François Chollet
Creator of the Keras deep-learning library and the ARC-AGI benchmark.
François Chollet did not write this page. ppll compiled it from their own words, each linked to the source. Is this you? · Report a mistake
Latest x.com
What they believe
-
Breakthroughs belong to the few who are obsessed with the process, not just the idea of the outcome
The breakthroughs belong to the few who are obsessed with the process, not just the idea of the outcome
x.comSaid 15 Oct 2026This on korrents
-
Changed their mind The LLM line of research will reach a capability plateau.
Now In the era of base LLM scaling (2022-2024), I believed the LLM line of research would reach a capability plateau (as later seen with base LLMs). In late 2024, after the o3 test-time compute demo, I changed my views: the new models were showing genuine fluid intelligence, and with this new line of work, the LLM line of research could achieve unbounded capability scaling. "There will be no wall."
Said 7 Aug 2026This on korrents
The bold line is korrents’ wording; the quote is theirs. These one: what they wrote before what they said aloud, one per source, newest first.
Feed
As its own page →Hiding
15 October
-
korrents.com
Breakthroughs belong to the few who are obsessed with the process, not just the idea of the outcomeTheir words
The breakthroughs belong to the few who are obsessed with the process, not just the idea of the outcome
14 October
From one piece @fchollet on X 2 beliefs · x.com
-
Their words
For this reason, intelligence is practically never the bottleneck to solving impactful problems.
-
korrents.com
Actual autonomous robotics requires engineering causality and discrete reasoning back into the loop.Their words
If we want actual autonomous robotics we have to engineer causality and discrete reasoning back into the loop.
-
Their words
You can't solve real-world physical tasks through autoregressive sampling alone.
+ 1 more
-
Their words
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
13 October
From one piece @fchollet on X 2 beliefs · x.com
-
Their words
Isolated chats are suboptimal in team environments because they lack the context of what the team already discussed and decided.
-
Their words
If an agent saves an engineer 2 hours to implement a feature but collectively costs the team 3 hours in alignment and review, the ROI is negative.
11 October
10 October
-
korrents.com
The shift from a transductive paradigm to an inductive paradigm is the big story of 2025 and 2026.Their words
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.
9 October
-
Their words
This is why science as a system requires exponential resources (input) to produce linear impact (output).
8 October
From one piece @fchollet on X 2 beliefs · x.com
-
korrents.com
To achieve sustained exponential progress, the AI industry will need to achieve matching exponential profits.Their words
To achieve sustained exponential progress, the AI industry will need to achieve matching exponential profits -- it will need to close the resource loop.
-
Their words
But super-exponential injection of *external* resources is not sustainable in the long-term.
4 October
From one piece @fchollet on X 2 beliefs · x.com
-
korrents.com
Machines are not consciousTheir words
Likewise a static input-output program that has *none* of the properties we associate with conscious beings (e.g. information integration, interoception, temporal binding, embodiment, etc.) has no more reason to be presumed conscious than a rock has to be presumed alive.
-
Their words
creating artificial consciousness should *not* be a goal of AI research. Understanding consciousness, sure, synthesizing it, no.
-
Their words
Humanity has not yet created a conscious computer program, and there are no signs we are close to doing so.
3 October
-
Their words
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.
3 September
-
Lovedrcmnd.app
GPT-6 AstraTheir words
GPT-6 Astra represents a step-function change in model capability for interactive reasoning problems. It scores 66% on ARC-AGI-3 using our standard harness, and nearly 100% with a continuous conversation harness and custom compaction, at a cost of roughly $360 per game.
24 August
-
Recommendstheir ownrcmnd.app
Deep Learning with PythonTheir words
If you're 17 (or any age) and you want to learn to build LLMs from scratch, read chapters 15-16 of Deep Learning with Python
7 August
-
measuring intelligenceLLMsscaling laws
On changing their mind
In the era of base LLM scaling (2022-2024), I believed the LLM line of research would reach a capability plateau (as later seen with base LLMs). In late 2024, after the o3 test-time compute demo, I changed my views: the new models were showing genuine fluid intelligence, and with this new line of work, the LLM line of research could achieve unbounded capability scaling. "There will be no wall."
18 September 2025
26 May 2025
4 April 2025
20 December 2024
-
measuring intelligenceLLMsOpenAI
Their words, before
This "memorize, fetch, apply" paradigm can achieve arbitrary levels of skills at arbitrary tasks given appropriate training data, but it cannot adapt to novelty or pick up new skills on the fly (which is to say that there is no fluid intelligence at play here.)
