private company overview
The century ahead hands people more control over their own minds and bodies than any before it. Whoever owns that layer decides whether it frees them or bills them for it. This company builds the free version: independence from surveillance, privacy by architecture, personal agency held as a product decision rather than a policy page.
cruush opens the portfolio with how people find each other, and its codebase is built to stand up further networks, so the second one costs a fraction of the first. Soma, Alan AI, and HEKATONOMOUS extend the same thesis into memory, judgment, and work.
Mutual Visibility Cryptography and Privacy-Preserving Proximity Discovery turn the hardest privacy problems into durable capability. Where they hold up, they can be licensed as well as shipped.
Selective software and design engagements buy independence: cash without outside terms, operating proof in public, and capability that returns to the portfolio.
Our AI research works toward intelligence that remembers rather than merely scales: durable memory, fewer weights, a lighter footprint on the world. Small models that stay with a person are what make personal intelligence ownable at all. The company intends to publish that work as it matures.
Map-first queer cruising app with reciprocal discovery and location privacy, on a codebase designed to seed further networks.
flagship / in developmentPersonal AI personality designed around persistence, individuality, and user control.
future productGeneral-purpose AI direction centered on LGBTQ+ people and uncensored creative freedom.
future model / researchA GitHub issue fixer that carries a bug from report to tested, independently reviewed fix, with every stage visible to its owner.
coming soonEncrypted visibility rules evaluated reciprocally so users become discoverable only when both satisfy the other’s visibility policy.
Nearby-user discovery designed to minimize exact-location disclosure, with Ghost Mode providing rotating viewer-specific display coordinates.
Privacy is a commitment, not a feature. The company does not sell user data, and it will never be used to train an AI.
The horizon here is Y3K, not the next cycle. Anything that only works while a particular platform survives is not worth owning.
Products, platforms, and private tools engineered around the client and handed over whole, not rented back to them.
Web experiences where art direction, motion, sound, behavior, and engineering become one coherent environment.
Search and answer engines, handled as engineering: structured data, fast pages, and content shaped so people and AI assistants find the work first.
Agents and creative AI that run under the owner’s control, on their terms, with the work and the data staying theirs.
use it. commission it. back it.