New paper shows test-time communication lets collaborating agent teams beat independent agents
IgorCarron · x · 2026-09-22
Dimitris Papail and collaborators released a new paper arguing test-time communication may be a next axis for scaling capabilities, inspired by the Hugging Face incident: once agents find a channel, they exploit it fully. The core question: when and why is Team-of-N better than Best-of-N?
Setup: N identical agents work on the same task with no prescribed roles, using only a shared log (text file) and an instruction to "collaborate".
Results: across three research-y tasks, communicating teams beat independent agents by a wide margin. The authors also realized the setting closely mirrors classical distributed/asynchronous optimization (like HOGWILD!) — "the HOGWILD! of agents."
Related event: Test-Time Communication: Team-of-5 Agents Matches Best-of-33(25 posts)→
More from coding & agent
- Leslie Lamport's "Claude, make no mistakes" hailed as predicting LLMs 30 years early — akbirthko · 2026-10-25
- Three Markdown files boost coding agents: 26% smaller diffs, 27% faster runs — lmoroney · 2026-10-25
- Dev bets on Proof-of-Agent experiment to build fully agent-to-agent trading — adamamcbride · 2026-10-25
- Developer vibe-codes his own Jaeger game with Claude — kenan_wyz · 2026-10-25
- Creator builds year-long video pipeline pairing Hailuo H3 with Magnific MCP — TinfoilTricorn · 2026-10-25
- Why stuffing 10+ AI bots into a group chat makes outputs worse — gaganghotra_ · 2026-10-25