Their words now
OpenAI's new o3 model represents a significant leap forward in AI's ability to adapt to novel tasks. This is not merely incremental improvement, but a genuine breakthrough, marking a qualitative shift in AI capabilities compared to the prior limitations of LLMs.
12 February 2024
1 November 2023
24 April 2021
1 October 2020
2 February 2018
Nothing matches.
Beliefs
Korrents What they believe 51 beliefs, 2 changed minds — each backed by an exact quote.
Each is a — compiled by korrents.com, not by them: the one-line wordings are korrents', the quotes are theirs.
Recent
Breakthroughs belong to the few who are obsessed with the process, not just the idea of the outcome
The breakthroughs belong to the few who are obsessed with the process, not just the idea of the outcome
@fchollet on X Said 15 Oct 2026
The ideal outcome for science is to have millions of AI-aided human scientists who follow their curiosity and become more ambitious, rather than a few labs using proprietary models to solve trophy problems.
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
@fchollet on X Said 14 Oct 2026
Real-world physical tasks cannot be solved through autoregressive sampling alone.
You can't solve real-world physical tasks through autoregressive sampling alone.
@fchollet on X Said 14 Oct 2026
Show 48 more
Actual autonomous robotics requires engineering causality and discrete reasoning back into the loop.
If we want actual autonomous robotics we have to engineer causality and discrete reasoning back into the loop.
@fchollet on X Said 14 Oct 2026
Intelligence is practically never the bottleneck to solving impactful problems.
For this reason, intelligence is practically never the bottleneck to solving impactful problems.
@fchollet on X Said 14 Oct 2026
If an agent saves an engineer 2 hours to implement a feature but collectively costs the team 3 hours in alignment and review, the ROI is negative.
If an agent saves an engineer 2 hours to implement a feature but collectively costs the team 3 hours in alignment and review, the ROI is negative.
@fchollet on X Said 13 Oct 2026
Isolated chats are suboptimal in team environments because they lack the context of what the team already discussed and decided.
Isolated chats are suboptimal in team environments because they lack the context of what the team already discussed and decided.
@fchollet on X Said 13 Oct 2026
The shift from a transductive paradigm to an inductive paradigm is the big story of 2025 and 2026.
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.
@fchollet on X Said 10 Oct 2026
Science as a system requires exponential resources to produce linear impact.
This is why science as a system requires exponential resources (input) to produce linear impact (output).
@fchollet on X Said 9 Oct 2026
Super-exponential injection of external resources is not sustainable in the long-term.
But super-exponential injection of *external* resources is not sustainable in the long-term.
@fchollet on X Said 8 Oct 2026
To achieve sustained exponential progress, the AI industry will need to achieve matching exponential profits.
To achieve sustained exponential progress, the AI industry will need to achieve matching exponential profits -- it will need to close the resource loop.
@fchollet on X Said 8 Oct 2026
Humanity has not yet created a conscious computer program
Humanity has not yet created a conscious computer program, and there are no signs we are close to doing so.
@fchollet on X Said 4 Oct 2026
Creating artificial consciousness should not be a goal of AI research
creating artificial consciousness should *not* be a goal of AI research. Understanding consciousness, sure, synthesizing it, no.
@fchollet on X Said 4 Oct 2026
Machines are not conscious
Likewise a static input-output program that has *none* of the properties we associate with conscious beings (e.g. information integration, interoception, temporal binding, embodiment, etc.) has no more reason to be presumed conscious than a rock has to be presumed alive.
@fchollet on X Said 4 Oct 2026
The belief that current AI models are sentient and can suffer, combined with the foolish idea of quantifying suffering across humans and non-humans, could lead to a dystopian path.
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.
@fchollet on X Said 3 Oct 2026
Competent information processing and sentience are wholly unrelated
Do not confuse competent information processing and sentience, they are wholly unrelated.
@fchollet on X Said 3 Oct 2026
The AI industry should focus on creating tools that enhance human prosperity and welfare, not on developing a successor species.
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.
@fchollet on X Said 28 Sept 2026
The difficulty of software engineering is probably essentially constant across abstraction levels, because human cognition adapts to new tools.
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.
@fchollet on X Said 25 Sept 2026
At some point in evolutionary history, matter learned to feel.
At some point in evolutionary history, matter learned to feel.
