HyperC is an applied AI research and productization company based in Silicon Valley, founded in 2019. We are a team who dream of truly useful AI systems that will eliminate poverty, save the Earth and propel humanity to the stars — and we measure ourselves in customer-realized profit, not demos.
CriticalHop starts in classical AI planning — Fast Downward distributions, the ordered library (entropy-controlled contexts in Python), metaplanner research. The lesson: automation fails at the decision, not the plumbing.
Seven years of foundational work on PARML — Profit-as-Regression Machine Learning — culminating in P34: a decision model trained against realized economic outcomes, deployed live in Amazon wholesale, micro-lending and crypto routing.
The working paper formalizes computable markets and the MSD data contract; the membership opens the model to builders; enterprise deployments carry it into governed production. The category: the Self-Driving Business AI Model.
Businesses make enormous numbers of decisions about what to buy, produce, price, move, finance and hold — and much of that coordination still runs on spreadsheets, fixed rules, meetings and intuition. Our long-range thesis is that economic intelligence can become programmable: AI systems that evaluate more possibilities, enforce constraints consistently, and automate more of the operational work of moving a business toward its objectives.
The civilization-scale narrative is an abundance economy: less wasted inventory and capital, more efficient supply chains, broader access to sophisticated decision tools — and more human time spent choosing goals rather than manually operating the machinery of commerce.
| Legal entity | CRITICALHOP INC., d.b.a. HyperC — Delaware C-Corporation, founded 2019 |
| Headquarters | Silicon Valley, California |
| Products | P34 · PARML · HyperC |
| Founders | Andrew Gree (Director, CEO, lead developer) · Tyler Hoffman (co-founder) |
| Backing | DVC (Davidov's Collective), which lists HyperC publicly in its portfolio |
| IP position | Filed provisional patent; multiple published defensive publications |
| Category site | computablemarkets.com |
We publish working papers, a proposed data contract for the category, and a defined falsification protocol — while shipping a commercial product measured directly in customer-realized profit. Research keeps the product honest; the product keeps the research grounded.
We hire researchers and engineers who want their models graded by a P&L. Write to careers@hyperc.com.