justin․searls․co

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.


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. 🤦‍♂️

Copied!

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.

Copied!

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

Copied!

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?

Copied!

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.

Copied!

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.

Copied!

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.

Copied!

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!

Copied!

Pro-tip: the codex CLI can't search the web by default (even if you bypass all sandbox restrictions). You need to explicitly enable --search.

If you ask codex to search the web without that flag, it'll literally guess domain names and try curling their homepages.

Copied!

I wish coding agents came with those Green/Red coasters they give you at Brazilian steakhouses:

🟢 Green: go ahead and pile stuff on my plate
🔴 Red: stop adding, we need to make room first

Copied!

I don't know who needs to hear this, but despite being bare bones from a feature-set perspective, Codex CLI with GPT-5 is much, much better at some coding ecosystems than Claude Code with Opus 4.1/Sonnet.

Codex writes competent Swift that does what I ask, nothing more. Claude hallucinates code all day.

Copied!

TIRED: Rebinding caps lock to escape because you use vim

WIRED: Rebinding caps lock to escape because you use Claude Code

Copied!

With Swift, I'm really speed-running the list of stupid things you do when learning a new language. 3 days ago I wrote a dependency injection framework, 2 days ago I convinced myself I'd found a compiler bug, yesterday I wrote my first macro, today I made a mocking library.

Copied!

The nice thing about server-side LLMs hitting the point of diminishing returns is that it gives local LLMs a chance to catch up and for their utility to approach parity.

Copied!

I would pay so much extra for a version of Claude or ChatGPT that paid the same toll I do whenever I fuck up. Make guilt a stateful property that decays over weeks or months. Trigger simulated self-doubt when similar topics arise. Grant my account bonus GPU-time so the chatbot works ridiculous overtime to make up for its mistakes, just like I would for my boss.

Copied!