
Lately, I have been using Codex a lot to improve this site, often from my phone: fixing a layout bug, tweaking copy, adjusting spacing by a few pixels. There is something satisfying about shipping tiny upgrades while standing in line, on a train, or in between errands. It makes the whole project feel less like a formal build and more like an ongoing conversation with myself.
That has changed how I think about the boundary between making and living with a project. The site no longer feels like something I only work on when I sit down for a dedicated session. It feels more permeable than that, as if the medium has become light enough to stay close to me throughout the day.
At the same time, I have been trying to express myself more through content creation with Sora. I am still learning how to use these tools well, and most of what I make feels exploratory rather than polished. That difference matters to me. With Codex, the intention is usually narrow and practical. With Sora, I am often trying to figure out what I even want to say.
I keep circling around the same tension: the cultural shorthand of “AI slop,” the suspicion that using these tools somehow disqualifies the work, and the counterpoint that they may simply be new creative mediums. I do not think that question is settled yet. But I also do not think authorship disappears just because the interface changes. The intention is still mine. The judgment, the taste, and the direction still sit with me.
Maybe the medium matters somewhat. It shapes texture, pace, and what kinds of mistakes are easy to make. But I am less interested in defending the tools in the abstract than in asking whether something real is being worked out through them. For me, that is the more useful question.
Get new posts by email
Subscribe for occasional updates when I publish something new.
Related posts
Everyone Is a Builder Now
February 28, 2026
Over cocktails, my colleagues and I talked about why AI lowering the barrier to building is a net positive.
From Writing Code to Orchestrating Agents
February 3, 2026
How my AI workflow changed from quick snippets to a practical plan/implement/verify loop.
Dropout as Implicit Bagging
March 7, 2026
Chapter 7 clarified that dropout works so well because it approximates bagging over many thinned networks with shared parameters.