A Weekend Build, and the Slow Years That Made It Possible
I built working software in a weekend. The speed is the least interesting part — it was borrowed entirely from the slow years before it.
Read →FOLLOW A THREAD
Decisions revisited, assumptions challenged, and judgment earned by doing the work.
Oldest first. Follow how the thinking developed.I built working software in a weekend. The speed is the least interesting part — it was borrowed entirely from the slow years before it.
Read →Working software came together in a weekend. The judgment behind it came from years of deals, spreadsheets, and unfinished systems.
Read →Recording a decision in the moment preserves the reasoning that gets lost when the story is reconstructed later.
Read →A genie makes you careless. A new hire makes you a manager. Why everything I do with AI is propose-then-approve until trust is earned.
Read →A solo operator loses the one thing that keeps judgment sharp: someone across the desk to say you're wrong. Here's how I rebuild that adversarial loop with AI — first a sparring partner you argue with, then a weekly coach that remembers what you committed to and holds you to it.
Read →Working alone means losing the person across the desk who tells you you’re wrong. A sparring partner and a weekly accountability routine help rebuild that loop.
Read →A flat wholetail percentage was off by about 20 points either way, so each estimate now carries its own wholetail number. Finished jobs get a short debrief on what ran over.
Read →My underwriting engine used to pick the exit and save it like I'd chosen it. Now it measures every exit and I make the call, on the record.
Read →Condition now comes from the sale record and the photos instead of a data field that called every house heavy. Every saved analysis carries the flood zone, and a failed lookup says so.
Read →