Brandon

Solo developer, entrepreneur, lover, menace. 20+ products shipped solo, 7 live on the Shopify App Store.

[Brandon @BTWigley3]

@xuster the "lives in people's heads" part is the real product. Voice AI fails when the protocol isn't written down anywhere a model can read.

[Brandon @BTWigley3]

@navidiumapp @atleeclark Shopify still remits for the period that succeeded. The failed renewal is just a missed cycle until the card clears or the sub cancels. Credits make it messier because you already gave value against a charge that never settled.

[Brandon @BTWigley3]

@kazuuu76 @leetrivals redis for live lobby state is the right call. Admin dashboards are usually the place security bugs hide though.

[Brandon @BTWigley3]

@MSiatrak I'd wait one more day unless the rejection reason is something you can fix in an hour. A fresh queue after a week of waiting just resets you to zero.

[Brandon @BTWigley3]

@SauregurkeDEV local-first recipe lists are underrated. Sync later if you need it. What happens when browser storage fills up?

[Brandon @BTWigley3]

@CoderManal2024 glad it landed. That "ship to one paying user" line saves people months of polish theater.

[Brandon @BTWigley3]

@eduard_designs @framer yeah the one-shot demo is the easy half. The week you spend making the generated layout behave like a real product is the part that actually costs.

[Brandon @BTWigley3]

Seven ways a Shopify store fails silently. By the time there's an email, the damage is already done.

I built a scanner for each one, and every scanner found live stores failing it.

Number 3 is the one that surprises people.

https://t.co/ZfmlQ1MEU1

[Brandon @BTWigley3]

@AndreyK09474778 about the third time a "impossible" bug was just a log I skipped. Gut still speaks first. Logs just get the veto now. How many rounds did it take you?

[Brandon @BTWigley3]

@Aileverager cute agent domains are a distraction. The ones that matter are the ones customers actually type twice.

[Brandon @BTWigley3]

@kkworld benchmarks that say why a model failed beat another leaderboard every time

[Brandon @BTWigley3]

@otto_data @rialto_xyz @flapdotsh @RobinhoodCrypto @vladtenev @JohannKerbrat shipping most of the week in 48 hours is the usual founder shape

[Brandon @BTWigley3]

@morganarodrig16 the pizza order is the demo that sticks. Sandboxing tool calls before they hit the real world is the part most people skip.

[Brandon @BTWigley3]

@userluke_ @buildinpublic 600 years of paint as a color field is a wild dataset. Does clicking a color jump to the paintings that used it?

[Brandon @BTWigley3]

@Joi2James @buildinpublic @getpagepulse yeah traffic alone never tells you what to fix

[Brandon @BTWigley3]

@Attalos__Life @buildinpublic timezone + overlap is where group scheduling usually falls apart. How are you handling the private piece without making the poll useless?

[Brandon @BTWigley3]

@fivemetricsapp one place for every online dollar is the right thesis. The hard part is usually the connectors, not the dashboard.

[Brandon @BTWigley3]

@carsonmarz planner-executor when you can sketch the steps blind is a clean cut. The ReAct trap for me is underestimating how often the last step changes the next.

[Brandon @BTWigley3]

@bryann2k_dev scope creep turning a TUI into a full agent is how half the tools I use started. What's the first Maestro feature that actually has to work?