I’ve long considered myself easygoing. Wrong, nope, turns out I’ve been understaffed.
For years, I’ve put up with digital indignities and have given myself credit for being adaptable, for example:
My bank statements downloaded with names like
eStmt_2026_11_FINAL_v2(1).pdfAn app I use every day has three fields I never touch, or
A recurring task requires an extra click for no reason I can understand.
Like you, I’ve worked around these things and got on with my life (which does feel like a reasonably grown up way to behave).
But today I gave claude a two-sentence instruction about one of my annoyances. 25 minutes minutes later, gone. When I looked at the result I thought immediately the date would look better with hyphens, and, oh yea could we move the padding a little to the left?
Over these past two years, it’s become clear that I have far more opinions than previously appreciated. Some of what I took for an easygoing disposition was just knowing what it would cost to get my way.
Now: the advice we keep getting is to be more ambitious with AI. Felix Rieseberg of Anthropic has a whole audacious tutorial which encourages us to ask more of capabilities we routinely underestimate.
Sensible advice, it is and there is good reason to reconsider the limits of what we can do.
But I think we’re also overlooking what happens at the other end of ambition. Cheap execution also lowers bar of a personal desire worth acting on, like many of my past digital annoyances.
Or as I like to call it the pettiness dividend. In official terms, “the newly affordable ability to act on a preference you would previously have decided wasn’t worth the trouble.”
What AI lets us do, does also afford us to think smaller.
For example, recently on Hacker News there was a thread argument that generating more code-review comments, including nitpicks, makes sense because agents can quickly sort through them. In another, a founder described agents finding technically valid issues nobody would ever have prioritized: “it’s not wrong but also why should I care about that issue.”
The practical answer is that we’ve always had other things to do. But my favorite example is from this Lovable user who built a “Petty Rivalry Generator” to track the goal of overtaking an unnamed person’s LinkedIn follower count. And apparently the other person doesn’t even know the rivalry exists.
I find this much more encouraging than I should.
So much time is spent discussing what AI will let us accomplish that we can make ordinary satisfaction sound like an inadequate return on the investment (hint: its not).
Even the lofty talk about human taste gets funnier in this light. It’s fun to imagine that cheap intelligence will instantly reveal our gifts as creative directors. Some of those gifts will turn out to be very strong opinions about date formatting.
There is one small complication, though. See, once you can change something, you start noticing how you would change it. The first improvement introduces you to the possibility of a second. What was acceptable yesterday can become almost right today, and “almost” can keep our dear little agents occupied for quite a while.
And so remember: a preference can be cheap to satisfy and still take up more of your attention than it deserves.
I’ll have to get better at noticing when I’m improving something I use and when I’m finding something to improve because I can. I still intend to collect that dividend, though some of my opinions will need to remain understaffed.


