I have a new idea that everything you should try to do to get ready for AI basically breaks down into three pieces.
The first is deeply understanding how things work, especially the things you care about. If you’re a philosopher or a politician, that means psychology, sociology, politics, government, and all that kind of stuff. If you’re a technical person who likes to build things with your hands, maybe that’s civil engineering. If it’s software engineering, then programming, systems, tech stacks, and all that.
So that’s the first thing: deep understanding. This is a thing you cannot outsource. You cannot sacrifice it. You must pursue and maintain and grow your deep understanding of how things work. And this is arguably the most important of the three.
You can’t tell what’s high quality or not if you don’t deeply understand things. You want to be able to see something and know that it’s bullshit, or see something and know whether it’s high or low quality, because you understand that topic in and out.
The second piece is that you actually have to want something. You have to resonate at a frequency. You have to have some sort of internal desire to see the world differently than it is. You have to want something to exist that doesn’t exist. You have to want something to exist in a different way than it currently does. You have to want the world to be a different way than it is.
That could be anything from the way healthcare is done, to the way the government works, to a business that fixes bike tires that doesn’t exist now and you think should be done a different way.
You have to look at how a thing works today and have a reaction to it—a visceral reaction. You have to deeply want it to be different, or want to create something that doesn’t exist. It’s a drive. It’s a passion.
This is the absolute most human component of this. It’s the reason I define art as the indirect communication of something that matters. And that “matters” component means it matters to humans.
This is the piece a lot of people don’t have, and it’s really the hardest part to get if you don’t have it. One of the worst situations you could possibly be in is to have skills because you understand how things work, but not actually want anything to be different because you don’t care about anything. You don’t want to make anything. You don’t have opinions. And that’s a really bad place to be right now in this phase of human evolution that we’re moving into.
The third piece is capability. Agency. The ability to actually make things happen.
When you’re at step two and you actually want something to exist—because from step one you actually understand how things work and you want them to work a different way, or you want a thing to exist that doesn’t exist yet, or you want to make a piece of art or a piece of technology—now you need the capability to make that happen.
That can include management skills, but I’m largely talking about technical skills that allow you to build. And that’s a broad thing. It could be the ability to create a new garden product, the ability to design a garden, the ability to design a new house style, or a new way of doing solar panels. So it can be technical, but it can also be any sort of skill that allows you to create or change things in the world.
So the first piece allows you to understand the world. The second piece is a desire to see something different in the world. And the third piece is the ability to make a change in the world—to build or create in the world.
Obviously number three is extremely highly tied to AI skills, because AI is the single greatest lever right now for someone to be able to move from steps one and two—from that understanding and desire—to actually taking action and actually changing or creating the thing. These three combine in a very powerful way, and they absolutely depend on each other.
Most problems—at every scale—can be reduced to current-to-desired state transition problems. AI’s Ultimate Use Case: Transition from Current to Desired State (2025)
You can have all the skills for number three, but if you don’t want to make anything, then you’re kind of just doing those skills on behalf of someone else, which can be fine. If you have number three but you don’t have number two or number one—so you don’t really understand how things work, but you’re able to make things, you’re able to execute—that kind of puts you in the category of a manager or a politician. You’re able to get people to do things. What I’m saying is you could have the ability to make things happen or to build things, but not really have ideas around what to build. Usually you do have a decent amount of number one if you have number three, but that’s not strictly always the case.
Not having number two, like I said, is kind of the worst. A lot of people have one and three, but they don’t actually want to make anything. They don’t actually know what to make. They might even wish that they wanted to make something. That’s a very sad place to be.
But the absolute worst is simply not knowing how things work.
The good news is that if you have the desire from number two to know how things work, you can actually study and learn how things work. You can solve the first one pretty easily, especially with AI now. You can just learn how things work. Traditional education is not even strictly required. In fact, it can be a hindrance because it’s so slow. You can learn a massive amount of stuff from YouTube and become a deep subject expert on things by studying, by practicing with AI, and by learning and building yourself through trial and error. So if you don’t have number one, there are lots of ways to go about getting it.
If you don’t have number two, that’s going to take some soul searching. It’s going to take some aggression. It’s going to take going back into yourself when you were a child. You’ve got to ask yourself: did you ever have this fire? Did you ever have passion? Did you ever want to make things? And if so, how do we get back to that? If you never did, well, there are some other exercises you could do there, but it’s going to be more difficult. You can read about meditation, you can maybe experiment with psychedelics or something like that. There are lots of different options, none of which I’m directly recommending—other than maybe poetry from Rilke.
And the third one: if you don’t have the skills to actually make and build and construct things and take action, that’s also a thing you can learn. And it’s a thing you can outsource quite a bit with AI. In fact, your skills at AI are in the third category—and in the first category as well, because deeply studying and learning AI skills starts with deeply understanding and learning AI itself. So I would put that into categories one and three. But essentially, three should be heavily focused on getting extremely good at AI so that you can delegate the parts of number one that you don’t understand while you’re learning them, if it’s one of the things you want to learn.
For example, if you’re not good at design, but you’re good at a whole bunch of other things and have deep understanding of a particular problem space, you can delegate that stuff to AI in number three. Importantly, though, you don’t want to be delegating pieces of the problem or the solution where you need deep knowledge from number one. It’s critical that you stay in touch with that deep understanding of the problem and your solution.
So I would say this trifecta is essentially how to move forward in the time of AI. When everything is changing, focus primarily on deeply understanding the things that you most care about. Deeply understanding the things where you need to be an expert. That’s all the way down—from the physics up—as much as possible. If it’s computers, you better understand hardware, software, CPU, memory, networking, operating systems, applications, programming. The deeper the better.
What you want is a deep fundamental understanding of the whole space that you’re in. Many problems come from people who should have that understanding and actually don’t, and that will cause them to fumble at multiple levels as they’re trying to build a solution.
