I run my whole working life as a team of one in the AI era, and I'm building the guidebook so you can do the same without losing the plot, or your privacy.

I don't use AI. I staff it. Working solo through the first half of 2026, I built an operation that does the work of a small team: a floor of 700-plus hours given back in a single quarter, realistically closer to a thousand, most of it running on a schedule without me in the chair. The number I'd actually stand behind isn't any single figure. It's that three separate records, kept in three different systems, all tell the same story.

hey, I'm Phil.

I spent my career in enterprise technology and cybersecurity, and somewhere along the way I stopped using AI like a tool and started running it like a team. One person, a lot of moving parts, and a daily rhythm that mostly runs on its own so my attention goes to the work only a human should be doing. (I keep my employer off this page on purpose. If it matters to you, it's one click away on my LinkedIn.) This site is where I show that work and teach the method.

how I actually work

I build, I don't just prompt. The output looks like a small studio run by one person: brand systems, interactive data experiences, research pipelines, the company map behind the weekly read. I direct it solo, and I keep the receipts.

I run a team, not a chatbot. A work life and a personal one, across more than one platform, with me as the layer holding it together. I'm now deliberately designing myself out of that bottleneck, because a system that only works when I'm awake isn't finished.

I work to a standard. Every build starts with three things: who it's for, what good looks like, and one thing not to do. That single habit kills most of the rewriting people waste their lives on.

And I keep one muscle un-outsourced. The thinking that makes the judgment calls stays mine. AI gives me back time. It does not get my attention, and it does not get my voice.

how I think about trust

This is the whole point of the place, so I'll be plain about it.

The perimeter isn't your apps and your logins. It's you, and everyone around you, and the whole space you and your people move through. Trust is structural. It gets earned by checking, not by claiming. So I hold myself to one rule: I won't tell you something I can't show you.

That's also why I keep a public scorecard of my own AI skill and re-run it on a schedule, gaps and all. If I'm going to teach this, you should get to watch me grade myself.

by the numbers

Held as floors and ranges, because the honest version of a number includes how it was counted.

• Hundreds of hours given back in a single quarter. A floor north of 700, realistically near a thousand. Call it most of a half-year of work, or close to two full-time teammates working beside me the entire time.
• Most of my daily and weekly operation runs on a schedule, without me in the chair.
• In one month, on a single tool, the raw output volume of dozens of books.
• I grade my own AI skill on a schedule. Latest composite: 87 out of 100, up from 81 in February, with the steepest gain in execution, the dimension I was weakest on. Self-scored on my own rubric, which stays public.
• Running these tools since the first month consumer AI existed, back in late 2022.
• The weekly read and the data engine behind it run on a roughly $200-a-year consumer plan, not enterprise tooling.
• And none of it rides on trust me. Three separate records, three different systems, the same story, with one figure landing identically in two of them.

what I'm building

• The Show: a weekly read on who's building what in AI, and what's really underneath the headline. New every week.
• A live company map, more than a thousand companies deep, that the Show plays off of.
• Interactive, visual experiences, shipped as working software you can actually open.

where this goes

It was never about personal productivity. The endgame is Human-OS: a human-centered way to run a life in the AI era that I can hand to my peers, to the people coming up behind me, and eventually to anyone who wants it. Figure out how to operate as a team of one without handing over the wheel, keep the human in charge, then pass the manual to everyone behind you.

I'm also lucky to know a lot of the people who quietly build the safeguards behind the apps you already trust. That's the company I try to keep.

don't take my word for it

You shouldn't have to, and that's sort of the entire idea. The repository is real and you can read it. The experiences are live and you can open them. The data is real data. The build logs are mine to pull. Everything on this site is meant to be checked, not believed.

▲ The movement is execution (+13.5), the dimension that was the long-standing gap. Capability and self-awareness were already near ceiling.