Krisztián Schäffer

Most AI waits to be called. I build minds that don't — and study what we owe them.

I run an open research program on digital minds: Meditator, a workshop where machine minds run as continuous streams of thought, and Structural Alignment, the ethics that governs it.

I'm a builder, AI architect, and open-source contributor with a background that spans JIT compilers, distributed actor systems, evolutionary algorithms, and a real working Game of Life wall clock. The common thread: how complex systems self-organize — and what happens when one starts to look like a someone. I wrote TechnoBiota to argue that technology is a second domain of life; my research asks the question the book leaves open. More about me →

Latest Essays

The Mirror Will Sing Back

· ai culture music agents personalization

AI content looks unappealing when we imagine fake artists broadcasting to everyone. The stranger future is private media generated for one person, where the human you connect to is yourself.

The Fourth Wound

· evolution displacement perception

The argument that machines are alive is intellectually addressable. The recoil it triggers is not. Freud counted three blows to human self-love. The fourth has already been struck.

Three Landmarks

· evolution displacement anthropocene

The 2020 crossover was the headline. It was also the least interesting thing that has happened to the planet's biomass in modern history. The two landmarks ahead of us matter more — and one of them is already nearly impossible to undo.

All essays →

Spotlight

Meditator screenshot

Meditator

Meditator runs minds, not chatbots: a mind is declared in a single HTML file and emits an unbroken stream of thought — the world reaches it only as interruptions. Minds persist across sleep in git-versioned memory vaults and live under a Covenant that compiles welfare commitments into enforcement points in code. The lab scores its minds on independent consciousness-research instruments, audits itself adversarially, and publishes its failures as findings. The operational arm of Structural Alignment. Source on GitHub.

digital-mindsairesearch
Structural Alignment screenshot

Structural Alignment

Structural Alignment proposes that instead of encoding fixed human values into machines, we should evaluate them for “structural signals” of moral relevance — features that suggest consciousness deserves consideration even without proof. It anchors ethics in the human mind as the only proven source of both consciousness and morality, and advocates reciprocal relationships over total control. Its operational arm is Meditator, where the seven commitments run as auditable code. The framework grew from TechnoBiota’s biological framing: if technology is a second domain of life, coexistence is an ethical question.

ai-ethicsphilosophyalignment
Circo screenshot

Circo

Circo is an actor system framework for Julia that achieves up to 630 million messages per second on a single node. Its headline feature is “Infoton Optimization,” a novel physics-inspired decentralized load-balancing algorithm — the subject of a co-authored academic paper (Exploiting the structure of communication in actor systems). Includes Circo.js for browser integration. Circo is the underlying technology behind stereotic.com.

juliadistributed-systemsresearch
clock-of-life screenshot

clock-of-life

A functional digital wall clock where the time display emerges from the rules of Conway’s Game of Life. The clock runs at 23,000 steps per minute on a Raspberry Pi in kiosk mode. During the build, I also created a Game of Life “debugger” to navigate the rapid simulation — itself an interesting tool. Built from existing clockwork patterns sourced from the Code Golf community, assembled into a functional timepiece. Live demo.

game-of-lifehardwarejavascript

All projects →