GPT-6 Astra Cracks a Decade-Old Voting Theory Problem Nobody Could Solve
Epoch AI marks the first Major Advance on its unsolved-math benchmark, with GPT-6 Astra driving the proof in an interactive session with three mathematicians.
20 stories
Epoch AI marks the first Major Advance on its unsolved-math benchmark, with GPT-6 Astra driving the proof in an interactive session with three mathematicians.
Meta engineers built a domain expert AI that separates knowledge from reasoning and learns from expert corrections without any model retraining.
A tiny 9,800-parameter Neural Cellular Automaton beats larger recurrent and feed-forward models on few-shot MNIST by reusing hidden-state features across classes.
A mechanistic look inside TaxiGPT reveals a faithful Manhattan map, a goal compass, and superposition interference behind its illegal turns.
A reverse-engineered toolkit lets AI agents drive CapCut's Chinese sibling Jianying entirely from the command line, producing native editable drafts and MP4 exports.
A community build strips refusals from a 27B ternary model by rewriting 2-bit codes in place, with no dequantization and zero size change.
A new theoretical framework decomposes W4A4 quantization error into two parts, yielding a backprop-free recipe that matches gradient-trained SpinQuant.
Convai Innovations released Laya, an open-source non-autoregressive decision model that returns typed answers with calibrated probabilities in ~33ms across 100+ languages.
Pydantic AI now supports Jev, a non-generative classifier that answers typed questions in about 180 ms, turning your output_type into the question itself.
A new open-source Claude Code plugin uses TypeSafe's Jev decision model to prune agent context per tool call instead of summarizing it away.
Shanghai AI Lab open-sourced Atria Dawn Preview, a 744B-parameter Mixture-of-Experts agentic model with MIT-licensed weights and FP8 checkpoints.
Alibaba's new omni-modal model reasons over audio and video, orchestrates tools across long workflows, and cuts video input costs by roughly 89%.
Sakana AI launches the Frontier Intelligence Group, a research collective betting that Transformers and scaling laws aren't the final answer to AGI.
A new preprint reframes recursive self-improvement as dreaming inside a replay simulator built from an agent's own discovery history.
Factory's coding agent joins Slack's new multiplayer channels in beta, letting teams delegate software tasks and get results delivered inside the conversation.
Epoch AI launched Benchmark Reviews, auditing 15 popular AI benchmarks with only 4 passing as Verified and 9 flagged as Flawed.
Claude accelerated over 30 open-source biology models by roughly 4x, letting a single GPU node predict structures once out of reach.
PrismML's ternary-quantized 27B model retains 98.2% of full-precision Qwen3.8 27B performance in a 5.9GB footprint, hitting 143 tokens/sec on an RTX 5090.
Anthropic published three transparency metrics tracking AI-driven R&D, agent oversight, and compute allocation, urging other frontier labs to adopt the same reporting.
OpenAI unveils a legal-specific configuration of GPT-6 Astra with a 230M-URL search index, firm-built workflows, and 73 plugins for practitioners.