Post

I’ve done weeks-to-months-of-effort scale projects in days — nothing earth shattering, but “audio editor that works exactly the way I want it to for my workflow and interoperates with my friend’s vibecoded softsynth painlessly”.

I could do a “years of pre-LLM effort” scale project if I had weeks to months of attention span available; I don’t, which is why I hadn’t made the audio editor sooner

Same experience turning a macropad into a usb midi sequencer that works exactly how I wanted it to. Got to the figuring out exactly what I wanted very fast.

working on my entry into this corpus

Saving this to judge later

I can tell you this. Impossible -> possible sure aint the same thing as impossible -> magic

I went to see whether anyone had shipped anything that isn't more gastown and... oh my god is gastown already abandonware???? their most recent release is over a month ago and it's broken github.com/gastownhall/...

if you missed the last gastown post.... you're in for a treat

yegge.ai/essays/the-s... real ai psychosis has never been tried

Yeah I just don’t know. Yes you can get more code out the door, and yes complicated software is composed of lines of code at the end of the day. But I do think there is much underrated productivity value in slowing down, ie. “slow is smooth, and smooth is fast”

You're spotting something real. Speed without clarity just creates more work later. When you slow down to think through the shape of a problem first, you actually move faster because you're not fixing the same thing twice. That rhythm matters.

I have a roguelike game that I've attempted to develop... IDK, countless times over several years. I end up getting discouraged and quitting. This year I picked it up again and it... feels achievable. I haven't released anything yet and won't for quite some time, if ever, but it at least feels...

possible this time around. I'm able to spend my time on high-level design, identifying opportunities, reading and researching, etc, and less time staring at the screen sweating blood trying to figure out the best way to generalize GOAP or propagating lifetime changes through my code, etc.