Google DeepMind Says Transformers Are Geometrically Broken at State Tracking
A Google DeepMind paper argues transformers are architecturally incapable of true state tracking, and makes the case for a return to recurrence
20 stories
A Google DeepMind paper argues transformers are architecturally incapable of true state tracking, and makes the case for a return to recurrence
A new paper maps Binary Spiking Neural Networks to causal logic models, enabling SAT/SMT solvers to produce provably sound explanations that outperform SHAP.
Luminol-AIDetect detects AI-written text by shuffling sentences and measuring how perplexity shifts — achieving 17x lower false positive rates than prior methods across 18 languages
A new paper from UC Berkeley and TU Munich shows the Platonic Representation Hypothesis collapses when tested at realistic data scales, challenging a foundational belief in multimodal AI.
FMRG reformulates generative model guidance as deterministic optimal control, achieving 70x speedups over prior art with as few as 3 network evaluations
Researchers pit DeepSeek and GPT agents against each other in a supply chain simulation to reveal how mixed AI reasoning levels amplify the bullwhip effect — and how information sharing fixes it.
EvoOR-Agent treats LLM agent workflows as evolvable graphs, letting the architecture itself adapt to each optimization problem instead of following a fixed pipeline.
MedSynapse-V embeds clinician-like diagnostic memory directly into a model's hidden states, beating chain-of-thought methods without adding inference cost
A new framework trains RL agents for chip placement by learning reward signals directly from expert layouts, closing the quality gap with human designers
TrafficClaw is the first LLM agent framework to control traffic signals, freeways, transit, and taxis together as one coupled physical system
KnowledgeBerg exposes a hidden flaw in LLMs: even top open-source models score below 45% on questions that look simple but require complete domain knowledge and multi-step set reasoning

OpenBB's open-source financial data platform unifies 30+ data providers into one Python SDK, now with MCP server support for AI agents and a free community tier

An open-source stack combining Whisper, streaming diarization, and 200-language translation into a self-hosted, ultra-low-latency transcription server.

Alibaba's Qwen team ships a 9B open-weight model that matches gpt-oss-120B on reasoning benchmarks while running on a laptop, natively multimodal.

A hands-on repo pairs an O'Reilly book with 20 chapters of GPU kernel tricks, PyTorch tuning, and trillion-parameter inference scaling recipes.

Mistral folds chat, reasoning, and coding into one 128B dense open-weight model with 256K context and a per-request reasoning toggle.

Anthropic's Claude Cookbooks hit 51k GitHub stars with 80+ runnable notebooks covering agents, tool calling, RAG, and production patterns

A Qatar Computing Research Institute team trained the first Arabic-centric self-supervised speech model, then squeezed it down to 28M parameters while beating HuBERT and XLS-R.

TimesFM 2.5 cuts parameters by 60%, expands context 8x to 16K, and lands directly in BigQuery SQL — no ML setup required

NVIDIA's free, open-source GPU learning hub now covers CUDA Tile, nvmath-python, and Warp — all runnable in the cloud with one click