Articles
I've built 30+ products over 20 years — some shipped and still running, some sold, some abandoned, some that never made it past a prototype. These articles cover what that actually looks like: where AI-powered products break in practice, how to scope an MVP when you've watched your own fail to validate, and what it costs to wear all the technical hats at once. Practical and specific — not startup advice for its own sake.
Posting the same message for two weeks got me nowhere new. Building it into a tiny app gave me weeks of content as a side effect — not because the app writes posts for you, but because shipping the small thing keeps paying off. Why building small beats writing one more post.
A design partner is not a beta user or an early customer — they are someone with deep domain expertise who helps shape what you build before you build it. Done right, it is the fastest way to capture knowledge you cannot get from surveys or user interviews.
Everyone says to put your answer in the first two paragraphs so AI engines will cite you. No primary study has tested this. Here is what the actual data — 560,000 AI Overviews and several academic papers — says about what gets an article cited, and what does not.
Founders who build a front end with a no-code tool and call it an MVP are not lying — they genuinely believe it. That belief causes real damage: wrong hiring conversations, wrong funding conversations, wrong timelines. Here is what the difference actually costs.
A free tool that turns a form into ready-to-post PNGs — Reels and Shorts at 1080×1920, plus square, portrait, landscape, Open Graph, and Pinterest sizes. 27 dark presets, 8 fonts, per-platform safe zone preview. Here is what it does and why I built it.
For nearly five months every article on this blog emitted [object Object] in its OpenGraph date tags — invisible on the page, visible only in the source. Here is how Claude Code surfaced it in a routine SEO review, what it actually cost, and the signal I had misread completely.