Alexandra Protsenko

Building and growing a portfolio of products: Paste, Wishes, ChattoCourt, PostSam. Reflections on running lean, and honest notes on what didn't work.

[Alexandra Protsenko @alexfoundercph]

AI didn't remove the thinking from research. It moved it to the front.

The effort I used to spend searching now goes into defining the question. I think that's a better use of it.

[Alexandra Protsenko @alexfoundercph]

Keep the handoff visible. If you can see what someone actually asked the agent for, you can tell whether the brief failed or the execution did.

Without that, every "the agent got it wrong" turns into opinion against opinion.

[Alexandra Protsenko @alexfoundercph]

Feeling the echo doesn't mean it's your storm.

When someone near you is going through something hard, the heaviness arrives and starts steering. The line doesn't push it away. It names it. Naming is the whole technique.

[Alexandra Protsenko @alexfoundercph]

@thisiskp_ The kids' agenda. I'd hand a bot every one of those awful parent apps and just get one message with what's on this week.

[Alexandra Protsenko @alexfoundercph]

@jasonfried This is the gap I notice most. With kids, most evening events don't happen for me, so a weekly call sounds like the realistic way to stay close to people a stage ahead.

[Alexandra Protsenko @alexfoundercph]

@tibo_maker Not a full solution, but the rule I try to keep: parallel sessions are fine as long as they're all on the same project. A bug, some pages and a refactor in three unrelated places at once is the part that fries my brain.

[Alexandra Protsenko @alexfoundercph]

"Envy is the one sin that isn't even fun." Charlie Munger.

It came back after a run of podcasts about other apps and their revenue. Podcasts feature winners, and winners share the highlight reel. Measuring my Tuesday against someone's best quarter isn't information. It's a mood.

[Alexandra Protsenko @alexfoundercph]

A question that fires in my head a dozen times a day now: how could this be automated?

Not in planning sessions. Before almost any task, as a reflex. The biggest surprises aren't the work tasks. They're the ordinary ones.

[Alexandra Protsenko @alexfoundercph]

My co-founder's attention is scarcer than my time.

Not more valuable. Scarcer. His is split across projects and hard to schedule. Mine is predictable, which means I can move it. That difference changed how I handle days that don't go to plan.

[Alexandra Protsenko @alexfoundercph]

My AI coach has a standing instruction: disagree with me.

The default is agreement, and agreement teaches you nothing. So: challenge weak assumptions, question optimistic conclusions, don't fold to be agreeable.

[Alexandra Protsenko @alexfoundercph]

One bad Tuesday is not evidence.

At work I'd never call a landing page wrong after one bad day. At home I do it all the time. One hard conversation: "we don't communicate". One slow month: "the market's gone".

Count the data points first. If it's one, try again next week.

[Wishes @wishes_app]

For the husband who swears he's not dressing up this Halloween. The Dude from The Big Lebowski. Bathrobe, sunglasses, sandals. So, his school drop-off outfit.
https://t.co/s9Wo4WQLNy

[Alexandra Protsenko @alexfoundercph]

A rule from my own notes: learned pessimism is as dangerous as blind optimism. When results come slowly it's easy to decide nothing works, and then you quietly drop the things that were working.

[DHH @dhh]

You don't have to commit your entire life to optimism, just try it on for a week. Start thinking "what if it all worked out?" and "what a time to be alive!". Then compare your state of mind to the week before filled with pessimism and doomer nonsense.

[Alexandra Protsenko @alexfoundercph]

@skirano Agree, and it lands hardest on product-led teams, where the product is also the marketing. If what people try for free can be recreated in a week, where do you put the steps ahead?

[Alexandra Protsenko @alexfoundercph]

@businessbarista Fitting pick, since Spider-Man's famous line is about responsibility as much as optimism, and from those quotes @dhh has both covered.

[Alexandra Protsenko @alexfoundercph]

@tdinh_me Curious how long hands-on mode stays hands-on. My weak spot is when the output starts looking clean and I catch myself approving instead of checking.

[Alexandra Protsenko @alexfoundercph]

People blame AI agents for their own mistakes.

I catch myself reaching for it too. Something comes back wrong and the first instinct is "the agent got it wrong", not "I briefed it badly" or "I didn't check".

The tools changed this year. Who makes the mistakes didn't.

[Alexandra Protsenko @alexfoundercph]

Small tasks steal focus from big ones, quietly, every day.

What finally worked for me: protected time blocks, and a "parking lot" for everything else. I get back to it in a dedicated slot later. Obvious. Also the hardest habit I've built this year.

[Alexandra Protsenko @alexfoundercph]

@amasad The prompt I'd want right after this one: stop when it isn't paying back.

[Alexandra Protsenko @alexfoundercph]

@levie At our size the first eval is a plain note on what each agent is for, including what it can't do. Most of our "the agent is bad" moments were someone asking it for a thing it was never set up to do.

[Alexandra Protsenko @alexfoundercph]

My rule this year: never keep going on something just because we've already invested a lot.

[Paul Graham @paulg]

One very important thing that even the smartest people seem to have to learn from experience is taste in problems. There are juicy problems that tend to lead to interesting discoveries, and nasty problems that tend not to.