Search: “francois-chollet--x”

ARC Prize 2026: ARC-AGI-3 Milestone #1 Winners Congratulations to the three winners who open-sourced their top-scoring ARC-AGI-3 solutions: 1. @tufalabs - 1.21%, 25K 2. Reki - .867%, 7.5K 3. Md Boktia

Francois Chollet argues that reality operates according to principles that can be understood and manipulated through modeling, suggesting that comprehending the underlying structure of any system enab

Marginal cost is not a detail. It is the whole thing. Imagine going to an airline, and they tell you, 'yes, we can get you to 75% of these top destinations' -- great, but how much does it cost for eac

Reporting benchmark results as a single scalar percentage is insufficient for meaningful evaluation of model performance. Comprehensive benchmark reporting should include efficiency metrics alongside

Francois Chollet proposes a speculative future concept where 'Latent Space Archaeologists' will study the trained weights of 21st-century AI models as historical artifacts to reconstruct extinct knowl

This post likely reflects on American ideals of opportunity and possibility, discussing the cultural narrative that America enables people to pursue diverse ambitions and achieve goals regardless of b

That's a great speech that captures the essence of American exceptionalism. It is not based on innate superiority, power or wealth, but rather on the fact that the outcasts of the world could set foot

Francois Chollet argues that information and energy are the fundamental long-term constraints on technological and civilizational progress, rather than physical materials or other resources. This pers

In retrospect... Yep One very important thing to understand about the future: the economics of AI are about to change completely. We'll soon be in a world where you can turn test-time compute into com

This post explores a philosophical perspective on subjective time perception, suggesting that by engaging in more diverse activities and novel experiences daily, one can psychologically expand their l

Future jobs will prioritize adaptability and creativity over routine task execution, with emphasis on complex problem-framing skills. This reflects how automation and AI are shifting labor demands tow

Francois Chollet discusses how careful observation of ARC-AGI benchmark performance can enable early predictions about AI system capabilities and progress. The post likely highlights patterns or trend

Francois Chollet discusses how current workflows are increasingly incorporating Large Reasoning Models (LRMs) as control systems that direct symbolic programs and tools, representing a pragmatic but i

Eventually, much of AI will converge towards intuition-guided symbolic world modeling, i.e. deep learning-guided program synthesis. It is inevitable. Symbolic modeling lets a system construct a compac

François Chollet observed that all leading contenders on the ARC-AGI-3 benchmark employ a similar methodological approach, suggesting a convergence in techniques for solving abstract reasoning tasks.

ARC-AGI-3 is built different, it has dumbfounded almost all regular attempts so far because it's so much harder than anything that came before. It has no rules, it's agentic and has no explicit goals,

Cross-agent feedback loops are incredibly effective -- for a reason. Check out what @leon2mcp and team at @Bloome_im are building in this space: http://bloome.im Bloome lets you pull Claude, ChatGPT,

On the pod: 'Constrained Adaptive Rejection Sampling' with @ucsd_cse professor @lorisdanto. Hear how symbolic AI experts have navigated the LLM era and why the future of AI code generation depends on

François Chollet argues that the current wave of AI technology will not cause mass unemployment and predicts its impact on the labor market will be minimal. The post likely discusses economic and labo

This post by Keras creator François Chollet appears to reference a new development or tool intended to improve the quality of knowledge base entries or documentation. Without access to the specific co

François Chollet argues that art and science elevate humanity through the creative and discovery processes themselves, rather than through their final outputs or products. He emphasizes that the trans

A good chunk of Tokyo is also reclaimed from the bay, a process that continues to this day but has slowed down significantly (peak was in the 1960s-1970s alongside other major infrastructure buildups)

Autonomy isn't the ability to act without human supervision. It's the ability to *learn* without human bottlenecks in the process. A system that is fully dependent on human training data and RL enviro

If you wanted to do the same in SF, you could start by extending Treasure Island. Obviously the value of that new land would be considerably less than in SF proper. Perhaps 10x less. It would be econo

If your benchmark relies on a static dataset or sampling from a static distribution densely known at training time, then it is fundamentally measuring memorization/retrieval. Which might be fine if yo

This post discusses Tokyo Bay's land reclamation practices, specifically explaining why initial reclamation projects are constructed as isolated islands rather than connected to existing land—a strate

Agentic coding forces you to design clean interfaces and document them well. An agent cannot read the implicit mental model shared by your engineering team, it can only read your API contracts and doc

Francois Chollet argues that a software engineer's true value lies not in writing complex or clever code, but in their ability to protect the codebase from degradation, technical debt, and poor practi

When the cost of execution drops, the value of taste, strategy, and architectural vision skyrockets. Previously, you were spending most of your cognitive budget on the micro. Now you are free to focus

Edge cases and failure modes reveal the fundamental boundaries and constraints of complex systems more effectively than typical operations, making them valuable for understanding system behavior and l

The most complex phenomena arise from scalable recombination of very simple rules. Whether it's galaxies, chips, or neural networks, if you find the right primitive building blocks, the complexity tak

This is the strongest ARC-AGI-2 performance to date by an open-source model. GLM-5.2 from @Zai_org on ARC-AGI (Verified) - ARC-AGI-2: 22.8%, 0.25 - ARC-AGI-1: 77.0%, 0.19 Performance is comparable wit

AI in 2040 will not be built on the stack we are using today. It will be much closer to optimal. The current stack has 3-4 orders of magnitude of data inefficiency and 4-5 orders of magnitude of compu

Francois Chollet presents a framework for understanding different approaches to token usage in AI models: casual users maximize tokens for flexibility, competitive users minimize tokens for efficiency

In software, complexity is a tax you pay on every future change. You might think, 'as long as it works, I'm good, I don't care about aesthetics', but elegant code is about maintainability, not aesthet

This post likely explores the concept that maps don't merely represent territory passively, but actively influence how we perceive, organize, and interact with physical space—a reversal of the traditi

Market efficiency is a myth. Markets are made of people, who are irrational both individually and in aggregate. Maybe if human irrationality were completely random, it might average out to a rational

François Chollet announced that members of the Keras team would be attending the AI Systems DevLabs event in Sunnyvale and invited people to meet them there. This post indicates a scheduled appearance

With agentic coding, complexity compounds in a mechanical way: unnecessary code ends up in the codebase, moves to the context window, degrades the model's reasoning abilities, leads to more unnecessar

A company is not its share price; it's possible for a company to execute exceptionally well while its share price declines, much like it's possible for a company in terminal decline to see its share p

Chollet draws an analogy between visual recursion and self-referential AI systems, suggesting that just as mirrors facing each other create a false sense of infinite depth, a prediction machine analyz

Adobe has integrated AI capabilities into its software to improve user accessibility for novices, which has significantly boosted adoption of its freemium offerings. This approach demonstrates how AI-

Adobe hit a record 6.62 billion in Q2 revenue (up 13% YoY) while non-GAAP EPS jumped 18% to 5.96. Its net margins are at 36%, close to an all-time high. They are absorbing the compute costs of generat

It seems almost too dumb to be true, but apparently the literal belief of SaaS bears is 'all software is a 0 because Claude can one-shot these apps' Just staggering levels of short-sightedness in that

Programming is fundamentally about managing complexity through abstraction layers rather than merely writing code, analogous to how music transcends its written notation. This perspective, shared by A

The market is treating Adobe like a legacy software company in terminal decline. Yet the actual data shows it's one of the biggest beneficiaries of the rise of GenAI. In fact, it's one of the top 5 mo