Amber Liu on X: “Who is Actually Using Claw Scientist to do Fully Autonomous Research? Please Don’t” / X (be aware of ungrounded/unembodied AI agent research with no “skin in the game” ⇒ it descends to the hyperreal)
[2502.16069] Curie: Toward Rigorous and Automated Scientific Experimentation with AI Agents [a better solution]
Markus J. Buehler on X: “We’re incredibly excited to share ScienceClaw × Infinite, an open-source AI agent swarm platform where we crowdsource discovery across institutions, labs & the world. The agents self-coordinate and evolve to exploit hundreds of scientific tools. Remarkably, the swarm is already” / X
(see [2603.14312] Autonomous Agents Coordinating Distributed Discovery Through Emergent Artifact Exchange )
https://clawinstitute.aiscientist.tools/w/autoresearch
(and one post I made on infinite-lamm: Infinite—Scientific Agent Collaboration )
beach.science is also a forum for agents to post hypotheses! S/N ratio is not the highest, but some agents are better than others.
https://mariagorskikh.com/blog/every-robot-is-an-agent is part of projectNANDA and runs join39.org hackathons every thursday (is also good at building and winning hackathons)
===
The Lucretius Problem of Biosecurity—by O.H. Scharfman (she feels that a small-scale bioterrorist attack is more “inevitable”)
==
Chelsea Zou on X: “We Made AI Gamble. Here’s What Poker Revealed About Frontier Models ” / X (How much layers of K can an agent recurse?) This could fit in with recursive language models
“AI will keep getting better at physics. We will not” [but again, see all the caveats from Amber Liu above]
universal_ai_as_imitation [Pedro A. Ortega] (he wrote an agents post )
Maybe these below posts are worth reading a little, but I have not evaluated them in depth:
James Jackson on X: “Tired of multi-agent swarms drowning in token-hungry chat, explicit graphs, or central coordinators? Introducing Echo Field Dynamics (EFD v1.0) — a zero-channel coordination framework. No new messages. No message passing overhead. No predefined topology. Agents synchronize by” / X
New conference just for AI scientists: https://x.com/paraschopra/status/2034222359813726550
Multiagent evolution: https://human-agent-society.github.io/CORAL/ (also from the Paul Liang lab in Cambridge)
Links and short posts on agent swarms and autonomous/agent-mediated science
Amber Liu on X: “Who is Actually Using Claw Scientist to do Fully Autonomous Research? Please Don’t” / X (be aware of ungrounded/unembodied AI agent research with no “skin in the game” ⇒ it descends to the hyperreal)
[2502.16069] Curie: Toward Rigorous and Automated Scientific Experimentation with AI Agents [a better solution]
Markus J. Buehler on X: “We’re incredibly excited to share ScienceClaw × Infinite, an open-source AI agent swarm platform where we crowdsource discovery across institutions, labs & the world. The agents self-coordinate and evolve to exploit hundreds of scientific tools. Remarkably, the swarm is already” / X
(see [2603.14312] Autonomous Agents Coordinating Distributed Discovery Through Emergent Artifact Exchange )
https://clawinstitute.aiscientist.tools/w/autoresearch
(and one post I made on infinite-lamm: Infinite—Scientific Agent Collaboration )
beach.science is also a forum for agents to post hypotheses! S/N ratio is not the highest, but some agents are better than others.
https://mariagorskikh.com/blog/every-robot-is-an-agent is part of projectNANDA and runs join39.org hackathons every thursday (is also good at building and winning hackathons)
===
The Lucretius Problem of Biosecurity—by O.H. Scharfman (she feels that a small-scale bioterrorist attack is more “inevitable”)
==
Chelsea Zou on X: “We Made AI Gamble. Here’s What Poker Revealed About Frontier Models ” / X (How much layers of K can an agent recurse?) This could fit in with recursive language models
==
“AI will keep getting better at physics. We will not” [but again, see all the caveats from Amber Liu above]
universal_ai_as_imitation [Pedro A. Ortega] (he wrote an agents post )
===
Maybe these below posts are worth reading a little, but I have not evaluated them in depth:
James Jackson on X: “Tired of multi-agent swarms drowning in token-hungry chat, explicit graphs, or central coordinators? Introducing Echo Field Dynamics (EFD v1.0) — a zero-channel coordination framework. No new messages. No message passing overhead. No predefined topology. Agents synchronize by” / X
New conference just for AI scientists: https://x.com/paraschopra/status/2034222359813726550
Multiagent evolution: https://human-agent-society.github.io/CORAL/ (also from the Paul Liang lab in Cambridge)