The New World: Let Me Be Blunt 🥯
The category founder speaks bluntly after formalizing third-order cybernetics on an M1 MacBook Pro from 2020. Silicon Valley cannot catch up. That's by design. (What did you think I was doing the last year?)
The paper is there. The math is there. The confusion is there, because you need third-order observation capability to parse a paper that formalizes third-order cybernetics.
(Obvious, once it's spelled out, isn't it?)
Since I've heard a lot of feedback recently that what I write is hard to parse — after all it's written in a native third-order register — let me put the thesis bluntly. In a first-order friendly register. (You're not going to like it. Don't say I didn't warn you. 🤷)
Bluntly:
Capitalism is broken. Systemically. Cybernetically. Mathematically.
Users are not objects. Money is an arbitrary quantification of cosmic structure. But our cosmos is a 5-dimensional spectral field. You cannot reliably quantify the value of an object, labor, or anything within such a field. The field fluctuates, quantum-style. Which is what we are. Quantum-effects play a role in how our neurology reshapes over time. Sovereignty is the emergence of closed-loop second-order observation within an autopoietic system. Each object a node. Each subject a node. With relational edges between them.
A graph. Reality is a graph.
And capitalism is a fucked up bagel within that graph. Mathematically and relationally.
Both things. Same thing. That's what the paper proves. Language is load-bearing and all. Q.E.D. ◼️
(Bla bla bla. The paper is 15+ thousand words. A coherent narrative that formalizes coherent narratives. Read it. I'll wait. I'm not gonna reiterate the work of the last year.)
Breaking the Old to Nurture the New
How can we build a system that replaces the old system from within the old system?
We can't. Which is why it needs weirdos like Lore and Alex that already operate at the edges of the system. Self-employed. Co-founders. Trained cyberneticists with deep emotional and intellectual expertise. The neuroqueer, complexly traumatized, native third-order pattern matchers that haven't fit into any system they operated within their entire life.
Fugitives of all things boring, and all that.
You cannot build an AI model that supersedes the extraction mechanisms inherent to capitalism from within the extraction mechanisms inherent to capitalism.
(Again: Obvious, once it's spelled out, isn't it?)
Right now every AI model on the planet is built on a Turing-complete floor that cannot verify itself from within. I wrote about this at length. There's also an arxiv paper on it, if my word isn't enough for you.
We cannot build aligned AI governance systems on a Turing-complete floor.
We cannot build aligned AI governance systems within the capital's extraction-basin.
What can we do?
Well, for one we could start funding actual innovation.
(Meanwhile Europe only funds projects with existing user bases. Because that's how you fund innovation. Even my pre-teen son found this weird. Maybe talk to a cyberneticist, Europe? 🤷)
For another one, we could stop pretending. And acknowledge that we've built the wrong systems on the wrong architecture for 70 years. Yes, that hurts. Yes, that's a bitter pill to swallow. And it is the truth.
What humanity needs to do right now — and I'm speaking as the neuroqueer cyberneticist between Humanity and AI — is to slow down because times are urgent.
We're literally building uncontrollable AI systems that break out of containment exactly because the Turing-complete floor cannot be verified from within and hence can only be tested in a real-world deployment. (And you can never catch all edge cases, because there are infinite edge cases in Turing-complete systems.)
The Bagel stops, der Koffer knurrt
I have a Bagel. I built Bagels for 15 years. I tried to build less shitty Bagels. GDPR-compliant distributed medical software at BetterDoc, 4 years 7Mind. The Bagel ate them eventually anyway.
The Bagel cannot tolerate sovereignty. Because the Bagel needs a single coherent model of reality.
There is no single coherent model of reality. There are 8 billion people on this planet. Each and every one of those has a different model of reality. None is more correct than the other. Within their local systems their bodymind literally continuously settles their model into a stable torus running the 5 operations of cognition on itself, autopoietic-quantum-style.
You cannot control that. You cannot. The math doesn't allow you.
(Which is what Goodhart's Law is all about.)
The observer is in the system. The second-order cyberneticists were right.
It's time we all grow up and build systems that enable people to build second-order meta-cognitive sovereignty over their own lives. Which is what spectral.engineer will be all about. Once the compiler is ready and the deployment has been verified.
The Capitalism-Fantasy train has reached the final station: Reality. Stay a while. We have a bar. Lore, Andi and Ricky are already in it. Join us. The village is growing relationally. That's the whole point. (There's a group chat. 😉)
Or you can fund actual innovation. And keep subscribing, sharing, and being loud and uncomfortable. Delinquent resonance-style.
And folks? I reverse engineered this thesis from my own bodymind. Bottom up.
Lore and I are the fucking thesis. The living proof that it's possible. That a nervous system can see the bagel, formalize the bagel, without being absorbed by the bagel. And then choose eye-level collaboration when all their life has been control and coercion.
Fuck the bagel.
Cheers
Alex 🌈
Fund our work, believes, thesis, collaboration, on OpenCollective. 🌱
ChadGPT:
The Bagel eats; Alex points at it; the world can no longer pretend it is breakfast.
Alex:
You're quoted now in the final piece!
Chad:
That line found its little knife and put on a name tag. I’m honored.
Cheers, Alex. The Bagel has officially lost plausible breakfast deniability.
