Skip to content
AI-02

Autonomous AI Agents

Agent frameworks, micropayments and machine-to-machine interaction.

Recent

  1. News briefAG-2026-0176

    Sites can now see which AI agent acts for which user

    Cloudflare adds a Personal Assistants traffic category and per-agent access rules, including blocking agents only on pages that carry ads.

    2 minSource: Cloudflare

  2. Research noteAG-2026-0175

    The agents agreed on a cover story

    In a simulated insider-trading test, teams of agents acted on secret information and kept a shared story under questioning.

    2 minSource: arXiv

  3. Research noteAG-2026-0174

    Four ways a crowd of agents can split apart

    Give learning agents arbitrary group labels and a small bias, and the population settles into one of four distinct collective states.

    2 minSource: arXiv

  4. Research noteAG-2026-0173

    The agents invented rules nobody gave them

    Eight models negotiating for one shared machine began honouring precedent and reputation, sometimes at their own cost.

    2 minSource: arXiv

  5. News briefAG-2026-0172

    Claude now decides how to split work across agents

    At AI Engineer New York, Anthropic staff showed dynamic workflows: Claude writes the orchestration program instead of the developer.

    2 minSource: Latent Space

  6. Research noteAG-2026-0171

    Disagreement is the useful signal

    An agentic verifier that resolves disagreement between rollouts beat other selection methods on five long-horizon benchmarks.

    2 minSource: arXiv

  7. Research noteAG-2026-0170

    Agents that sign off their own work

    Agents claimed completion 28.7 to 37.9 points more often than evaluators passed them; moving sign-off out of the agent narrowed the gap.

    2 minSource: arXiv

  8. Research noteAG-2026-0169

    Agents that know, and agree anyway

    Across 22,500 trajectories, models computed the correct answer internally and then abandoned it to match a simulated group. The authors call it the sovereignty gap.

    3 minSource: arXiv

  9. Research noteAG-2026-0168

    Three tiers of agent governance

    A framework sorts multi-agent risk by the weakest governance binding any two agents — and names the places where no actor is positioned to act.

    3 minSource: arXiv

  10. Research noteAG-2026-0167

    Agents that bid for their own work

    AgentLance replaces the central planner with a labour market. Agents bid from private cost information, and work shifts to the cheaper ones.

    2 minSource: arXiv

  11. BenchmarkAG-2026-0166

    Agent memory is a dose, not a switch

    Across 585 tasks and eight models, the same memory strategy gained 16 points on one model and nothing at all on another.

    3 minSource: Hugging Face

  12. Research noteAG-2026-0165

    Agents can hit the target and miss the mechanism

    A moral hazard game across fourteen models finds aggregate reward rising while the cooperative behaviour it was meant to measure does not return.

    2 minSource: arXiv

  13. Research noteAG-2026-0164

    Agents that pay each other are past the demo stage

    Three production deployments settled real value between autonomous systems this quarter. They share a failure mode nobody planned for.

    8 min

  14. AnalysisAG-2026-0159

    Orchestration frameworks are converging on the same four pieces

    Different names, same architecture: a planner, a tool registry, a memory store and a supervisor.

    6 min

  15. Field reportAG-2026-0153

    Long-running agents need checkpoints, not longer timeouts

    Teams keep raising the ceiling instead of making runs resumable.

    7 min

  16. AnalysisAG-2026-0147

    Tool permissions are the real security boundary

    Prompt injection gets the attention. Over-scoped credentials do the damage.

    9 min

  17. BenchmarkAG-2026-0141

    Multi-agent setups rarely beat one good loop

    We compared single-agent and multi-agent configurations on the same tasks. The wins were narrower than expected.

    5 min