Eighty years ago, a machine at Penn proved computation was real. I walked those halls. Now I’m finishing what it started.
The University of Pennsylvania was founded by Benjamin Franklin in 1740 — named not for him but for William Penn, the Quaker who believed a colony could be governed by principle instead of force. Penn’s first principle was that governance precedes settlement. You write the charter before you build the city.
Franklin took that further. He founded a university on the radical premise that practical knowledge — science, commerce, medicine — was as worthy of study as theology. In 1765, Penn established the first medical school in North America. In 1881, the first business school. And in February 1946, in Room 100 of the Moore School of Electrical Engineering, Penn built the first general-purpose electronic computer.
ENIAC. Thirty tons of vacuum tubes. A ballistic trajectory calculation that would take a human twelve hours, finished before the reporters uncapped their pens.
Yesterday was its 80th birthday.
First medical school. First business school. First computer. A Quaker’s governance principle, a polymath’s practical university, and a machine that proved logic could live in hardware — all on the same campus, all in the same lineage.
The Halls
I walked those halls for a decade.
1999 to 2009. MSE in Systems Engineering. PhD in Genomics and Computational Biology under Junhyong Kim. MD. Post-doc in Molecular Ophthalmology with Dwight Stambolian. Ten years at the University of Pennsylvania — the same campus, the same brick corridors, the same institution that built the machine that built the digital age.
Those were my first principals. Kim taught me that biology is computation — that a genome is a program and evolution is its compiler. Stambolian taught me that the clinic is where computation meets consequence. And Warren Ewens — the Warren Ewens, author of the Ewens sampling formula, one of the founders of mathematical population genetics — sat on my thesis committee. The man who proved that neutral drift governs genetic variation sat across a table from me and asked if I understood what I was proving.
I didn’t. Not then. Not for another twenty years.
Ewens’s insight was that most genetic change isn’t adaptive — it’s neutral. Drift, not selection, dominates variation. That same insight now runs through CANONIC’s code evolution theory: at fitness equilibrium, all change is neutral. Drift wins. The mathematics Ewens wrote in 1972 describe the governance dynamics of software in 2026. The professor who grilled me on my thesis gave me the theorem I’d need two decades later.
Systems Engineering taught me the rest: that systems are governed or they fail. Requirements. Specifications. Validation. Acceptance criteria. The discipline of making whole greater than parts — and provably so.
Penn doesn’t announce what happened there. The plaques are small. The students rush past. Moore School sits quiet, the way important things sit — waiting for you to notice. I noticed. Not immediately. You don’t think about ENIAC when you’re debugging PennCNV at 2 AM or studying for boards. You think about it later, when the lineage catches up with you.
I’m a product of Penn. Every neural network, every transformer, every clinical AI system I’ve ever built — the ancestry traces back through Kim and Stambolian to that room. Room 100. First principles. First principals.
What ENIAC Proved
ENIAC wasn’t just fast. It was governed.
This gets lost in the mythology. Everyone tells the speed story — seconds instead of hours, revolutionary, paradigm shift. True. But the deeper truth: ENIAC was built to a specification. Army contract W-670-ORD-4926. Every component had a requirement. Every requirement had a test. The six women who programmed it — Kay McNulty, Betty Jennings, Betty Snyder, Marlyn Meltzer, Fran Bilas, Ruth Lichterman — knew every circuit path because they’d wired the logic by hand.
The specification preceded the machine. The governance preceded the computation.
Sound familiar?
The Arc
1946 — Penn. ENIAC proves computation. Thirty tons of vacuum tubes and an Army contract. Can logic live in hardware? Yes.
2001 — Penn. A medical student starts an MSE in Systems Engineering. Learns that systems are governed or they fail. Doesn’t know yet what he’s learning.
2007 — Penn. PennCNV. First real publication. Genomic copy number variation — the computational biology that maps variation to structure. The same campus where ENIAC mapped trajectories to tables.
2013 — Stanford. First commit on GitHub. Code is a ledger. Work is evidence.
2026 — Orlando. CANONIC proves governance. 255 bits and three primitives. Can you prove your AI works? Yes.
Five points. Three cities. Eighty years. The thread runs through Penn like a wire through a vacuum tube.
How You Discover CANONIC
You don’t invent a governance framework. You discover it. And when you trace the discovery back, every piece was already there — sitting in the halls of a university founded by a Quaker’s principle and a polymath’s bet on practical knowledge.
William Penn said governance precedes settlement. Franklin said practical knowledge is real knowledge. ENIAC proved computation could be governed. Ewens proved drift governs variation. Kim proved biology is computation. Stambolian proved the clinic is where computation meets consequence. Systems Engineering proved that the whole is greater than the parts — and provably so.
CANONIC is what happens when those ideas collide. Not in a flash of invention — in a slow accumulation of insight, the way Ewens said it would happen. Drift. Neutral change. Twenty years of walking halls and writing code and building systems that couldn’t prove themselves, until one Sunday in December the pieces fell into place and governance became compilation.
I didn’t invent CANONIC. Penn taught it to me. It just took two decades of drift before the signal emerged from the noise.
Somewhere between 1946 and 2026, the industry forgot what ENIAC’s builders knew: the machine is only as trustworthy as its governance. They had a contract. They had specifications. They had acceptance tests. They had six women who knew every circuit path because the specification demanded it.
ENIAC moved fast. ENIAC didn’t break things. Because ENIAC was governed.
Happy birthday, old machine. The product of Penn is finishing what you started.
Figures
| Context | Type | Data |
|---|---|---|
| post | audit-trail | items: Penn 1740 → Medicine 1765 → Wharton 1881 → ENIAC 1946 → CANONIC 2026 |
CANONIC — Governed since Room 100.