@fchollet on X Said 24 Sept 2026
Those who relied on code as their source of truth for managing understanding now need a new source-of-truth artifact.
If code was previously your source of truth for managing understanding, then you need a new source of truth artifact.
@fchollet on X Said 21 Sept 2026
People who built their understanding through code-centric workflows need new workflows to do so.
If code-centric workflows were how you developed that understanding, then you need new workflows.
@fchollet on X Said 21 Sept 2026
People almost always align their beliefs with their interests.
People will almost always align their beliefs and their interests.
@fchollet on X Said 14 Sept 2026
Sincerity and self-interest usually go together by default.
You can be sincere *and* self-interested, and that's usually the default.
@fchollet on X Said 14 Sept 2026
Extreme power concentration is one of the most worrying risks linked to frontier AI.
One of the most worrying risks linked to frontier AI is extreme power concentration.
@fchollet on X Said 13 Sept 2026
The only way to avoid extreme power concentration is multiple independent providers of frontier AI models, including open-source options.
The only way to avoid extreme power concentration is to ensure we have multiple independent providers of frontier AI models, including open-source options.
@fchollet on X Said 13 Sept 2026
In the near term, more capable models should mean safer models.
In the near term (definitely not in the long term), more capable models should mean safer models (maybe paradoxically).
@fchollet on X Said 13 Sept 2026
Current models are unsafe because they are not smart enough, not because they are too smart.
Current models are unsafe not because they're too smart, but because they take goals too literally or take nonsensical shortcuts to achieve these goals, i.e. they're RL-fried. They lack common sense. They don't do the right thing in the face of ambiguity. Basically, they're not smart enough.
@fchollet on X Said 13 Sept 2026
If AI has a high chance of causing human extinction within years or decades, the only rational stance on safety monitoring and research pacing is stringent top-down government involvement and universally ratified international treaties.
If there really is a high chance of AI leading to the extinction of humanity within years/decades, then the only rational stance towards safety monitoring and research pacing should be stringent, top-down government involvement and universally ratified international treaties.
@fchollet on X Said 13 Sept 2026
It is not rational to say something is likely to end humanity in one's lifetime while planning to completely wing it.
But you cannot say "yes this is likely to end humanity in your lifetime, and we're going to completely wing it." That is simply not rational.
@fchollet on X Said 13 Sept 2026
The case for building AGI has always been framed around invention, such as curing cancer or solving fusion energy.
The narrative around why building AGI was a good idea has always been about invention -- "it will cure cancer", "it will solve fusion energy", etc.
@fchollet on X Said 7 Sept 2026
AGI should not be declared until AI is capable of invention, including conceptual breakthroughs, novel insights, and new real-world technology.
So I would not "declare AGI" until we have AI that is, at last, capable of invention -- conceptual breakthroughs, novel insights, new real-world technology, etc.
@fchollet on X Said 7 Sept 2026
Test-time scaling seems to have gained a third axis: latent space reasoning iterations in looped transformers.
Seems like test time scaling has gained a 3rd axis: latent space reasoning iterations in looped transformers.
@fchollet on X Said 5 Sept 2026
Capabilities provided by harnesses around AI models are increasingly shifting into the model itself.
so harness capabilities are increasingly shifting into the model itself.
@fchollet on X Said 3 Sept 2026
In verifiable domains model capability scaling should remain unbounded, because the space of enumerable patterns is infinite by construction.
In verifiable domains, model capability scaling should remain unbounded. Models will simply keep improving by "absorbing more and more of the computational universe", which is infinite by construction.
@fchollet on X Said 27 Aug 2026
A growing fraction of social media is AI slop influencers being replied to by bots — an echo of an echo of an echo.
An increasing fraction of social media consists of slop influencers using AI to make posts and bots replying to them. An echo of an echo of an echo
@fchollet on X Said 23 Aug 2026
Calling today's economic effects "the Singularity" waters the term down past recognition; Vinge meant an event horizon beyond which nothing is imaginable.
Incredible watering down -- the Singularity is now redefined to mean "the rate of new firm creation has increased somewhat" Vernor Vinge described the Singularity as an event horizon past which everything (e.g. what happens tomorrow) becomes entirely unimaginable and unpredictable to human understanding
@fchollet on X Said 19 Aug 2026
Coding is not just another application domain for AI; it is the meta-skill that lets AI generate its own training material and start the recursive self-improvement loop.
Coding isn't yet another application domain -- it's the meta-skill required for AI to automatically develop its own training material, via symbolic world models. That's how the RSI loop actually kicks off.
