I joined Twitter in 2007 and my brain slowly morphed over the next 15 years from hopelessly verbose to nihilistically pithy. I've kicked the Twitter habit, but the takes keep flowing. That's why I post them here and format them as a social network of one. You're welcome to bookmark any of these takes, though I'm not sure why you would.
By the way, the hearts and like counts are fake. They're just there to make you feel safe.
Use chatbots as tools, not friends. If you use ChatGPT or Claude, turn off memory and the ability to reference past chats. It only wastes context, assuming present-you wants what past-you wants.
Turned it off months ago and responses are way better. When you have memory enabled you effectively never one-shot anything. Every request is inherently multi-shot.
I now know three people who've decided to return iPhone 17 Pro and buy iPhone Air instead. Love to #influence people. justin.searls.co/posts/why-i-bought-the-iphone-air/
My coding agent started stopping to tell me to run the tests instead of running them itself. Kept happening. Got mad. Replaced my 800-word AGENTS.md with "I'm not your mom."
It started running the tests.
Generative AI is like the "Draw the Rest of the Owl" meme. Before, I'd draw two circles and give up. Now, I draw the two circles and it can get me something I can iterate on and ship.
The hard part shifts from drawing the owl to getting the two circles right—success depends on the quality of the foundation you lay for it.
For a real weird time, get yourself an MRI of your thumb.
Seeing a lot of Fractional CTOs and CFOs out there, but limiting your market to people who understand fractions is foolish. That's why I'm a Decimal CTO.
TFW a virtually identical app to the one you've been working on for months was basically already created by someone else. gg echofeed.app
Easiest smell test of AI features in software:
If it's not one of those, it was probably shoehorned in because somebody was told to go do an AI.
This post by Danny Bolella perfectly captures my reasoning for getting started as an Apple developer ini 2025: after 10+ years of transition, you can finally build things without leaving one foot behind in the old world: captainswiftui.substack.com/p/the-great-shift-in-apple-development
Smart change Apple Software Update: if you were subscribed to the OS 26 public beta, you have been automatically unenrolled now that it's out. In past years, users would (usually unwittingly) remain on the public beta channel for every point release forever.
Less great: a bug preventing those devices to enroll in developer betas. 🤦♂️
Playing Super Mario Odyssey and just realized you collect moons because in Sunshine you collected suns.
The cowards at Nintendo should have called it Super Mario Moonshine.
So glad iPhone Pro goes to a 4x telephoto camera. The 5x lens was a real reach.
Thanks to exercise and clean living, my resting heart rate is 35-45. Got an EKG, all looks good.
Cardiologist: damn you're incredibly healthy
Also cardiologist: here are four follow-up appointments to blow through that pesky deductible
Feel like Becky's podcast is starting to hit its stride. Really enjoyed today's episode.*
*Conflict of interest: I am married to Becky and therefore predisposed to judging it more harshly gram.betterwithbecky.com/podcasts/9
Just fired up my Oculus Quest 2 for the first time in two years and—this will shock you—my $3500 Vision Pro has completely spoiled me.
How did I ever use this thing?
There are fully six podcasts devoted to Curb Your Enthusiasm and not a single one is named Curb Appeal.
So much for that intelligent audience.
While I'm complaining about LLMs, another one: the overwhelming preference for creating dead code by keeping around old code paths "for compatibility" in case anyone depends on it, despite their being not only duplicative but literally unreachable. Another searing indictment on the incompetence of the countless professional programmers whose work served as training data.
One of the most pernicious habits of LLMs (that I can simply never get them to stop doing) is to sprinkle in useless code comments everywhere. No amount of prompting or instructions ever really helps.
Says a lot about the code they were trained on.
GPT-5 + Codex is so fast that when I expressed suspicion that a script was returning too few results (via | wc -l), Codex corrected me that I should have passed --count instead. Sure enough, that worked.
Checked git status and realized Codex implemented the --count flag in the script concurrently as it corrected me for not having used it! Gaslit by a robot!