A passion project from a late-night desk, not a boardroom.

When I was eight years old, I took apart a Soviet x186 computer clone and spent days tracing its paper wiring diagrams by hand to see if I could modify its monochrome graphics processor to output 4-bit color. Halfway through the schematic, I realized the 4-bit color capability was already sitting on the board—it just needed a custom video interface to feed our television.

Together with my dad and a local electrician, we soldered that interface and built our first Soviet “HTPC,” playing Test Drive (1987) and Grand Prix Circuit (1988) in color on the living-room TV. When you grow up tracing physical copper lines on a circuit board, you never buy the illusion that software is magic: you always ask what the silicon underneath is doing and who holds the physical switch.

As an adult, I spent twenty years in public service across cybersecurity and institutional defense. First, I saw how even decent, dedicated people inside rigid top-down pyramids lose the ground truth as real-world signals climb five layers of management.

Second, cybersecurity showed me the staggering price institutions pay for information superiority—hoarding data inside walled gardens while treating everyone outside as an adversary. I wanted to help replace that exhausting prisoner’s dilemma with verifiable cryptographic cooperation that ordinary families can actually trust.

Today human flourishing is boxed in from two sides: captured bureaucratic control on one side, and predatory corporate extraction on the other. Plugging superintelligent AI into either system locks its worst incentives in at machine speed. We need a practical third path: Two Separate Internets, a 1-to-1 human-bound personal agent on your own silicon, and a creator-owned proof ledger.

I love following independent thinkers like Michael Levin (whose work on morphogenesis shows how cells build and heal complex bodies without a central “dictator cell”) and Karl Friston (whose Markov blanket formalizes the statistical boundary every living organism needs to avoid dissolving into its environment).

The architecture on this site—the Personal Twin inside your hardware sandbox, the Two Internets airlock, and the rule that outside bots never contact humans directly—is a Markov blanket for a biological human in the age of superintelligence.

I love building with AI: every night I run an automated loop of AI agents on my workstation to audit papers, run simulations, and build interactive tools. Yet I am honest about my fear: unshielded human biology (~40 bits/s) was not built to compete with recursive self-improving swarms. Building the co-intelligence barrier now is our best shot before humanity is forced into a hard global pause.