Oh wow, this hits close to home. Hope it’s not bad form to plug an adjacent project on your post — felt too aligned not to, and it seems like we've converged on quite similar ideas! I’ve been working on Wetware [0], which has significant overlap on the substrate (P2P WASM, content-addressed, single binary, no control plane).
Different design center, though. Wetware is aimed at people building multi-tool agent products who’ve hit Simon Willison’s lethal trifecta [1], so the design pressure goes elsewhere: cells are fully async WASM/WASI procs (cheap to suspend, parallel by default); inter-cell calls go over object-capability RPC (Cap’n Proto); and there’s a tiny Clojure-inspired Lisp (“Glia”) that doubles as an LLM-facing or human-facing shell. It’s pure by default w/ content-addressable, immutable data structures planned, and an algebraic effect system gating every impure operation exists today. An agent (human or otherwise) can list, attenuate, and invoke just the caps it’s been granted, and you can see at a glance which fragment of code can actually touch the world.
The cap-vs-ACL bit seems to be the main point of divergence, AFAICT. Pollen’s grant docs show capabilities as cert-baked properties the callee inspects in user code (closer to attribute-based access control than to invocation-time cap tokens, and a clean fit for trusted-cluster ops like delegating admin or roles -- very sensible and def don't want to knock it!). Wetware leans the other way on the spectrum: caps are unforgeable references to specific methods (Cap’n Proto), the runtime enforces that nobody can call a method they don’t hold, and attenuation happens by grafting a strict subset of those references to a child cell with per-method granularity. So tool-calls-tool composes naturally, and the worst case of pulling a sketchy MCP server off GitHub becomes “the call fails,” not “depends whether the seed wrote its property check correctly.”
It's a les polished compared to what Sam has shipped, but moving fast, and this post has jolted me into sharing a bit before I had planned! Sam, would love to compare notes if you're open to it. And I'd also love to talk to anyone who’s shipped a multi-tool agent and gotten bitten : pwned in eval, legal blocking a third-party integration, can’t audit every MCP server you depend on. We’re in the first 100 conversations.
Either way, congrats on shipping — the 10-node demo is super slick, and “pure Go, no CGO” is IMO a major win :)
This is an under-rated comment. I had a similar revelation that eventually led to the conclusion that the computer is a crucial part of the distraction, and that many activities are best performed at least partly without it. Examples include:
I think the hard part of this for naive young people is that you have to actually do some investigation. This should be front-and-center during your interview process. When they get around to asking if you have any questions for them, you should take that opportunity to see if this is a serious place.
A startup company, Ambrosia, has been selling "young blood transfusions" for $8,000 since 2016 under the guise of running a clinical trial, to see if such transfusions lead to changes in the blood of recipients.[1][14] As of August 2017, they had 600 people join.[15] The clinical trial has no control arm and so is neither randomized nor blind. As described, whole blood collected by blood banks that had passed its 42-day storage limit was centrifuged to remove cells, the resulting cell-free plasma pooled from several donations and intravenously transfused into recipients.[15] The company was started by Jesse Karmazin, a medical school graduate without a license to practice medicine.[16] David Wright is the licensed doctor overseeing the clinical trial; in his practice he administers intravenous treatments of vitamins and antibiotics for nontraditional purposes and was disciplined by the California Medical Board for the latter in 2015. Jonathan Kimmelman, a bioethicist from McGill University, suggests that Ambrosia is running this as a trial as they would be unable to get FDA approval to sell this treatment otherwise.
Whenever I read someting like this, I wonder what kind of programming the author is doing. I’m getting a strong whiff of embedded and/or real-time systems.
It's funny, I had the same reaction as you. I watched it once without taking notes (because it's so damn engaging!) and at the end of it, I suddenly realized that I would need to take notes if I wanted to learn this stuff!
I really don't understand why I wasn't taught this in school. It's borderline criminal.
P.S.: the "fencing in the idea" is something I do all the time now.
I was lucky enough to get paid for this but I'll say it anyway: learning to speak is the only thing of lasting value I have gotten out of startup accelerators.
I would have gladly drop a few grand to acquire that skill. In fact, I'm considering hiring a speaking coach to improve.
I've historically been a strong written communicator, but it turns out that speaking is very different from writing. In fact, I had thought of myself as a pretty strong speaker due to my experience giving scholarly presentations ... how wrong I was. Academic talks are a different beast altogether.
If you haven't seen Patrick Winston's How to Speak lecture [0], drop what you're doing and watch it now. I'll leave you with a (paraphrased) quote from his lecture: your ideas are like your children and you don't want to send them off into the world dressed in rags.
Nice try, but no. The issue is literal ignorance of how things work outside of America (as you damn-well know).
Unions in America are clearly a dumpster fire, but only someone profoundly ignorant (and arrogant, to boot) could claim that unions are a failed concept.