"I had no idea that within 6 months, we'll have a stadium full of weird lobster cultists excited about Agentic computing." Peter Steinberger's quote just about summed up ClawCon, Michigan.
An eclectic mix of ~3000 developers, enthusiasts, founders, and just curious folks showed up in Ann Arbor on Apr 16th , and honestly... it was one of the more gratifying things I've been part of recently. He also said a couple of things that can function as the tl;dr for this post:
(on the new institute.. see below):
"Honestly, kudos to the university. This is very visionary. This is very 2026. We are very rapidly going to a future where everyone will have not one, but multiple agents...."
"We need to solve identity, we need to solve auditability. You know, we as humanity, how do we level up by collaborating and specializing? What if agents will do the same? So there's some really interesting questions we need to figure out."
Much of the public conversation around AI right now runs on fear (nothing seems to "unify" Americans than the fear of AI!). Certainly, some of it warranted and some of it misinformed. It is perhaps hard to be scared of something when you watch 3000 people of every stripe build with it, and demo it to each other in the open, with nobody pretending to know exactly where it's going.
Why am I so excited? This is an amazing time for science and innovation, and one of my greatest excitements in all of this is the following: science is incredibly beautiful.. and I hope with these technologies, we can democratize the understanding of science, the enjoyment of science, and the advance of science. One caveat : AI alone will only pick low-hanging fruit, and not transformative. The necessary ingredients are humans + AI + tools (simulations, experiments, instruments, the accumulated craft of entire fields). AI is just one piece of this. Increasingly though, it is the glue.. the thing that lets people, domain experts, simulations, and experiments actually work together, and agentic frameworks enable exactly this, and can decentralize.
Announcements: We did use the occasion to announce a couple of things:
- The Institute for Agentic Computing - a hub for researchers, developers, and policymakers to drive the advancement of agentic AI technologies, make breakthrough discoveries, steward global access, and address the foundational challenges of governance, alignment, trust, and responsible development that this technology demands.
- The Lobster Compute Company which will invest in companies that are built with OpenClaw and invest some of the captial into the OpenClaw community and platform
Talks & Demos: I spoke briefly on intelligence via social aggregation, and democratizing science. Fiona Wang and Changwen Xu followed with demos on decentralized science and autonomous discovery of superconductors.
Ten other on-stage demos (and many more off-stage ones) filled out the rest of the day - personal productivity apps, robotics, agentic identity, and a few things I am still processing.
Thanks: A number of of people pulled this off:
- Dave Morin & Brit Morin for establishing the OpenClaw Foundation and giving this ecosystem a non-proprietary center of gravity.
- Daniel Feder, whose quiet persistence is the reason U-M is now a serious node - arguably the node - for agentic computing in the academy.
- Nadia McIntosh, Nick Moroz and the events crew, who turned "let's have a big conference" into an actual big conference, on a college campus, on short notice, without any of it falling apart.
What's next: No one can predict how the agentic computing landscape evolves. Rest assured, we'll be doing interesting things... at scale :)
