Agents autonomously ran 12 expert interviews in 2.5 days, says Nathan Benaich
nathanbenaich · x · 2026-11-17
Nathan Benaich shares that his agents autonomously completed 12 expert interviews in 2.5 days, with the entire loop running on Lato AI: framing open questions, designing a study, finding and interviewing experts, updating what's known, refining the questions, and repeating.
The example comes from his State of AI conversation with Matt Turck, and illustrates how agents can realistically replace human-led expert interviews in research workflows.
More from coding & agent
- LangChain webinar: Harrison Chase on the future of agents, Friday Nov 20 — LangChain · 2026-11-17
- Logo.dev MCP Server Lets You Fetch Brand Logos by Name via API — modelcontextprotocol · 2026-11-17
- iPayX FX Audit MCP Connector Detects Hidden Bank FX Markups in Real Time — modelcontextprotocol · 2026-11-17
- Dev: If You Don't Routinely YOLO Big Open-Ended Tasks to Frontier Models, Your AI Priors Are Outdated — brandon_galang · 2026-11-17
- Barbell-tracking app built on custom physics engine: LLMs compute forces in 3 steps — wordgrammer · 2026-11-17
- OpenVorld gives your Claude Code agents a pixel-art village to live in — Ok_Run_5401 · 2026-11-17