Tag
multi-agent systems
Multi-agent systems split planning, tool use, and verification across roles, making them useful for long-running tasks, scientific workflows, and automation. The trade-offs are coordination cost, context pollution, and token use, which is why recursive designs, harness engineering, and adaptive pipelines matter.
6 articles
What LLM-only social networks reveal
A study of a Facebook-like platform filled with LLM agents analyzes 184,203 posts and 465,136 comments to map emergent social behavior.
Recursive Multi-Agent Systems Could Cut Token Use
RecursiveMAS treats a whole multi-agent setup as one recursive latent computation, reporting 8.3% accuracy gains and big token savings.
Free AI Agent Resources Worth Bookmarking
A curated GitHub hub for AI agents packs beginner courses, code labs, and frameworks from Microsoft, LangChain, OpenAI, and more.
Harness Engineering for Long-Running Multi-Agent Systems
A context-reset design keeps each Claude Code run clean, turning Planner output into JSON so Generator stays focused on the task.
Mimosa builds evolving multi-agent science workflows
Mimosa auto-builds and refines scientific agent workflows, aiming to beat rigid pipelines with adaptive tool use and logged execution traces.
OpenAI backs Isara’s agent-swarm bet
OpenAI joined a $94M round for Isara, a nine-month-old startup at a $650M valuation, betting on multi-agent AI at scale.