Projects, design, translation, sales — anything I feel is worth writing down.
Should a team be built out of people who resemble each other, or out of people who do not. My lean is toward the second, and this piece is mostly an attempt to work out how much weight that lean can carry. The literature turns out to agree with half of what I believed and to push back fairly hard on the other half, which is the more useful outcome of the two.
I have worked with three people in human resources, and between them they landed almost exactly on the three points I would have drawn if somebody had asked me to sketch the range — one above what I expect, one well below it, one flat in the middle. What the exercise keeps telling me is not about the three of them. It is about the chair: the one seat that writes the rules, enforces them, and decides who else gets a seat.
For a long time I held a quiet, unexamined view of the administrative department — a place where not much value is made, and not much is lost. Two small things at my current company have taken that view apart. The department with the widest reach inside a company is also the one with the weakest feedback loop, and a bad one does not fail neutrally. It fails onto everyone else.
The two-image-week piece was the mechanics of the rule. This is the why. The rule makes three mistakes at once — at the level of the individual (overjustification), at the level of the process (redundancy), and at the level of the relationship between the manager and the team (trust). None of them is visible from inside the rule. All of them are why the rule keeps failing in the way it fails.
A small creative team is not a factory floor. The two-image rule, the quota, the count — these are workarounds for a manager who cannot see the work, and the workarounds fail, in a small creative team, in four predictable ways. This is a note on the four failure modes, the four principles I have come to hold instead, and the team I am trying to learn how to manage.
The agent era drops the cost of building for every team. It does not drop the cost of paving. The argument here is for the role the agent era is going to make most under-priced: the person who abstracts across business domains and lays the paths everyone else walks on — the paved paths, the unified auth, the security guardrails, the infrastructure modules. Without that role, the unbundling produces a fleet of Frankenstein products.
A note on what the agent era changes, what it does not, and the kind of literacy the next decade will actually require. The argument is not against agents. The argument is for one old principle: professional matters should be handled by professionals.
A note on two failure modes I have been watching this year — the version of KPI where the form has no teeth, and the version where it has too many. Neither is working the way the people who designed them hoped.
There is a category of work that looks like work, fills the calendar, and produces no creative output. It comes back every day, in slightly different forms, and never quite goes away. It is called dirty work, and it is the silent killer of every team that is supposed to be doing creative work. The system — or the agent — is the only thing that can move it out of the way.
When I need to think, I leave the office. Park, grass, an empty field — the wider the better. The principle has hardened into a team policy: 16:00 to 16:30, every day, no exceptions. The half hour is not a coffee break. It is a deliberate swap into the mode the work actually happens in. This is what the long literature on attention, rest, and creative incubation has been saying for fifty years. I am just trying to take it seriously.
Instagram gives every account exactly one link in the bio. For a personal account, that is fine. For a brand running two entities, six channels, four subdomains, and an international competition, one link is not a link. It is a bottleneck. Here is how I think about replacing it.
A working note on the question I am chewing on: in the AI Agent era, can a sales lead or an operations lead file, refine and ship a real requirement, end to end, without a developer ever being the bottleneck?
A quiet place to put the work and the thinking behind it. Project postmortems, design notes, translation reflections, observations from operations and sales. I add things here as I figure them out.