Every processed story in chronological order, with the newest coverage first. Filter by tag, source, or score to drill in.
The post demonstrates a concrete path from single-agent discipline to parallel multi-agent orchestration, showing how the author's own role contracted from writing code and reviews to tuning workflows — a practical illustration of what the "conductor" layer of agentic development looks like in practice.
The benchmark shows that for autonomous coding agents, the choice between GLM 5.2 and MiniMax M3 reduces to a concrete cost-accuracy tradeoff: GLM's correctness edge is real but narrow and concentrated in greenfield packaging, while MiniMax delivers nearly the same results on modification tasks at roughly one-third the cost and half the latency.
ENPIRE removes the need for continuous human supervision and manual algorithm engineering — identified in the paper as the central bottleneck in physical robot learning — by giving coding agents a fully automated, closed-loop path to self-improve real-world manipulation policies.
Redteam introduces a human-gated, dual-model review structure that directly counters the single-model blind spot of an AI both writing and approving its own code.
The paper resolves a contested debate by showing that guidance production method — not guidance presence alone — determines whether `AGENTS.md` files help or hurt coding agents, and provides a concrete tuning procedure that raises SWE-bench Verified resolve rate by 7.5 percentage points over an unguided baseline.
The RATs framework demonstrates that self-directed play — without any explicit task instructions — can build a reusable, transferable code skill library that improves both in-distribution and real-world robot task performance without retraining the underlying model.
The skill replaces subjective style guidance with empirically weighted pattern detection, giving Claude a data-driven basis for avoiding the specific design defaults that real users most frequently identify as markers of AI-generated UIs.
Kimi K2.7 Code delivers substantial benchmark improvements over its predecessor while cutting reasoning token usage by 30%, making a capable open-weights coding model more efficient and freely accessible.
Iris replaces screenshot-based or assumption-based verification with runtime evidence from a live app, giving coding agents a concrete, structured verdict on whether their changes actually worked.
The benchmark reveals that frontier AI models — including those augmented with Code Agents — effectively fail at large-scale game project engineering, with runtime pass rates collapsing to 5.7%, exposing architectural design as an unsolved bottleneck that compilation-focused improvements cannot address.