The A.I. Daily
François Chollet@fchollet
Dopamine does not spike highest for guaranteed rewards. It spikes highest at reward prediction error. If you want to have fun in life, you should stop avoiding risk and uncertainty.
François Chollet
36 storiesFrançois Chollet is the creator of Keras and a driver of the ARC-AGI benchmark, known for his work on intelligence and reasoning. Recent posts discuss Gemini 4 Argon's scores on ARC-AGI-3, ARC-AGI-2 and ARC-AGI-1, the effectiveness of symbolic learning, why robotics needs causality and discrete reasoning, and LLM post-training experiments with KerasHub, JAX and TPUs. He also critiques AI doomer culture, argues intelligence needs grounding in reality, and highlights conversational AI in education.
Related topics
François Chollet@fchollet
Symbolic learning works
François Chollet@fchollet
A common sign that you are in an insane millenarian death cult is that you axiomatically convinced that the Apocalypse will happen within a few months, you are constantly on the lookout for signs that it may have started, you feel both fearful and *excited* a…
François Chollet@fchollet
When enough people believe something will happen, it doesn't need to happen at all for all of its consequences to be felt
François Chollet@fchollet
Some observations about the "total" automation of software engineering: 1. There are more software engineers now than in 2025, and demand is growing. 2. Software engineers have never worked harder than they do now. 3. They largely don't write code anymore, ag…
François Chollet@fchollet
Argon has a similar performance profile on ARC-AGI-3 as Astra. Google is back at the frontier. [Quoting @arcprize]: Gemini 4 Argon from @Google on ARC-AGI (Verified): - ARC-AGI-3: 97.5%, $4.8K (provider adapter harness), 49.8%, $5.8K (standard harness) - ARC-…
François Chollet@fchollet
I truly believe conversational AI will drive the biggest educational leap of our generation. 1-on-1 tutoring is the gold standard for education, but it doesn't scale. AI can change that with truly active conversational learning. Annie and team set out to solv…
François Chollet@fchollet
Everyone wants to have solved a hard problem, but almost nobody wants to live inside the grind of actually solving it. The breakthroughs belong to the few who are obsessed with the process, not just the idea of the outcome
François Chollet@fchollet
You can't solve real-world physical tasks through autoregressive sampling alone. The physical world is entirely unforgiving to hallucination or approximate pattern matching. If we want actual autonomous robotics we have to engineer causality and discrete reas…
François Chollet@fchollet
String theory is what happens when an intelligent system is starved of sensory input and starts hallucinating. Intelligence needs grounding. For this reason, intelligence is practically never the bottleneck to solving impactful problems. The bottleneck is ite…
François Chollet@fchollet
The ideal outcome for science isn't a couple of labs using proprietary models and millions of dollars to crack trophy problems, it's to have millions of AI-aided human scientists who follow their curiosity and become way more ambitious
François Chollet@fchollet
Congrats on the release! Looks like a fantastic model. [Quoting @sundarpichai]: Gemini 4 Argon is here. It’s available globally today for developers and enterprises in the paid tier of the Gemini API and the Gemini platform, for Ultra subscribers in Google An…
François Chollet@fchollet
If you explained agriculture to folks in SF, you'd blow so many minds. It's... self-replicating, recursively self-improving nanotech for the physical world. And it's all open-source! One instance can turn into hundreds of seeds, that turn into hundreds of ins…
François Chollet@fchollet
I've been having a lot of fun trying a wide range of LLM post-training experiments using KerasHub + JAX + TPUs. Might open-source some of that
François Chollet@fchollet
How AI shifted from a transductive paradigm (token-by-token answer completion) to an inductive paradigm (on-the-fly reasoning chain synthesis), in one chart. This is the big story of 2025 and 2026. This is what made coding agents work.
François Chollet@fchollet
The real world abounds with recursively self-improving systems, but one in particular deserves attention: Science, modeled as a system (perhaps even as an agent, with goals and resources). If you want to really understand AI RSI, science should be your refere…
François Chollet@fchollet
Right now, using AI agents still feels very single-player. They do your work, but the results are trapped in isolated chat histories and nothing compounds. The Busabase team (@Busabase4agent) is launching an open-source (MIT) database and workspace for agents…
François Chollet@fchollet
We're reaching really good ARC-AGI-2 scores on Kaggle now! [Quoting @arcprize]: New ARC Prize 2026 - ARC-AGI-2 High Score 88.06% by @tufalabs The $150K Bonus Prize, on top of the guaranteed prizes, will be split among all teams scoring over 85%
François Chollet@fchollet
Most measures of AI progress closely fit an exponential over the 2023-2026 period. Meanwhile AI capex over the same period has been slightly super-exponential (i.e. the rate of growth itself is increasing). So if you model AI as a system that takes as input e…
What happened to 2024's BABA is AI paper?
A Reddit user revisits a 2024 ICML paper that found GPT-4o, Gemini-1.5-Pro and Gemini-1.5-Flash fail dramatically when generalization requires manipulating and combining game rules, asking whether this still holds for today's agentic systems.
François Chollet@fchollet
Humans provide a fun comparison point here. In the 80s, the "cognitive consequences of programming" was a major psychological research focus. The prevailing hypothesis was that learning formal logic and algorithmic structures through computer programming woul…
François Chollet@fchollet
The notion that current AI models are sentient and can suffer, combined with the foolish idea that suffering can be mathematically quantified and weighted between humans and non-humans, could lead us down an incredibly dark and dystopian path. But before it g…
François Chollet@fchollet
The critical distinction between base LLMs (2024 and earlier) and modern LRMs is not symbolic tool use. It's the switch from a transductive paradigm (intuit the answer to the query) to an inductive paradigm (intuit the program/instructions that produce the an…
François Chollet@fchollet
What is a way to test human-level generality in artificial intelligence? "The meta-benchmark of being able to pass ARC-AGI-(n+1) immediately upon release" [Quoting @dylantmoore]: @Jabaluck @DAcemogluMIT The meta-benchmark of being able to pass ARC-AGI-(n+1) i…
François Chollet@fchollet
I don't read/write code these days, I only instruct an LRM. But it's not because I think LRM code quality is perfect, or even good. And it's not because I think my instructions are always being perfectly acted upon -- I know they aren't. The thing is: LRMs do…
François Chollet@fchollet
The purpose of the AI industry should be to produce tools that, in the human hand, will improve human prosperity and welfare. It should not be to create a "successor species" to the human race. Entertaining such a thought makes you, de facto, the enemy of all…
François Chollet@fchollet
I think the "difficulty" of software engineering is essentially constant no matter what abstraction level you move to, because human cognition adapts to new tools until it can fully utilize itself. Tools are only affordances, not a magic wand that makes work…
François Chollet@fchollet
At some point in evolutionary history, matter learned to feel. We don't know how. It wouldn't be scientific to dismiss the possibility that it could happen again. (I don't believe it happens in current AI systems, because they don't have the charactistics I w…
François Chollet@fchollet
The tweet below got enormous pushback at the time I posted it, more than 5 years ago. But this prediction is looking obvious by the day now. [Quoting @fchollet]: Within 10-20 years, nearly every branch of science will be, for all intents and purposes, a branc…
François Chollet@fchollet
Summer 2026 has been a crazy time in AI... [Quoting @arcprize]: ARC Prize 2026 - ARC-AGI-3 Progress Prize - 9 day left $37,500 in prizes are being awarded on *Sept 30th* to the top open source solutions Current standings: 1. Tufa Labs 2. Lord Han Solo 3. NVAR…