Pay attention to this one, AI devs. If you're building multi-agent systems, you're probably wiring static org charts. New research argues they should look more like a labor market. The paper introduces OneManCompany (OMC). Instead of fixed teams, it defines "Talents," portable
@dair_ai
-
AI Agents Token Costs: Systematic Study on SWE-bench
By
–
How do AI Agents spend your money? Most teams treat agent token costs as a rounding error even though the data says they shouldn't. New paper presents the first systematic study of how agents actually spend money on coding tasks. They ran 8 frontier LLMs on SWE-bench Verified
-
Top AI Papers of the Week: April 19-26
By
–
The Top AI Papers of the Week (April 19 – 26) – Skill-RAG
– DeepSeek V4
– Autogenesis
– Attention to Mamba
– Stateless Decision Memory
– Self-Evolving Logic Synthesis
– Self-Generated World Knowledge Read on for more: -
PARE Framework Evaluates Proactive AI Agents Anticipating User Needs
By
–
Great paper on improving proactive agents. (bookmark it) Proactive agents act before you do. But how do you evaluate something that's supposed to anticipate needs you haven't expressed? This work introduces PARE, a framework that models applications as finite state machines
-
Structured Memory Systems for Long-Horizon LLM Behavior
By
–
Good agent memory paper. And great insights on the benefits of structured memory for long-horizon behavior in LLMs. Why it matters: It treats memory less like search and more like a system that will need maintenance (which they often do). Flat memories are cheap to write.
-
Multi-Agent LLM Systems Show Diversity Collapse Over Time
By
–
Cool paper on diversity collapse in AI agents. It's a common issue with all the deployed multi-agent systems. New paper shows that multi-agent LLM systems converge on near-identical outputs over time, even across different architectures and different starting prompts. They call
-
Building Production-Grade AI Agents: Essential Reading
By
–
Worth a read if you are building production-grade AI agents.
-
Agents That Self-Generate World Knowledge via Outcome-Based Rewards
By
–
How far are we from agents that can self-generate world knowledge? The work proposes an outcome-based reward that measures how much an agent's self-generated world knowledge actually improves its task success rate. The external guidance is then removed at inference. Result: A
-
Survey on Multi-Agent Systems from Classical to LLM-Based Paradigms
By
–
// Survey on Multi-Agent Systems // The paper traces the landscape from classical paradigms (consensus, distributed control, swarm intelligence, cooperative learning) to foundation-model-enabled MAS (LLM-based planning, role specialization, task decomposition, multi-modal
