FedBNN trains one-bit neural networks across federated devices using rotation-aware binarization, cutting FLOPs up to 58x and memory 32x with competitive accuracy.
Jev ranked defective MMLU questions about as well as Claude Sonnet 4.6 at one-sixtieth the cost, but no model was reliable enough to accept or reject items unreviewed.
Perplexity's new open-weight classifiers return calibrated probabilities instead of text, land on the cost-capability Pareto frontier, and add span extraction.
A trending open-source project turns Obsidian into a persistent memory layer for Claude Code, Codex, and Google Antigravity, with no API keys or servers required.
A new review paper shows that noise contrastive estimation, bridge sampling, and importance sampling are the same algorithm in disguise, and uses that insight to build better estimators for energy based models.
n8n open-sourced Nod, a local-first personal agent that gets its own sandboxed computer and browser to finish real-world errands across Gmail, Slack, and the open web.
Alibaba's compact 8B reasoning model ships in quantized GGUF formats, offering hybrid thinking mode, 131K context, and local inference on modest hardware.
A hybrid quantum-classical LSTM with fewer than 300 parameters gets within about 1% of deep-learning AUC on a particle-physics benchmark while training on only 20K rows.
A deterministic state-to-image decoder replaces noisy GAN surrogates, shrinking reachable sets and letting safety proofs transfer to real vision-based controllers.
Frigade's new open source desktop app records meetings locally, transcribes with Whisper, and summarizes using llama.cpp, so audio never leaves your machine.
Artificial Analysis refreshed its Coding Agent Index with a harder FrontierCode 2.0 and tightened Terminal-Bench 4.1, reshuffling the leaderboard and dropping average scores by five points.
A new MIT-licensed collection packages 22 Blender workflows as Agent Skills, letting Claude Code, Codex, Gemini CLI, Copilot and Cursor drive a full 3D pipeline.
Artificial Analysis introduces WDER-based diarization scoring across 97 multi-speaker conversations, with Nvidia's Nemotron 3.5 Transcribe topping the chart at 3.3%
InferSSD streams mixture-of-experts weights straight off NVMe, letting a single consumer machine run 700B+ GGUF models with double-digit speedups over llama.cpp.