This course is not about managing your agent’s context window or teaching you all the shortcuts and tricks in Claude Code. It teaches you a timeless approach to building with AI, regardless of what tool you use, or how the technology evolves.
This way you can come back to the course time and time again, be reminded of the core ideas, and apply them to your work even years later.
The three principles below are woven into the lessons in the course, they are the foundation of everything you will learn. We will learn more about them in the upcoming lessons, but please read through them carefully. They are important to me, and I hope they will become important to you as well.
Friction slowly goes away as we use AI more and more. We can build everything and build it fast. We no longer have to think through the possible options and choose the best one. We can build them all without thinking too much about it.
This is not what we want. We want to wrestle with the problem and come up with the best solution. We want to think through the possible options and choose the best one. We want to be able to justify our decisions and explain them to others.
As you invite more agents to your work, and create “software factories” to be more productive than ever, you’ll soon outsource everything to an agent, even the thinking.
The work you do will feel frictionless, it will feel amazing on the surface, but once you go a bit deeper, it’ll feel extremely shallow. It will no longer remind you of you and the work you aspire to. You will turn into just another extremely efficient producer of things, but not a creator like you were before.
You shouldn’t add friction to everything you do though, you need to be selective. More on that in the upcoming lessons.
Whether you use Codex, Claude Code, or OpenCode, it does not drastically change your results. Same goes for models, Astra, Fable, Opus or Grok, it does not matter as much as you think.
Some models are better at say 3D than others, that’s true. Some people say that Claude Code is miles better than OpenCode and vice versa. We pay so much attention to all of this while it usually gives us marginal gains at most.
Same with people talking about how they run lots of agents at once, pushing 100s of PRs a day. Don’t feel bad if you don’t do that. You don’t need to, I don’t do that either.
These people also usually don’t have strong enough proof to back their software factories, their products are not significantly better than the competition in my opinion. They might ship more, but is that the measure of success?
Keep in mind that people pushing this narrative are usually working at AI companies, companies that sell tokens as a business. They are incentivized to convince you to use more tokens.
To sum this up: focus on what matters, your output. Don’t get caught up in the hype. If there’s a new revolutionary way of using AI, you will hear about it. But keep in mind that most of the excitement online is usually just noise. Noise that offers marginal gains at best, but mostly just wastes your time and energy.
The amount of information available was already overwhelming for a while, but now with AI, it’s through the roof. It’s hard to know what to pay attention to, and it’s even harder to know who is actually credible, since now with AI, everyone can fake expertise easily.
One of the most important skills in the AI era is being able to filter through the noise and find the good stuff. This skill compounds fast, since you can absorb the right ideas easily with AI and apply them to your own work.
This is one of the most important skills, a timeless one as well. We will learn about how to develop this intuition, and how to properly absorb the right information in the upcoming lessons.