Staying oriented in strange days: three essays on AI, craft and trust

We’re living through a regime change that throws you off. Machines write code at a speed no person can match, and around that speed a fear-laden conversation has grown: that AI atrophies you, that the craft is ending, that the human is redundant. These are strange days — and at StrangeDaysTech we take them seriously, because our job is to build tools for getting through them without losing your bearings.

Over the last few weeks, our founder, José Villaseñor, has been working out a response to that fear on his personal blog. Not a passing comment, but an arc of three essays that read as one. Here we gather them, tell you what each is about, and explain why this argument is, at bottom, the reason StrayMark exists.

The argument, in three acts

The bet — I don’t want an AI that needs me less. The industry sells AI on how little it needs you: agents that work for days without asking, without waiting on anyone. The essay flips the frame. You have to distinguish coding —typing, wiring, translating intent into syntax— from engineering —deciding what to build and why, what to reject, where the work is headed. Automate the first; the second stays with people. The human in the loop was never the inefficiency to remove: it was the foundation of trust.

The myth — The immaculate programmer. A common reply says the best programmers would never make the mistakes AI makes. But we built an entire discipline —code review, linters, postmortems— precisely because they do. The honest comparison isn’t AI against an idealized human who never existed; it’s a process run with AI against one run with people. The debate was never about raw capability, but about whether you keep the engineering around the code or expect the code to hold itself up.

The fear — It’s not atrophy, it’s a mismatch. The word “atrophy” medicalizes what is really a mismatch of speeds: the machine runs, and we don’t have the map to keep pace. The cure isn’t to slow the machine down or resign yourself to being left behind, but to pair up: to put knowledge —not just information— in front of the person, a map they can stand inside, see what was decided and why, and decide without holding the whole system in their head.

Why this is StrayMark

The three essays land in the same place. If the problem isn’t AI’s capability but keeping a human oriented and in command, then you need a tool that leaves a record and a map where that person can actually stand. That’s StrayMark: it doesn’t exist so you use more AI, but so the AI you already use leaves a trail — what was decided, what’s in motion, what’s still pending, where it’s headed. It’s the same bet we made when Google published the OKF, in Who is your AI’s knowledge for?.

The question is no longer how capable AI is. It’s whether the people responsible for the outcome can still see what it does — and decide from there.

A response to strange days

Strange days are the ones where the machine runs and you fall behind, overwhelmed, with the ground shifting under your feet. The response isn’t to compete with the machine or to delegate blindly: it’s to stay oriented. That’s the work, and it’s the one we build.

If you want the full argument, the three essays are on José’s blog: the bet, the myth and the mismatch.