justin․searls․co

My ARSE® process for building apps with coding agents: [A]mbitiously create by handing an agent a HUGE task to do over hours/days [R]age at your computer as you uncover all its stupid decisions [S]culpt the app by fixing things in shorter, bursty sessions…

I think the iPhone Duo might serve me well in the Post-Programming Era.

Normally, the best way to watch an Apple event is the TV app on an Apple TV, but damn if their use of maxed-out HDR for visualizations combined with sudden cuts back to SDR footage didn't make it feel like every shot of a human was dim and washed out.

Gonna go on record one last time and say that I hope Apple names its folding phone the iBook.

Maybe it's because I got to see it in IMAX, but I really liked Christopher Nolan's Castaway reboot.

Biggest knock on Astra so far is that when I raise an issue during a long-running task, it will acknowledge my concern and then stop all work.

First Astra impressions: it's like Fable, but faster.

Direct link to podcast audio file The summer heat has finally broken here in Shizuoka, and I'm enjoying a cool, wet, and cloudy day huddled around my microphone for warmth.

Finally figured out a bulletproof system for landing on the best model for the job.

Asked Codex and Claude Code to each try their hand at a tricky Core Text layout issue for Japanese vertical writing: GPT 5.6 Sol xHigh: 1,581,445 tokens and 2189 lines changed Fable 5 medium: 37,000 tokens and 1 line changed Fable's solution worked, Codex…

/goal is a special command for OpenAI Codex in which it will keep working toward whatever goal you set until it gradually convinces itself that the stated goal is not what you wanted but rather some much easier objective was the real goal all along.

Whenever I need to eke out some extra intelligence from an LLM, I ask it to speak to me in a British accent.

[Translator's note: buffet dining is often referred to as バイキング ("Viking") in Japan.

Talking about a conversation you had with an LLM is like telling people about your dreams: absolutely unacceptable no matter what.

Many of you sent feedback about your experiences with agents at work—but it was all too positive!

Demoed a component of my unreleased memorization app today.

Do you work on a team or at a company where most people are actually using agents in earnest?

The Opposite of Brooks's Law: Removing manpower from a late software project makes it less late.

Now that I stand to gain very little financially from posting correct takes on the Internet, I find myself less motivated to post correct takes on the Internet.

"You can't afford bad engineers anymore." blog.florianherrengt.com/ai-removing-middle-class-software-engineering.html

This week, Anthropic shipped a new messaging feature to Claude Code.

Direct link to podcast audio file We cut our trip a day short, so I found myself with a totally free day with no plans.

Finally, Netflix you can chill.

In 2025, I bought four WD Black 8TB NVMe SSDs, slammed them in a Thunderbolt enclosure, and set them up as a 32TB Raid 0 volume for my Mac Studio.

I realize this is Very Rude, but I've noticed a lot of the people who think AI sucks at writing code never seemed very good at writing code themselves.

I joked last year that by leveraging coding agents I now "played the orchestra", by directing multiple programmers at once.

When I'm AFK, I want to just create issues and reply to them and have an agent spin up, do stuff, and wind down. order_taker seems to do that p well. github.com/searlsco/order_taker

I'd describe my approach to work as, "light-hearted and heavy-handed."

Direct link to podcast audio file The summer heat ended up thawing out my podcast, so I had no choice but to go buy a new microphone and grumble into it for a couple hours.

This is a copy of the Searls of Wisdom newsletter delivered to subscribers on April 10, 2026.

Direct link to podcast audio file The promise of Breaking Change is that with every major version (this being the 60th episode of the program, but only the 53rd such release), I will break something.

Update on using MacBook Neo for dev tasks: this is definitely a device you should to reboot daily, even without Chrome or any Electron apps.

Using MacBook Neo for "real" development work and it's simultaneously juggling: • Running Claude Code in three tabs at once • Compiling multi-package Xcode builds • Automating two iOS Simulators and a Mac build of my app Hasn't missed a beat.

Claude sure goes down a lot for being a product targeting businesses.

Sent the DLSS 5 video to some not-very-online friends and they thought it looked incredible.

Direct link to podcast audio file Quincy Larson over at freeCodeCamp had me on their podcast to discuss how the rapidly changing software industry is impacting junior developers and what they can do about it.

We blew past this milestone without much fanfare, but it bears repeating: building awareness & goodwill by releasing open source no longer makes strategic sense for many companies.

Ben Thompson's latest rests on a single load-bearing assumption: that the harness and the model are tightly coupled, the way Apple's hardware and software are.

Marathon is the first PVP-heavy game to get its hooks in me since… Unreal Tournament in 2001?

This one goes out to all the testing neophytes who only recently realized that it's useful to have an automated means of verifying their code does what it claims to do.

Pro-tip to any devs who only discovered TDD thanks to coding agents: refactoring is inherently directional.

I'm still iterating on my experimental Claude Code verification harness, prove_it.

Claude's electron app for macOS is such a buggy mess that I've uninstalled it and sequestered it to a Safari tab, just like I did to Slack, Discord, etc., six years ago.

I'll admit, it's hard not to get frustrated by all the posts about coding agents going viral lately that are saying the same shit I've been blogging and podcasting for multiple years at this point with little to no fanfare. blog.lmorchard.com/2026/03/11/g…

GPT 4.1?

Want to buy Becky's iPhone Air?

I need a new blog to subscribe to.

Barely a month goes by where I can check the news without being reminded of the Cracked skit about two comedy writers who were hired to run a fake Donald Trump campaign and then accidentally got him elected youtu.be/8qIQbydyHwc?

Direct link to podcast audio file Today, we're joined by a very special guest, Len Testa!

For the extremely narrow Venn diagram of people who love both I Think You Should Leave and Disco Elysium, this is amazing youtu.be/6OKgdTXLWIg?