kelly Quilodrán

Last week I spent two days polishing an edge case in a settings panel maybe 2% of users would ever hit. Meanwhile my onboarding flow still has a broken redirect I've been ignoring for a month.

I think it's a comfort thing, polishing a small, contained feature feels productive and safe. Fixing the core flow means touching stuff that's scary to break, so it keeps getting pushed down the list.

Lately I've been trying a dumb gut-check before every commit: "would a user even notice if I stopped right here?" If the honest answer is no, I stop and move to something that actually blocks someone.

Curious how you all draw that line, is it a rule, a gut feeling, or do you just ship when you're sick of looking at it?

I ve been thinking about a product-design problem that seems simple at first, but gets complicated very quickly:

When should an AI agent just do the thing and when should it stop and ask for confirmation?

The more actions an agent can take, the more useful it becomes.

But the more autonomy you give it, the more important trust becomes.

Both major labs switched on text watermarking within 48 hours at the turn of August. My first thought was that this is very hard to do in a way that actually holds text has no pixel layer to hide anything in. So I read Anthropic's own documentation on how it works.

It does work. It also breaks in a lot of ordinary situations, and they say so themselves: short text leaves too little signal, heavy editing or translation thins it out, screenshots and format conversion strip it entirely.