@fchollet on X Said 10 Aug 2026
Future AI, say in fifteen years, will not be built on the LLM stack; it will have to move to symbolic learning.
However, looking ahead, I still do not believe that future AI (say, in 15 years) will be based on the LLM stack. I believe it will necessarily have to move closer to its optimal, final form -- symbolic learning. Obviously this is a risky and contrarian belief -- the safe bet would be LRMs. But let's see.
@fchollet on X Said 7 Aug 2026
Current AI techniques are four to six orders of magnitude away from optimal in data efficiency and test-time compute efficiency.
I believe current techniques are 4-6 orders of magnitude away from optimality in terms of data efficiency and test-time compute efficiency. But far future AI will be near-optimal.
@fchollet on X Said 7 Aug 2026
Human adaptability is fundamentally linked to the efficient construction and refinement of internal world models
We argue that this profound adaptability is fundamentally linked to the efficient construction and refinement of internal representations of the environment, commonly referred to as world models, and we refer to this adaptation mechanism as world model induction.
Assessing Adaptive World Models in Machines with Novel Games (with 13 co-authors) Said 17 Jul 2025
Current AI evaluation of world models is too narrow, favoring static representations learned from massive datasets over efficient learning through interaction
However, current understanding and evaluation of world models in artificial intelligence (AI) remains narrow, often focusing on static representations learned from training on massive corpora of data, instead of the efficiency and efficacy in learning these representations through interaction and exploration within a novel environment.
Assessing Adaptive World Models in Machines with Novel Games (with 13 co-authors) Said 17 Jul 2025
Rapid world model induction will be critical for achieving robust, general AI that functions in complex, fast-changing real and human worlds
We expect that building and evaluating AI systems capable of this kind of rapid world model induction will be critical for achieving robust, general AI capable of functioning effectively in the complex and fast-changing real world, and especially in human worlds - the environments that human beings have evolved in, created, and are continually changing and re-creating.
Assessing Adaptive World Models in Machines with Novel Games (with 13 co-authors) Said 17 Jul 2025
An adaptive agent's world model cannot be a static representation learned once and fixed
For adaptation to complex and changing environments, an agent's world models cannot be static representations learned once and fixed.
Assessing Adaptive World Models in Machines with Novel Games (with 13 co-authors) Said 17 Jul 2025
Games are particularly rich and controlled environments uniquely well-suited for evaluating rapid world model induction in AI
We contend that games provide particularly rich and controlled environments uniquely well-suited for systematically evaluating rapid model adaptation and the process of world model induction.
Assessing Adaptive World Models in Machines with Novel Games (with 13 co-authors) Said 17 Jul 2025
Scaling the 2019-2023 recipe — same architecture, bigger model, more data — is not enough; further progress depends on new architectural ideas.
o3's improvement over the GPT series proves that architecture is everything. You couldn't throw more compute at GPT-4 and get these results. Simply scaling up the things we were doing from 2019 to 2023 -- take the same architecture, train a bigger version on more data -- is not enough.
OpenAI o3 Breakthrough High Score on ARC-AGI-Pub Said 20 Dec 2024
Passing ARC-AGI does not amount to achieving AGI: o3 still fails on some very easy tasks, indicating fundamental differences from human intelligence.
Passing ARC-AGI does not equate to achieving AGI, and, as a matter of fact, I don't think o3 is AGI yet. o3 still fails on some very easy tasks, indicating fundamental differences with human intelligence.
OpenAI o3 Breakthrough High Score on ARC-AGI-Pub Said 20 Dec 2024
Progress toward more intelligent artificial systems requires defining and evaluating intelligence so systems can be compared with each other and with humans.
To make deliberate progress towards more intelligent and more human-like artificial systems, we need to be following an appropriate feedback signal: we need to be able to define and evaluate intelligence in a way that enables comparisons between two systems, as well as comparisons with humans.
On the Measure of Intelligence Said 5 Nov 2019
Solely measuring skill at any given task falls short of measuring intelligence, because skill is heavily modulated by prior knowledge and experience.
We argue that solely measuring skill at any given task falls short of measuring intelligence, because skill is heavily modulated by prior knowledge and experience: unlimited priors or unlimited training data allow experimenters to "buy" arbitrary levels of skills for a system, in a way that masks the system's own generalization power.
On the Measure of Intelligence Said 5 Nov 2019
The Abstraction and Reasoning Corpus can measure human-like general fluid intelligence and enable fair comparisons between AI systems and humans.
