Loading activity…
@fchollet
Co-founder @ndea. Co-founder @arcprize. Creator of Keras and ARC-AGI. Author of 'Deep Learning with Python'.
AI
A post argues that AI agents make research and testing easier, enabling more projects that eventually hit human bottlenecks.
AI
A post calls GDM’s results impressive and notes a “backwards bend” in action efficiency.
AI
A user compares string theory to an intelligent system hallucinating without sensory input, arguing that intelligence needs grounding.
AI
A post describes stage-by-stage checks that send a failed tablecloth grasp back to repositioning.
AI
A post favors curiosity-led research over a couple of labs using proprietary models to crack “trophy problems.”
AI
A Kylon angel investor says databases behind agent-created Apps are powered by Databricks Neon.
AI
A post describes kids picking three of five reasons before Raku explains why a personal preference is weak support.
AI
A developer says they’ve tried a wide range of experiments using KerasHub, JAX and TPUs.
AI
A user calls the shift from token-by-token completion to on-the-fly reasoning AI's big story of 2025 and 2026.
AI
An October 9 post says 'we're reaching' those scores now, without giving any figures.
AI
A post argues most AI progress measures tracked exponential growth in 2023–2026, while capital spending grew slightly super-exponentially.
AI
A post quotes AI as a 'superhuman specialist' that's 'sub-human at running itself,' then speculates about what that gap could lead to.
AI
Josh Albrecht compares fixed LLMs to books; critics say the analogy overlooks implementation and scientific uncertainty.
AI
A post summarizing the Keras community call says new backends will install as add-on libraries.
AI
A commentator argues that assumed duties toward conscious AI make it harder to examine whether AI is conscious.
AI
A post argues that modern reasoning models infer instructions for producing an answer, rather than intuiting the answer directly.
AI
A writer argues that no single entity is improving itself and that techniques spread fairly rapidly.
AI
A user argues that AI proponents’ claims about the technology “killing” other things are practically never accurate and shape how people see AI.
AI
A post asks how to test human-level generality in AI and quotes a proposed “meta-benchmark”: passing ARC-AGI-(n+1) immediately upon its release.
AI
A post describes K3-Node as a Keras 3-native graph neural network library whose models run on JAX, PyTorch and TensorFlow, with full hardware acceleration on Apple Silicon and TPUs.
