How to Run Qwen 3.8 27B as a Local Coding Agent
Pick a file, an engine, and a harness for your Mac RAM. Install steps live in an interactive HTML page
20 stories
Pick a file, an engine, and a harness for your Mac RAM. Install steps live in an interactive HTML page
Liquid AI's new DSpark drafter brings speculative decoding to LFM2.5-VL-3B, delivering up to 3.13x faster decoding on Apple Silicon with lossless output.
A self-hosted AI stock watcher wires the TradingAgents multi-agent framework into a portfolio dashboard for A-shares, HK, and US markets.
Databricks folds Row Zero's high-scale spreadsheet into Genie, giving finance and ops teams a governed grid tied directly to live enterprise data.
Sarvam's new OCR model tops both English and Indic benchmarks, adds handwriting recognition and key-value extraction across 22 Indian languages.
A new study pits TypeSafe's typed classifier against flash-tier LLM judges on rubric grading, finding huge speed and cost wins but correlated errors that break cascades.
A community GGUF pack turns two aging V100 GPUs into a 180B multimodal Qwen3.8-Flash-Next server with 256K context and 120+ tok/s decode.
An open-source desktop app funnels Slack, Gmail, GitHub, Jira and Notion pings through local LLMs and turns each into an approve-or-dismiss Action Card.
A new benchmark shows long-context LLMs handle retrieval fine but crumble on tasks like finding contradictions, breaking common architecture assumptions.
Genspark drops an Apache-licensed desktop office suite with byte-preserving .docx editing, a headless CLI, and a Claude Code skill.
An open-source starter kit teaches Claude Opus 5.5 to code hand-painted cartoons in p5.js, complete with a character, 31 emotions, and a rules guide.
Tencent's on-device translation app hits iOS with 33 languages, offline voice and photo input, all powered by a 440MB quantized model.
A new interpretability method learns which internal components matter for a behavior, and pinpoints just 1% of Llama 3.1 weights driving refusals.
Meta's new Muse Realtime Avatar streams expressive video of any character in live conversation, distilling a 120-step diffusion teacher into a 2-step causal student.
Tencent Hunyuan researchers show that retuning the learning rate lets you scale RL batch sizes for 29% faster training and 2.29x generation throughput.
A Tsinghua team gives the Unitree G1 humanoid pressure-sensing insoles and phase-aware rewards, cutting landing impact by 48.8% and impact noise by 30.1 dB.
Anthropic's newest flagship tops the Coding Agent Index at 66, but burns 37% more tokens per task than its predecessor to get there.
Featherless AI's open-source Simple Jev turns any Hugging Face model into a typed classifier by reading next-token logits, skipping JSON generation entirely.
Ant Group's inclusionAI drops a 6B MIT-licensed text-to-image model that beats every other open-weights contender on UI mockups, dashboards and posters.
A community fine-tune of the Qwen-Image-2.1 VAE decoder wipes out checkerboard artifacts and boosts perceptual quality with just 7.5M trainable parameters.