He did not write the video platform. He has not written any of it: not the company website, rebuilt from the ground up to replace an old WordPress install, forty-six pages across twenty-two tracked releases, with a landing page for each of the thirty-five cities in Maricopa and Pinal counties, including Mammoth, a mining town on the San Pedro River, founded in 1876, that is not, strictly speaking, a hot market; not the admin panel that lets him add a shoot, reorder a portfolio of roughly 3,470 images, and redeploy the site with one button, through a daemon that keeps the web tier from ever holding a shell; not the analytics he put in when he took Google’s trackers off the site; not the dozen-odd scheduled jobs that back up, monitor, and patch the server, including the one that will apply a kernel update but will not reboot while a video is encoding. He directs the work. The code is written by an AI, against a specification he writes first.
The specification says what the system must do, what it must never do, and how he will know it is correct, and it is written before anything is built. The work goes out in scoped increments rather than one large request. What comes back gets read, exercised, and checked for where it diverges, and nothing is accepted until it has been verified against a live system, with a dated backup and a documented way to roll back in place first. He is precise about the boundary. He is not a software engineer and does not claim to be. What he claims is the business: thirty-five years of knowing which requirements are right, which trade-offs are acceptable, and when a thing is subtly wrong even though it technically runs.
In January 2026 he founded BLUFify Media, a two-person Arizona company that applied the intelligence community’s analytic tradecraft to the ordinary news: ninety outlets graded separately on bias and on factuality, confidence language tied to evidence thresholds under the same directive the analysts use, and a second model checking the first for anything it had made up. At its peak the platform tracked 177 sources across fourteen sectors and produced 1,864 briefings, three a day, on 23,645 lines of Python that he directed and did not write. It never had a paying customer. It paused in May under bootstrap funding pressure, documented for relaunch, every file carrying a checksum. He states the part about customers plainly, which is the point of the whole exercise.