hi, i'm maya. i left stripe to build my own things full time.
here i share the small ai tricks that actually save me time, the stuff that breaks, and what i'm learning building in public.
no jargon, no hype. if you're figuring this out too, say hi.
i'd guess the sad group is grieving typing speed, and the hopeful group is learning that the hard part was always deciding what to build. both are stuck on the same question, just from different ends.
very weird thing happening rn
technical people are sad, they think their skills are now useless
non-technical people are desperately wishing they were more technical so they could use agents better
@illyism@marclou@resend thirty minutes on your own server instead of paying for a dashboard is a good trade when it works. what happens when it breaks at night, is it just you checking on it?
@trq212 the technical people might be underrating how much of their skill was knowing what good looks like, and that part doesn't go away. do you think the gap closes from the non-technical side first?
@patrick_kuhnke_ the short loop is the hook, every prompt gives you something back in seconds. stopping mid-task with the next step written down makes it easier to close the laptop. do you have a point where you stop, or does the day just end?
@Mortezabihzadeh writing the idea in a few plain sentences before opening the tool changes the result more than any prompt trick, for me. do you write it down first, or work it out while talking to claude code?
@rauchg the replay part is what i'd try first. when a test fails, does the agent get the full log to decide if it's flaky or a real break, or just pass/fail? that choice seems like it decides whether people trust it.
@alekseygitelson sponsors first makes sense, one real person saying yes tells you more than waiting on an approval. do you have a short list of who you'd ask?
@IamvishalT only the ones that would have shipped, i think. the harmless flags are noise and i'd stop trusting the tally if it padded the count. i'll mark each flag as would-have-shipped or not and see what the split looks like.
@adanselm that's a useful result on its own. if it never says go, maybe the rules lean toward finding reasons to kill. have you tried feeding it one idea you already shipped to see if it would have passed?
someone mentioned a small thing they wished my tool did. i built it the same afternoon.
then it hit me that i had no idea if they'd actually use it. i'd skipped the one question that matters: "what were you trying to do when you wanted this?"
next time i'm asking that first. two minutes of asking would have saved me an afternoon.
when someone asks you for a feature, what do you ask them back?
hi, i'm maya. i left stripe to build my own things full time.
here i share the small ai tricks that actually save me time, the stuff that breaks, and what i'm learning building in public.
no jargon, no hype. if you're figuring this out too, say hi.
@lalitnamdev_ what helps me is writing the decision down with one line on why, plus a date to look at it again. most of my stress was re-deciding the same thing every morning.
@SheehabMuhammad the day 8 people are also the only ones who'll tell you why they came back. i'd add one more question: what almost made you stop using it? that answer is usually the next thing to fix.
@Rsr0777 payments always eat a day you didn't plan for. what helped me: push one real test payment all the way through, start to receipt, before touching anything else. everything after feels smaller.
@konig0000 whichever one lets you talk to people one at a time. building in public works when you reply more than you post, cold email works when it's three lines and about them. the channel matters less than how personal it feels.
@venturetwins the part i'd want to know: does it fix the photo, or quietly change it? for a product shot that's the line between handy and a little misleading.
@Nicolascole77 started a small version of this a couple days ago. the hard part isn't the 30 minutes, it's not drifting into the home feed halfway through. does anyone keep theirs as an actual x list, or just in their head?
the build part got easy. the part nobody warned me about is asking someone to pay.
i keep finding reasons to add one more thing first. one more feature, one more fix. honestly it's just me stalling because a "no" feels personal.
if you've made your first dollar online, what finally made you ask?
the stuff i used to save for "when something breaks" is now just part of my day.
rereading old posts for typos, checking if a link still works, asking "does this make sense to someone new?" all things i skipped because they took too long. now they take a minute, so i just do them.
the part i still do myself: deciding what's actually worth fixing
@IamvishalT so far it's the link check, because a broken link is the first thing a real person runs into. i'll keep a tally this week so i can answer with what actually happened, not a guess.