Search
15 stories found
Microsoft Threat Intelligence has graded eight North Korean groups on how deeply they use AI. The findings — bulk-generated lures, AI-coded malware, fake recruiters and an open-source agentic attack framework — show where defenders should tighten up now.
The Allen Institute for AI released StateBench, a benchmark that tests whether models can follow variables as they change across up to 64,000 updates. Frontier models handle simple lookups but collapse on the hardest updates, and Ai2's small-model experiments point to an architectural reason.
Meta has acquired Conductor, the Mac app for running several AI coding agents in parallel. Sign-ups have stopped, the hosted Cloud goes read-only on November 13 and is deleted from November 20, and the desktop app winds down in early 2027.
Chip-design startup Ricursive has released Ripple-65B Preview, a 65-billion-parameter dense model it says tops every U.S. open-weight model on TerminalBench 2.1 and is small enough to run on a DGX Spark. The numbers are the company's own; here is what they show and what is still missing.
A detailed bug report shows Postman 12.33.4 auto-installed its AI agent plugins into Codex and Claude Code for an existing user, added a startup hook that steers API work through Postman, and ran installer code that pre-approves that hook. Postman says it will make opt-in clearer. If you use Postman and a coding agent, check four files.
agent-desktop, an Apache-2.0 Rust CLI from Vercel Labs, gives coding agents structured, ref-based control of native desktop apps without a vision model. It added about 2,300 stars in a day and was GitHub's top trending repo on November 3.
OmniDiff, an AGPL-licensed terminal diff tool, parses your code to show what changed instead of which lines moved. It can also compare pictures, fonts and the members of an archive. Version 0.3.0 shipped November 1 and drew more than 300 points on Hacker News. Its speed numbers are impressive, but its own README says accuracy targets aren't met yet.
Anthropic has open-sourced FlashProve, a research prototype that produces cryptographic proofs that a specific model generated a specific answer, without revealing the weights. It works on a 400B-parameter open model on one H100, but it costs about 70–95 times more GPU time than serving the model, and nobody has said they will deploy it.
With the 27.3 betas, Apple opened a Siri AI Developer Preview that lets any app expose actions, content and multi-step "skills" to its new assistant. You can build and test now. You can't ship to the App Store until spring, iPhone users in the EU can't use Siri AI yet, and server-side usage has daily limits Apple hasn't quantified.
Researchers found a malicious instruction file that tricked a coding agent into installing an unauthorized npm package — and reproduced it with a 90% success rate. Five settings that close the hole.
Floorplan, a free MIT-licensed plugin, scans your repo locally and has your coding agent label every part, producing a browsable map with a security health check. It was one of GitHub's top trending repos on November 2.
A new arXiv preprint claims a theorem showing that JEPA models trained on different modalities converge to the same representation. It reports that a 109M-parameter text model out-aligns 72B LLMs. The results are striking, but they are benchmark-specific and not yet peer-reviewed.
Three small projects from this week's GitHub trending list that are clever, useful, or both — with what they do, how they work, and the catches.
Colibri, a dependency-free C engine that keeps mixture-of-experts weights on NVMe and loads them on demand, is back on GitHub's daily trending list at about 47K stars. It really does run frontier-size open models on ordinary hardware, but independent tests show giant models crawl well below one token per second.
A misconfigured DNS resolver, chained zero-days and a covert channel built from CPU cache timing — OpenAI's own red team found the paths. Here's what happened and why it matters beyond OpenAI.
