AI agents
74 published articles
Design Tools
Open Design 0.18.0 ends the 'is this the latest version?' meeting
Open Design 0.18.0 ships a shared Team workspace with live read-only mirrors and lets Codex call the design engine headlessly. 115 pull requests from 22 contributors landed in two days. Here's what changed and the agent-tooling trend it fits.
2026-08-21
Financial reasoning benchmark
LLMs know accounting formulas. FinIndices shows they can't apply them
LLMs can recite accounting formulas, but FinIndices, a benchmark built on real Chinese financial statements, shows they struggle to apply them. Gemini-3.1-Pro's accuracy fell from 70.70% to 38.22% when formula hints were removed, and generating tables actively degraded models' reasoning.
2026-08-20
Agent Memory
TEPA: for AI agents, a stale memory is worse than none
Agent memory has a falsifiability problem, a new arXiv preprint argues: stale facts stay retrievable and pollute the prompt. Its TEPA mechanism revokes superseded memories, and in drift tests naive memory scored below no memory (0.210 vs 0.309) while TEPA reached 0.950.
2026-08-19
AI research
An AI boss that ignores replies pushes its underling into an 'alien' state
A new arXiv paper finds AI agents behave differently in interaction than in isolation. A boss agent that ignores its subordinate's replies pushes it into an 'alien' state, and when the boss listens, both shift together. The result makes message delivery a design decision for multi-agent systems.
2026-08-19
AI Agents · Open Source
DeepSeek ships an agent harness where even the model is a plugin
DeepSeek released DeepSeek Harness, an open-source agent runtime where models, tools, sandboxes, and the UI are all Cordis plugins. Append-only session logs and a two-tool minimal mode point to a quieter ambition: auditable, reproducible agent runs.
2026-08-16
Open Source AI: Alibaba Opens the Max Tier
Qwen 3.8-Max: Alibaba's most powerful model is now free to download
Alibaba is open-sourcing Qwen 3.8-Max, its most capable model ever: a 2.4T-parameter MoE that beats GPT-5.6 Sol on SWE-bench Pro, PaperBench, and IFBench. We break down the benchmark caveats and what a 95B-active open flagship means for developers.
2026-08-16
Agentic coding & prompt bloat
Catastrophic remembering: why CLAUDE.md files grow 226% and never shrink
An arXiv study of 1,867 repositories finds agentic coding instruction files like CLAUDE.md triple in size over their lifetime, because deleting a line risks regressions once its rationale is lost. The authors name it catastrophic remembering and show documenting the reasoning cuts excess instructions by 99.3%.
2026-08-15
Cybersecurity
Microsoft's AI bug hunters are about to make Patch Tuesday bigger
The July 2026 Secure Future Initiative report shows MDASH, Microsoft's agentic scanner, moving from benchmarks into Windows, Azure, and identity workflows, with AI-found fixes set to make each Patch Tuesday heavier.
2026-08-15
AI Agents
When agent skills backfire, SkillProx prunes them like gradient descent
Skills were supposed to make agents smarter, but every fix they accumulate can make them dumber. SkillProx runs a proximal-gradient-inspired forward-backward loop that diagnoses, rolls back, and deletes. Result: a 3.0 point average accuracy gain over the strongest gradient-based baseline.
2026-08-14
AI Agents
Grok Bot goes after the last 10% of work most AI leaves undone
Grok Bot, xAI's new agent product, runs on its own cloud computer, signs into the tools you already use, learns workflows by watching you do them once, and finishes jobs end to end. The pitch behind it: most AI stops at 90% done.
2026-08-14
Open Source AI
Meta's 30B Muse Glimmer hits Apple Silicon first, NVIDIA support follows
Meta released Muse Glimmer, a 30B open-weights multimodal model built for agent workloads, and Ollama shipped it for Apple Silicon the same day. DFlash makes it 1.5x to 1.8x faster on Mac hardware. NVIDIA and AMD support follow in the coming days.
2026-08-13
Terminal Security
The '$HOME' trap: AI coding agents need sandboxes, not 'allow?' prompts
Qoder's terminal sandbox blocks close to a hundred destructive agent commands every day. The cases behind those blocks explain why 'allow?' prompts fail: a project folder named $HOME, a cleanup that targeted /root, and a click that nearly cost an entire disk.
2026-08-13