You Can't Afford to Dabble Anymore
I'm writing this plugged straight into a terminal.
Not a document. Not a polished editor with a blinking cursor and a word count in the corner. A terminal. The plain text window your computer has always had, the one most people never open in their lives. That's where I work now. That's where this post is being written.
I never studied computer science. I don't have an engineering background. A few years ago I couldn't have told you what a terminal was for. Today I do something like 95% of my work from one. If that sounds like a strange place to end up for a non-technical person, good. That's the whole point.
Here's where I've landed in the summer of 2026: if you do knowledge work for a living, you can't afford to dabble anymore.
By dabbling I mean what most people are doing. Opening ChatGPT once a week to clean up an email. Asking it to summarize a PDF. Treating these models like a slightly better search box and then closing the tab. That was fine in 2023. It's not fine now.
The newest models from Anthropic, OpenAI, and Google aren't a slightly better search box. They're a multiplier on everything a knowledge worker does. And the gap between the people who understand that and the people who don't is widening every single month. If you're not spending at least an hour or two a day actually working with these tools, you're not standing still. You're falling behind people who are.
Let me explain what I actually do, because the words scare people off before they understand how simple it is.
A terminal is the window where you talk to your computer in plain instructions instead of clicking around menus. VS Code is a free editor made by Microsoft, the place where the actual files a project is built from get opened, read, and changed. Those two things used to belong to programmers and nobody else.
Inside the terminal I run a tool called Claude Code. I describe what I want in normal language, the same way I'd explain it to a colleague. It writes the code, edits the files, runs the commands, checks its own work. I steer, it builds. I don't write the code. I decide what gets built and whether it's any good.
That's the part that took me a while to believe. You don't need to know how to code. You need to know what you want, and you need to be able to look at the result and say "no, that's not it." The machine handles the rest.
If you want to start somewhere, start with Claude Code in a terminal. It's the fastest, most direct way to work I've found, and the fact that a non-technical person can live inside it is the best argument I have. If the terminal feels like too much, at least start with Claude itself, the regular chat app. It's a friendlier wrapper around the same intelligence. Either way, the rule is the same: use it daily, for real work, until it stops feeling like a novelty and starts feeling like a limb.
A few days ago something happened that I keep thinking about.
The US government told Anthropic to cut off access to its two most capable models, Fable 5 and Mythos 5, for any foreign national, citing national security. Anthropic disagreed publicly, then shut both models off for everyone to stay compliant. One of them, Mythos, is so good at finding flaws in software that the worry was it could become a weapon in the wrong hands.
Whatever you think about that decision, look at what it tells you. We've reached the point where a government looks at a commercial AI model, one that was available to hundreds of millions of ordinary people, and decides it's powerful enough to be a matter of national security. That's the curve we're on. These tools are getting strong enough that the people building them are no longer fully in control of who gets to use them.
If that doesn't make you want to understand this technology, I don't know what will.
I'll say the uncomfortable part out loud, because it's where I've landed.
Work is going to become more like a sport. Something a certain group of people do at a high level, by choice, because they're good at it and they want to compete. Watching how much I get done as a solo founder with these models, I don't see how it goes any other way. A huge amount of knowledge work can already be automated, and the amount only grows.
That isn't a grim prediction. It's the opposite. If the standard of living holds, and I think it will, then plenty of people simply won't need to work the way we do now. They'll make art. They'll raise their kids. They'll watch TV all afternoon if that's what they want. None of that is lesser. One of the things I care about most is that nobody suffers unnecessarily, and a world where every person gets to spend their days doing what they actually want is a world with a lot less suffering in it.
But if you're in your thirties, forties, or fifties, and you still want to be in the game, to keep adding value as a knowledge worker, then this is the moment to take it seriously. The people who stay in that game will be the ones who kept their minds sharp.
Which is the other thing I've gone back to lately: reading.
Not skimming. Reading. We're living through a stretch of history where a handful of AI companies are quietly deciding what the future looks like, and I want to understand the people doing the deciding. So I'm reading about them.
Right now I'm working through The Scaling Curve by Claude St. John, about Dario Amodei and Anthropic and the race to build something more capable than us. I started it this week and I plan to finish it before the week is out. Alongside it I'm reading A Mind at Play by Jimmy Soni, about Claude Shannon, the man who basically invented information itself. I've got The Information by James Gleick waiting, and The Book of Elon by Eric Jorgenson, Musk in his own words.
That last one matters more than it might seem. A few days ago Elon Musk became the world's first trillionaire on paper, off the back of the SpaceX IPO. You can have whatever opinion you like about the man. I'd still argue it's worth downloading his mental models into your own head, the same way it's worth knowing how Charlie Munger thinks. The world is getting more competitive, not less. Filling your head with the sharpest thinking you can find is how you keep up.
Read the right things. By right I mean the things that actually make you sharper, that change how you think, not the things that just pass the time.
A bit of housekeeping, since this is a Sunday note and I like to keep the people who follow along in the loop.
A few months ago my attention was scattered across five or six projects. I've pulled that way in. Two things have my focus now. The first is SailWP, the WordPress theme that makes WordPress fun and easy and ready for 2026 again. The second is LinkPulse, my own answer for affiliates who are quietly leaking revenue they don't even know they're losing.
Start24 is on the back burner, but it isn't going anywhere. It's still one of my most important income streams, and I'll keep maintaining it and building on it. It helps people who are right at the beginning, making their very first website, get that first site live. I'm not letting it drift.
So that's the Sunday. I'm writing it from a terminal, reading about the people building the models I write with, and trying to stay in a game that gets harder every month.
The models will keep getting better. The window where knowing how to use them well is still rare won't stay open forever. Keep your mind sharp. Read the right things. And stop dabbling.
If you want to know when I write something new:
