · François Chollet: Why Scaling Alone Isn’t Enough for AGI

“So performance of Bezel lamps on V1 stayed very, very low, even though in the meantime, we had scaled up these models by 50,000x, right? So it was really telling you that more scale, scaling up pre-training alone, was not going to crack the benchmark.”

François Chollet · François Chollet: Why Scaling Alone Isn’t Enough for AGI

“We are trying to measure agentic intelligence. So it's interactive, it's active, like the data is not provided to you. You must go get it. The idea is that your agent is dropped into a new environment, which is kind of like a mini video game. And it's not provided any instructions... It must figure out everything on its own via trial and error.”

François Chollet · François Chollet: Why Scaling Alone Isn’t Enough for AGI

Report last updated: Apr 21, 2026

Get started free

Back to Entities Intelligence Report