this is spot on. llms at work aren’t about making old workflows faster (or even automating them). they’re about discovering entirely new ways of working that weren’t possible before. reinvention v optimizatikn

spent last week in france. contrary to the online hivemind, it’s still a lovely place full of wonderful ppl & culture. sometimes i wonder if markets will ever produce anything with that kind of depth again. then again, ppl prob said the same 1,000 years ago.

half baked thought: passive computing is a new medium

i give grokbot an ungodly amount of unfiltered, unstructured product direction. 5m rambles on feats, ui to iterate on, remapping of core ux. it just delivers. taking the technical layer out and all that's left is direction. r feels like software is only just getting started.

LFGGGG. Love to see the experiment continue 🏌️🏌️🏌️

first nounish idea in two years hit me today, let’s see if it has legs

design minions catching up to coding minions

the vast majority of ppl will be more than ok w a free plan that has access to basic models. just means ads-esque monetization are the only way for the big labs

i use llms every day and feel lucky to have been born just in time to see it happen. it truly is the invention of our times. but it's just tech. nothing about is fundamentally diff than everything else that came before it. it's a continuation of man inventing fire. sure the curve is of acceleration is steeper and the desire to be fully engulfed by it is harder to fight. but it's just tech. what is most human lives in a place where neither tech nor markets can go. and what a wonderful thought it is to know that nothing but ourselves stands in the way of basking in it. "please use ai" by shawn smucker. please give it a read.

prob a good time to re-read this bb

my personal expense tracker app is getting ready for review. i promise you it does just one thing really well: answer the q “where did my money go?” (it also does not enable restaurant reservations)

not a believer in "ai will gives us back time". the work is infinite.. there’s no amount of speed that gets you faster to a place that doesn’t exist.

2026 weird little guys are not weird enough - go harder

clippy was so ahead of its time re: ux. not part of the ui itself, but a companion to the user navigating it.

this is such a good take. if you are good at anything, use this magic-like tech called ai to build anything useful. if its useful for you, it can be useful for other people. but just do it

the part of me that loves this lives in abstractions. the part that doesn’t lives in the body. there’s a paradox in augmenting exploration: the moment you put technology between a kid and the world, you change the thing you’re trying to preserve.