Search: “francois-chollet--x”

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

Their AI-first ARR tripled year-over-year to surpass 500 million. Very few enterprise software companies are reporting this level of direct, paid AI adoption. The breakout driver is Firefly, now at 30

TL;DR: Adobe is currently the most profitable it has ever been, and it is using AI to accelerating its adoption and earning growth (now 13%, up from 10-11% last year). For all the talk about 'Adobe is

To be clear: 1. No it can't. I've used Fable while it was available, it was a good model but still less than 1% of the way there. 2. If it could, that fact would generally benefit SaaS companies, not

Francois Chollet likely discusses why books are exceptionally valuable or 'overpowered' as learning and knowledge resources, possibly contrasting their effectiveness against other media or modern alte

MaineCoon is the first video model that focuses on social interactions: facial expressions, emotions, fluid conversation, audio-lip sync, etc. Really impressive inference specs: 22B params, 47.5 FPS o

This post from AI researcher François Chollet discusses the psychological principle that initiating a task requires the most effort (activation energy), while momentum carries you through the remainde

The more you embrace AI, the more you need SaaS. This is not obvious to armchair market analysts who love disruption narratives, but it is obvious to people actually running companies. Levie now uses

This post by François Chollet, creator of Keras and AI researcher at Google, likely discusses the iterative learning principle of 'pray, read, re-read' (Ora, lege, relege), possibly applying it as a m

Some considerations that many folks seem not to get: 1. It can be a bubble even if the tech works. (For instance, if the tech doesn't have a high-demand use case.) 2. It can be a bubble even if the te

A French engineer who lives quietly in Paris has spent 30 years writing software that the entire internet now runs on without knowing his name. He wrote the code that streams every YouTube video, ever

I wrote my first neural networks in pure C, then in Matlab, then in NumPy, before eventually upgrading to Theano. Since then I have seen and tried pretty much every NN framework ever developed. Some a

Code volume is not a reliable metric for measuring developer productivity, as writing more lines of code does not necessarily correlate with delivering more value or solving problems more effectively.

Knowledge scaling—increasing the volume of learned information—provides reliable performance on familiar tasks but limited flexibility when facing novel problems. Intelligence, by contrast, enables ad

Anthropic Opus 4.8 is new SOTA on ARC-AGI-3 Score: 1.5%, ~$10K ARC-AGI-3 analysis notes: * Opus 4.8 read the environment an abstraction *above* Opus 4.7, as objects & systems, not pictures * Opus 4.8

Francois Chollet's post likely expresses a philosophical perspective on existential risk or civilizational decline, suggesting that humanity's downfall would be precipitated by a loss of human values,

This post likely references the popular internet trend where people ask others if they've thought about the Roman Empire today, which became a viral meme highlighting how frequently some people contem

Einstein reflects on his creative process, asserting that language plays no significant role in his mechanical thinking and invention, suggesting that his intellectual work occurs through non-linguist

Creativity often thrives within limitations rather than unlimited freedom, as constraints force innovative problem-solving and focused thinking. This insight from AI researcher François Chollet sugges

Developer productivity is hard to manage. Uber’s COO has said that it’s getting “harder to justify” its AI costs because there was no way to show a link between AI spend and any meaningful increase in

Thinking of AI as a productivity booster for prior workflows is the wrong framing. Like all of the previous waves of computerization/softwarization, AI is a tool that lets you do new things in new way

This statement from AI researcher François Chollet suggests that the ability to learn something genuinely novel demonstrates a fundamental learning capacity that generalizes across all domains. The cl

We saw our first meaningful jump in the ARC-AGI-3 competition today @tufalabs went from 0.68% > 1.17% My notes: - .68% is the score of the best template (which is why so many people have this score) -

Francois Chollet reflects on the frequency of his mistakes and incorrect predictions, likely discussing humility in scientific work and the importance of intellectual openness in AI research and devel

Francois Chollet observes that AI systems which always affirm user correctness without nuance or disagreement reduce his confidence in their reliability. This reflects a concern that uncritical agreem

The Codex 'goal' feature will take any silly shortcut possible in order to avoid doing the work (including rewriting your external checks), but if you manage to sufficiently constrain it so that it ha

When an agent is allowed to decompose a goal into smaller sub-tasks, it frequently suffers from goal drift. Left unchecked, it will redefine the optimization metric to favor a simpler, useless sub-tas

Gemini Gemini 3.5 Flash ARC-AGI (Verified) ARC-AGI-2: - High: 72.1%, 0.85 - Minimal: 8.9%, 0.11 ARC-AGI-1: - High: 92.5%, 0.42 - Minimal: 48.8%, 0.06 Gemini 3.5 Flash is on par with GPT-5.5 (Medium) o

Most human tasks are not Markovian, the optimal next action cannot be determined solely by looking at the current state. It depends heavily on the past trajectory, the original intent, and context con

This is true of all agents, not just coding agents. Probably the biggest challenge that most companies run into in their agent strategy is getting agents the right constrained context to work with for