Back
Happenlist
Happenlist
#happenlist
What to do in your city, reinvented. This is the first time this is going to work & it will restore offline life.
Load previous page…
Got some great pull from friends on 
#happenlist. Obvious what the current bottleneck is, and I have a confession: I'm already working on the WRONG thing since SATURDAY. That's almost a week of ruined growth. Loser me. Will switch immediately.
Will increasingly make a point of publicly confessing stupid prioritization, because it's my biggest problem.
I love to build, but that's also my achilles heel. I'm too often subconsciously optimizing for the wrong the goal function: short-term satisfaction/fun in building.
Paradoxically, perceived long-term quality in what's being built is what gives me that short-term satisfaction. I enjoy thinking about the problem deeply and building ultra-strong foundations. But in the limit that leads you to reinventing the entire tech stack from the bottom up – yes the deepest tech stack of human history, which includes semiconductor tech itself. "There's plenty of room at the bottom" – not just for product, but also for design ambition. An ambition sponge of doom.
new app with collapsing similar events released #happenlist. Also refined layout further, better utilizing screen space. Every pixel matters! The app's main screen is the central product.
Redefined the key metric of #happenlist. Added backend basis for collapsing event groups into a single row in the main list
With some distance from building this morning, a “Most Important Question” in the back of my mind on what of #happenlist to work on next after yesterday’s improvements is slowly crystallizing its answers! If you don’t know MIQ, look it up. My mind has worked like this for years and has done its best work with it, but I was only made conscious of the pattern recently.
Yesterday #happenlist went through a significant improvement. It got much better, much more useful, but it made new problems evident.
while showering this morning, figured out what was must be causing the deadlock in my promptgun library if an ai call http-fails. it was in my asyncronous reactor library for node, monoflux. Walked to the computer, wrote a test to test if that was the problem – it was. Wrote the fix – it worked. Updated monoflux. Update promptgun. Updated #happenlist. Feels good 🤣.
Family weekend. #happenlist crawler improvement moving into hard mode. Have multiple designs to test. Will need a metric to systematically measure improvements.
Initiating release of a new #happenlist client. Probably will come through tomorrow, by which time I'll have solved some important parts of the server side algorithm. The release will be a "notify the users" moment so they can experience the server side fixes.
Wooow the improvements I've made to Promptgun (npm i promptgun) are incredible.
Solved data-aware streaming of LLM responses and it cleaned up #happenlist code majorly. See screenshot on how to do it.
#promptgrid is still the project Promptgun is tied to, but maybe i'll rename everything to Promptgun as i have the .com. What do you think?
Making impactful fixes to crawler. #happenlist. Started as a vibecoding project so it’s a target rich environment. Having fun.