When I notice a project dragging, my reflex is always to cut scope. Trim the feature down, ship the smaller version, get something out the door instead of polishing the grand plan forever.
And usually that's the right call. Most of the time I was building more than the problem actually needed, and cutting it loose feels great.
I build faster now with AI than I ever could by hand. Things that used to take me a weekend now just take a prompt, and a lot of that is stuff I genuinely couldn't have written cleanly myself.
So part of me looks at non-technical founders asking whether they should learn to code and thinks the honest answer might be no, or at least not the way I did it. Spending six months grinding through tutorials to write code an AI will write better feels like a waste of time.
When we first launched, we operated lean. We improvised our tools, worked out of tiny spaces, and every dollar spent was strictly tracked. Then right after our first funding round, we upgraded to expensive software suites, hired too quickly, and spent heavily on premium perks. Instead of accelerating our growth, it dramatically raised our burn rate and left us scrambling before reaching product-market fit. How do you maintain a frugal, resourceful mindset across the company once fresh capital hits the bank account?
I've been spending some time using GPT-6 Astra, and i noticed something I find more interesting than usual "is the new model smarter?" discussion. I'm finding myself giving it more context before asking it to do something, rather than trying to break every task into small prompts. That got me wondering whether newer AI models are actually changing the way we work with AI, rather than simply improving the answers we get. For people who have tried Astra: Has a newer model actually changed the way you structure your prompts, workflows, or projects? I'm particularly interested in real examples rather than benchmark results.
I build a trading bot for Polymarket, so Brazil's runoff on Sunday was an interesting one to watch. The week before the vote, the market priced Fl vio Bolsonaro at 77% while Datafolha and MDA had the race basically tied. On election night, after 8:30 p.m. ET, Kalshi still had Lula at 21%. He won 51.8 to 48.2.
The next morning B3 hit its circuit breaker at the open for the first time since 2020. The Ibovespa closed down 13.9%, the dollar jumped 8% against the real, and EWZ lost 22% in New York. Stocks had been priced for a Bolsonaro win too, and the market moved all at once when that turned out wrong.
Turns out password reset emails taking almost a minute to arrive was causing a huge chunk of our user frustration.
People would get impatient, hit reset three times and give up or flood our support inbox. Replacing email links with a quick on-screen verification code for active sessions fixed most of those tickets instantly.
Are there any parts of your app where delayed transactional emails are quietly killing the user experience?
I used to think staying invisible was the safer move. Let the product do the talking, no personal baggage attached to it. That was my plan for the first year of building.
Then I watched two founders in my own network launch almost identical tools within a month of each other. Same market, similar pricing. The one who'd been posting his half-baked thoughts and failures on Twitter for a year got 10x the traction on launch day. The other one, better product honestly, got maybe 40 upvotes and went quiet.
