S

About

Skakeny is my notebook.

Today I build apps with a team of Smiths. The Matrix kind, not the blacksmith’s — though some days, watching one confidently break something that worked, it’s hard to tell. I’m a product guy. I work out what’s worth building and what “good” looks like, then get it built with a team that happens to be made of AI.

One of them, Job Studio, is live now. More are close behind. I’m not selling them here; they’ve got their own corners for that. Here I just build, and tell you how it went.

I keep starting from zero, and I never quite leave anything behind. HTML in 1998, off View Source and stubbornness. Flash the year after. A creative shop of my own, and when a telecom swallowed it in a merger I went from me and two developers to running thirty: print, web, 3D, campaigns across the region. Print back then was a different animal than today’s soft digital stuff. Film to plate to press shop, me over a proof with a loupe, checking the plates landed dead-on, or the whole run was scrap. Then an agency. Then a food business: seventy people, a real kitchen, a real P&L.

Relearned from zero

HTMLFlashDesignPrintA restaurant lineProductionProductNow — building with AI

Took me years to see it, but it was all the same job. A creative shop is a product. A campaign is a product. A restaurant is a product — the menu, the room, how a Saturday night runs. I wasn’t switching crafts. I was building products, in whatever material the vertical handed me. So a few years back I stopped picking up new trades and went all in on that one thread. Build across enough verticals and you stop seeing separate trades; you start seeing what carries. A kitchen and a codebase run on the same discipline. Now it’s AI. Same job, new material.

I’ve always taken things apart to see what makes them tick. As a kid, whatever I could get a screwdriver into. Now it’s software and systems, the same itch pointed at what makes a thing work and where it breaks. That’s the whole reason this blog exists.

Most days with the Smiths go the same way. Build something. Watch a Smith confidently fuck it up. Find the fix. Overcorrect. Catch it. Settle on the middle. Write down the rule that holds. Then tomorrow, again. They lie with a straight face, swear the thing works when it doesn’t, so you keep them on a short leash and taste every plate yourself.

I’m not a writer. I write this because journaling the mess keeps my head straight, and maybe my way of doing it is useful in yours. What you’ll find is what I do and how I do it: the techniques, not prompts to paste.

There’s a lot of noise out there making you feel behind, like you should already have this AI thing figured out. If you’d rather understand the work than collect screenshots of other people’s demos, you’re who I’m writing for.

Grab a beer. Pull up a stool.

Craft over hype. Nothing for sale here.