3 of 113 works · Research.
Our standing argument that the interesting question about AI is not which jobs get faster, but which jobs become possible that nobody was ever staffed to do. It underwrites how all four units pick their work.
Every family in the register, rewritten so an operator can tell in two minutes whether it applies to their problem. A standing publication program, not a marketing page.
Training is treated as a broadcast: publish the course, record the completion, hope for the change. We treat human knowledge as a control loop instead, where every step from delivery to outcome produces evidence and the evidence decides what the person gets next.
A family of USPTO filings around one invention, granted or pending, with its real status.
An operating company chartered by Studios, with an entity, an operator, and a closure path.
A capability built for the whole portfolio and designed to be spun out on its own.
A company funded, built, or spun out. (Lapsed filings remain in the patent register.)
A pool of capital with a written thesis and a stated deployment model.
Long-form publishing that makes the case for an invention or an operating model.
Standing research programs and publication series, including peer-reviewed work.
A landscape read that changed a filing or funding decision.
A defensive publication that puts prior art in the open so a bad patent cannot be filed later.
A structured intelligence asset that decides where capital, filings, or ventures go next.
A reusable operating method that the rest of the portfolio is built on.
A piece the founder published on his own site, tagged by type, that argues a finding in the open before it becomes a filing or a fund.
An owned online publication that makes the case for our work to a standing audience.