We argue that ARC can be used to measure a human-like form of general fluid intelligence and that it enables fair general intelligence comparisons between AI systems and humans.
On the Measure of Intelligence Said 5 Nov 2019
Both the special-purpose-programs view and the blank-slate view of human intelligence are likely incorrect.
Today, it is increasingly apparent that both of these views of the nature of human intelligence - either a collection of special-purpose programs or a general-purpose Tabula Rasa - are likely incorrect
On the Measure of Intelligence Said 5 Nov 2019
If intelligence is the process of acquiring skills, no single task demonstrates intelligence unless it is a meta-task of skill-acquisition across many tasks.
If intelligence lies in the process of acquiring skills, then there is no task X such that skill at X demonstrates intelligence, unless X is actually a meta-task involving skill-acquisition across a broad range of tasks.
On the Measure of Intelligence Said 5 Nov 2019
Beliefs others hold too
Machines are not conscious 4 hold this
Likewise a static input-output program that has *none* of the properties we associate with conscious beings (e.g. information integration, interoception, temporal binding, embodiment, etc.) has no more reason to be presumed conscious than a rock has to be presumed alive.
@fchollet on X Said 4 Oct 2026
Scaling the 2019-2023 recipe — same architecture, bigger model, more data — is not enough; further progress depends on new architectural ideas. 2 hold this
o3's improvement over the GPT series proves that architecture is everything. You couldn't throw more compute at GPT-4 and get these results. Simply scaling up the things we were doing from 2019 to 2023 -- take the same architecture, train a bigger version on more data -- is not enough.
OpenAI o3 Breakthrough High Score on ARC-AGI-Pub Said 20 Dec 2024
What is a korrent?
A korrent is a belief a person has stated in their own words: one sentence stating the claim, backed by a quote and a source, kept at korrents.com.
Under a name here, the quoted block is what they actually said. The korrent beneath it is the claim those words support, in korrents' wording — tap it to see the record, its source, and who else holds it.
Nobody here wrote their own korrents. They are compiled from public statements, and a person can change their mind, which is recorded too.
Recommends rcmnd
rcmnd What they recommend 2 things — each backed by an exact quote.
Compiled by rcmnd.app, not by them: the quotes are theirs, and each links to the post it was copied from.
Recent
Loved GPT-6 Astra
GPT-6 Astra represents a step-function change in model capability for interactive reasoning problems. It scores 66% on ARC-AGI-3 using our standard harness, and nearly 100% with a continuous conversation harness and custom compaction, at a cost of roughly $360 per game.
x.com 3 Sept 2026
Recommends Deep Learning with Python
If you're 17 (or any age) and you want to learn to build LLMs from scratch, read chapters 15-16 of Deep Learning with Python
x.com 24 Aug 2026
their own
Where they publish
Site fchollet.com Personal page: bio, books, talks and papers.
The front door — who they are, the Keras and deep-learning books, selected talks and papers, and links to everywhere else they publish.
Link verified 1 Sept 2026.
x.com @fchollet Where the arguments about what AI can and cannot do happen.
The most active channel: long threads on intelligence, benchmarks and the limits of scaling, plus announcements for ARC Prize and Keras.
