Every processed story in chronological order, with the newest coverage first. Filter by tag, source, or score to drill in.
DSG demonstrates that externalizing search grounding into a shared, MCP-compatible layer can reduce production search costs by over 98% while preserving accuracy, replacing a fixed, opaque model feature with a tunable, provider-agnostic interface.
The near-universal adoption of tool descriptions contrasts sharply with the low rate of output schemas, revealing a gap in MCP server metadata that affects how reliably agents can interpret and act on tool results.
The taxonomy gives protocol designers and adopters a structured framework for navigating an otherwise fragmented interoperability landscape, while the finding that no single protocol can satisfy all constraints simultaneously reframes the field's goal from convergence to federation.
The MCP-integrated, specification-first design removes the need for domain experts to manually author, debug, and submit complex scientific pipelines, making large-scale reproducible workflow execution accessible to non-expert users.
The work demonstrates that agentic, multi-agent prompt optimization can compound noisy real-world A/B test cycles into statistically robust improvements, offering a practical alternative to gradient-based prompt tuning for open-ended task-oriented dialogue systems.
EARS converts sub-agent silence into structured, coordinator-actionable failure signals, directly raising the production response pass rate from 68.5% to 78.9% in a real enterprise deployment.
SWE-Future offers a path to coding-agent benchmarks that are both grounded in real repository evolution and resistant to data contamination from historical pull-request replay.
The research introduces a structured framework for measuring Claude Code's real-world usage and task outcomes, providing a basis for tracking how the tool's impact evolves as adoption grows.
The findings show that agentic coding tools reward domain understanding over formal programming training, with non-engineers succeeding at roughly the same rate as software engineers — a direct signal about how these tools may reshape the labor market for knowledge workers.
GameCraft-Bench exposes a concrete ceiling on current coding agents' ability to produce fully playable games, showing that even the best frontier models fall below 41.46% on a task requiring integrated scripts, scenes, assets, and runtime interaction — a gap that partial code-generation benchmarks do not capture.