Using AI to spin up new features is awesome for speed, but letting it handle the test suite usually creates a mess of brittle tests. They either pass by mocking out half the application or fail because the agent didn't understand the broader system architecture. Lately I've started treating AI-generated code as untrusted until I manually write or review the tests myself.
What kind of review workflow or safety checks do you use to make sure your test suite doesn't slowly rot away?
I spent a day trying to fix something and could not. Nothing worked. I tried everything I could think of.
Then I remembered I had deleted a block of code months ago because I thought it was useless. It was not useless. It was doing something I did not understand, and I had removed it because it was old and messy.
I wasted months trying to force myself to learn complex financial modeling and detailed project management tools just because I felt I "should" be good at them. I barely got any better but it drained all my creative energy for product building.
Eventually I realized it's way smarter to just hit a basic "good enough" level on your weak spots or bring in a freelancer/cofounder to handle them so you can spend 90% of your time on the stuff you're naturally awesome at.
How do you handle the parts of running a project that you're just plain bad at?
During our days we would send surveys asking our users what new tools they wanted us to add to the platform. We ended up building a reporting dashboard because a dozen vocal customers requested it.
After spending three weeks developing the dashboard only two of those customers actually used it more than once. We learned the way that customers are excellent at identifying their frustrations but terrible, at designing software solutions.
We changed our feedback process to focus entirely on their workflow instead of feature wishlists. Now we only ask them to describe the tedious part of their week or the task that takes them the longest to finish.
This subtle shift gave us a deeper understanding of their underlying problems. We recently built an automated export function based on workflow feedback and it immediately became our most popular feature of the year.
Our email list hit 40k subscribers, but open rates dropped below 12%. I was terrified to clean the list thinking a big subscriber count meant more potential buyers.
We finally deleted 12k people who hadn't opened a single email in six months. right after open rates jumped to 28% and weekly email revenue went up by 15% because main inbox placement improved.
How often do you prune inactive contacts from your newsletter list?
Something I've run into repeatedly and I can't tell whether it's the platform, my settings, or just me.
I found a comment on my own intro thread that had been sitting there 20 days. It was the most useful piece of feedback anyone has given me on this platform, someone had taken the time to pull apart my positioning properly, and I only saw it because I happened to be scrolling my own activity. No notification appeared. Same thing again this week with two people who tagged me on launch day. So I was three weeks late to both.
In the beginning, you celebrate every single sign-up like crazy. But after a while getting a notification that someone canceled their subscription feels like a direct gut punch.
We looked through our exit feedback recently and almost nobody left because of bugs or broken features they just stopped opening the app. It's wild how hard it is to actually bake your tool into someone's daily routine. If you've been running a product for a while, does the anxiety around churn ever actually calm down or do you just get used to it?
We learned this the hard way after spending weeks building a complex integration that our most active, non-paying community members kept asking for.
Once it shipped, none of them actually used it while our paying subscribers were quietly dealing with basic bugs we hadn't fixed yet. It's so easy to get distracted by loud voices on forums who want everything for free but at the end of the day paying customers have totally different priorities.
How do you decide whose feedback actually makes it onto your product roadmap?
