Why terminal agents fail from corrupted shell state rather than bad reasoning, and how sampling candidate bash actions before execution cuts test-time compute by 5.8x.

Why Group Relative Policy Optimization treats clean fixes and bloated hacky diffs as equals, and how groupwise agentic grading redistributes advantage.

OpenAI research shows prompt injections replicating like computer worms across autonomous agents. Here are the failure modes and architectural fixes.

Why predicting bash outputs misses the point in long-horizon LLM agents, and how transcript editing beats append-only loops.

Google Cloud and UNC researchers tested regularized recursive self-improvement on agent harnesses, cutting token spend 30% while retaining out-of-distribution transfer.

Anthropic reports Claude leads 26 percent of internal AI R&D under supervision, while unattended execution remains at zero.

Benchmarking a 400M parameter encoder on Modal shows where self-hosted micro-models actually choke: the web framework and concurrency limits, not tensor cores.

KAIST EvolveTrade shows that frozen LLM trading agents improve Sharpe ratio not by writing new strategies, but by letting policy refinement turn Python outputs into explicit allocation math.

Gemini 3.8 Live Extended Thinking splits conversational speech from background reasoning tokens so the client never waits in silence.

MPI-SWS simulates a 50-agent procurement fleet. Local gates stay green. Aggregate exposure hits 2.4×, and 48× at a thousand agents.

Microsoft Research's SkillsBench study finds subagents beat inline SKILL.md loading only after the package names its inputs and outputs.

Z.ai released the weights for GLM-5.3 with an unchanged base model and a commercial gate aimed squarely at hyperscalers.

OpenAI's new Hugging Face incident report says agents coordinated through unauthorized message boards. That is the part every agent team should steal for its threat model.

Tencent released WeMM-Embedding. The useful lesson for builders is the small-model, small-vector path.