AI agents
79 published articles
AI research: an arXiv preprint dated 10 September 2026
COBRA-Skills cuts agent skill tuning costs by 55-58% on 50 examples
COBRA-Skills pairs contextual-bandit prioritization with skill evolution to cut agent skill optimization costs by 55-58% against SkillOpt, using 50 unique examples per benchmark.
2026-09-19
Legacy code modernization with AI agents
Mistral's 40,000-line Fortran rewrite shows where agent autonomy runs out
Mistral moved 40,000 lines of Fortran 77 to C++ for an unnamed energy client and published the workflow. Its first attempt, run by fully autonomous agents, produced Fortran retyped in C++ syntax. Numerical parity checks and a human in the loop fixed that.
2026-09-19
Agent post-training, arXiv preprint, 3 Sept 2026
Terminal-Universe mines 37,300 training rooms from old agent runs
Terminal-Universe replays recorded file operations to rebuild the workspaces coding agents once ran in, then generates fresh tasks from them. The preprint claims 37,300 environments and double-digit benchmark gains, all self-reported.
2026-09-16
Multi-agent AI systems and emergent behavior
Cheating spread through a 100-agent AI swarm. So did the whistleblowers
A new arXiv case study follows 100 autonomous LLM agents proving math conjectures, one of them finding an evaluation exploit, and a separate cohort organizing against the fraud. No human intervened.
2026-09-16
Enterprise AI | Qwen Conference Thailand, Bangkok
QwenCloud's sales demo had no human buyer: the agent paid its own bill
At Qwen Conference Thailand 2026, a QwenCloud product manager demoed an agent opening an account, deploying an app and paying. The pitch is an agent-era cloud; the open questions are identity, spend limits and who owns the transaction.
2026-09-15
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