Cohere Labs releases ATE dataset of ~700K MCP tools to reveal what agents do
Cohere_Labs · x · 2026-09-03
Cohere Labs released the Agentic Task Ecosystem (ATE) dataset, cataloging nearly 700K tools from public MCP servers to study what developers are actually building AI agents to do.
Key points:
- The team argues for "measuring more and predicting less" about AI's impact on jobs.
- The signal: not what AI could do or is asked to do, but which tasks developers choose to automate.
- Their analysis suggests AI's impact on work depends on those choices, not just model capabilities.
- Dataset and blog are publicly available.
More from coding & agent
- Hamel Husain on decision models for evals: don't trust the 0.5 threshold default — HamelHusain · 2026-10-21
- HF Staff Open-Sources Self-Iterating 3D Asset Prompt With Vision Critique Loops for Small Quantized Models — mervenoyann · 2026-10-21
- Open models below 4-bit struggle at self-iterating 3D asset generation, HF dev finds — mervenoyann · 2026-10-21
- Developer Moves AI Agents to a Dedicated Linux Workstation Using Incus System Containers — kevinkern · 2026-10-21
- Motion Launches a General Video Agent That Edits Footage and Mimics Any Reference Style — mhdfaran · 2026-10-21
- MCP Platform NitroStack Wins Product Hunt Product of the Day, Focuses on Retention — EstablishmentFun4373 · 2026-10-21