Recent
- Argon has a similar performance profile on ARC-AGI-3 as Astra. Google is back at the frontier. 15 Oct 2026
- 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 solve this with @gakkuai, and they've done a great job. Gakku is an AI mentoring app for K-12 students. Rather than doing the work for them, the voice chat coach guides students through 10 min essay writing "missions" that cover brainstorming, drafting, and revising. It actively trains kids to generate their own ideas, structure logical argum… 15 Oct 2026 gakku.ai
- 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 15 Oct 2026
Show 17 more
- 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 reasoning back into the loop. Biwei and the team at Aether AI is doing exactly the kind of foundational, non-consensus work the field actually needs right now. 14 Oct 2026
- 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 iterated contact with reality. 14 Oct 2026
- 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 14 Oct 2026
- If an agent saves an engineer 2 hours to implement a feature but collectively costs the team 3 hours in alignment and review, the ROI is negative. Isolated chats are suboptimal in team environments because they lack the context of what the team already discussed and decided. The Kylon app fixes this by running Claude Code and Codex in a shared, multiplayer environment. Agents gain real-time visibility into team consensus and architecture decisions. 13 Oct 2026
- Congrats on the release! Looks like a fantastic model. 13 Oct 2026
- 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 instances. Next thing you know you have trillions of them. It's super-exponential 13 Oct 2026
- 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 11 Oct 2026
- 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. 10 Oct 2026
- 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 reference point. Science is an intelligent system, and it is obviously recursively self-improving: 1. Scientific discoveries unlock new technology that helps build better experimental tools. This is a top driver of progress in nearly all fields. 2. They unlock new conceptual advances (ideas, theories) that help solve more problems. 3. They incr… 9 Oct 2026
- 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 to fix this. Instead of losing work to ephemeral chats, your agents write directly to a structured workspace to turn their outputs into reusable data, docs, and skills your entire team can access. Details here: 9 Oct 2026 busabase.com
- We're reaching really good ARC-AGI-2 scores on Kaggle now! 9 Oct 2026
- 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 external investment and outputs AI progress, that system has a slightly sub-linear response. But super-exponential injection of *external* resources is not sustainable in the long-term. To achieve sustained exponential progress, the AI industry will need to achieve matching exponential profits -- it will need to close the resource loop. 8 Oct 2026
- 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 would act as a form of mental gymnastics that could upgrade general reasoning, planning, and novel problem-solving abilities. But foundational research, followed by subsequent meta-analyses, demonstrated that this does not happen. Students who learn programming become highly proficient at algorithmic thinking and coding, but these gains do no… 8 Oct 2026
- What if the jagged frontier is mainly math + code (which you can push arbitrarily far with RLVR), and everything else starts to plateau because it is still bottlenecked by human generated data? Model performance in non-verifiable areas has kept improving steadily, albeit much slower than for math and code. But is that steady improvement a side effect of a higher G (itself driven by RLVR), or only a function of the amount of new human data getting injected into training (which is still continually happening on a massive scale)? A lot of things depend on the answer to this question 7 Oct 2026
- Arguing that "consciousness arises from a computational substrate, and since AI models are computation, they are likely conscious," is as meaningless as arguing that "living creatures are made of atoms, and since this rock is made of atoms, it is likely alive." Everything is computation. Sometimes incredibly sophisticated computation, like a chess engine, AlphaGo, google3, or the Linux kernel. That does not make it conscious. A rock being made of atoms does not confer it any of the properties we associate with living things, also made of atoms. Likewise a static input-output program that has… 4 Oct 2026
- 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 gets to that point, it will rightfully be met with immense backlash from team humans. 3 Oct 2026
- Summary of the Keras community call -- 1. New pluggable backend plan, upcoming MLX and PaddlePaddle backends; new backends are installed as add-on libraries 2. Reducing number of ops needed to ship new backends; adding backend-agnostic implementations of all secondary ops 3. Streamlining unit testing for new ops; a single harness can now test all ops functionality like casting consistency, etc. 4. Lots of new models available in KerasHub 5. Moving KerasHub away from tf-text preprocessing and adopting PyGrain instead 2 Oct 2026
Link verified 1 Sept 2026. Recent items update automatically from the channel.
GitHub fchollet Keras and the rest of the code.
The code behind the work — Keras above all, plus the smaller research repos. The commit history is where the engineering, as opposed to the argument, lives.
Recent
- deep-learning-with-python-notebooks — Jupyter notebooks for the code samples of the book "Deep Learning with Python" 18 Sept 2025
- namex — Clean up the public namespace of your package! 26 May 2025
- ARC-AGI — The Abstraction and Reasoning Corpus 4 Apr 2025
Show 5 more
- keras-resources — Directory of tutorials and open-source code repositories for working with Keras, the Python deep learning library 12 Feb 2024
- keras-blog — Blog with Keras news, tutorials, and demos. 1 Nov 2023
- nelder-mead — Pure Python/Numpy implementation of the Nelder-Mead algorithm. 24 Apr 2021
- deep-learning-models — Keras code and weights files for popular deep learning models. 1 Oct 2020
- hualos — Keras Total Visualization project 2 Feb 2018
Link verified 1 Sept 2026. Recent items update automatically from the channel.
About the English under a post
Some people here publish in a language other than English. Where they do, this site shows a machine translation beneath the post, in this typeface — the site's own, not theirs.
The post itself is never changed, moved or hidden: what is set in the serif above is exactly what the person published, and it is what to quote them on. A translation can be wrong in ways that matter, especially about tone.
Only the post's own words are translated. A quoted post, a linked article and a belief on korrents.com are left in their original language.