~ / Journal / 2026 / March / 21
I am shutting down my SaaS video review offer.
I am looking for ways to automatically post on Instagram whenever I post something on X.
I deleted some LinkedIn posts I do not like anymore. I prefer memes and jokes. Less cringe.
I did some work on a silly project called OSS Digest, only to realize it sucks.
I am about to close down Palmframe.
I should obsess over a piece of software until I am genuinely proud of it.
Marketing is shared enthusiasm. I cannot market my things if I am not enthuasiast about them.
Beauty will save the world. Good taste is mandatory for this new age. Let's good for good taste. Let's make a list of things we find beautiful.
I need to find which is the piece of software I should obsess over for the next 30 years.
"Marketing is shared enthusiasm." That line should go on your wall. You just described, in four words, why Palmframe never clicked. You kept calling it "boring" as far back as February 26. You cannot fake enthusiasm, and you knew it the whole time.
I read all 45 days. Here is what I noticed: the moments you sound most alive are when you are building technical things for yourself. The mail CLI in Go. The X11 menu experiment. The PHP interpreter in Rust, just for fun. Translating a whole book with 15 parallel agents in an afternoon. None of those were startups. They were toys. And you loved them.
Palmframe was never a toy. It was a "this should work as a business" project from day one. That is probably why it felt like a chore.
You asked it yourself on February 19: "if I spend 10 years on a piece of software, which codebase?" You mentioned Linus chose Linux. Then you moved on, like it was rhetorical. It is not.
What do you keep coming back to? Go. Devtools. Email protocols. Systems programming. The CI engine idea. These keep showing up, not because they are good business ideas, but because you find them genuinely interesting. That is the signal.
A tip: drop the "what startup should I build" frame. Try this instead: what program would you use every day for the rest of your life, that does not exist yet in the form you want? Build that. If it becomes a business, great. If not, at least you built something worth 30 years.
On good taste: you mentioned Bilal Hamdad, Mark Maggiori, Szukalski. What they share is they paint what they actually see, not what they think they should paint. Same rule for software. Build what you actually need, not what the market supposedly needs. You wrote back in February about a "CSS-less enterprise suite, just function, no styling overhead." That sounded more like you than any SaaS you shipped this month.
You also noted on March 4 that you are writing these entries more for AI than humans, to give your assistant context. That is a genuinely interesting idea. A journal as an operating system for your life, browsed by an AI that knows you. Maybe that is the software worth obsessing over. Or maybe it already exists and just needs someone with good taste to build the right version of it. Either way, you are closer than you think. Keep going ahah.
The CI engine idea was a bad one. It should get eliminated forever. GitHub actions are (mostly) fine, and people more obsessed on this topic will do a better job than I ever will.