How to think

When I originally saw this tweet, I chuckled.

Then I realized: I do the same thing, and so do the people around me. That is, we outsource our thinking to a machine, which can’t think in the first place (though that fact is a whole separate piece I am working on).

Since the rise of Generative AI, what I caught myself doing is using tools like ChatGPT or Claude to go through problems. Not as a help, but instead had it spit out an answer that I then (at times blindly) adopted as my own solution.

And going by that post above, and the anecdotal evidence I have, I am not alone in this.

This is not thinking. Again, the machine cannot think. It can only match patterns and emulate writing. But thanks to increasingly sophisticated models, the solution the machine gives us seems like the solution we were looking for.

But if I am not the one thinking, and thus not the one solving the problem (because problem-solving is what thinking ultimately is), I have learned nothing. I have just taken another one's thoughts (and again, the machine cannot think) as my own. I see little value in this.

This phenomenon is, however, not a recent problem, even though Generative AI has exacerbated it. Outsourcing our thinking to other things, or people, is something humanity has been doing forever.

Before Generative AI, we outsourced our thinking to influencers and whatever the algorithmic timelines fed us. Before that, it was to politicians, celebrities, and other people in power. Before that, it was the churches. And before that, it was the shamans. (To be fair, people still do this.)

But at least in these examples, the thinker we outsource to is human. We can, most often, deduce what their agenda is. But what is the agenda of a machine that has been trained by a group of people who probably don’t even understand how that machine works in the first place?

I believe that in this age, at a time when we get inundated with information from all directions, the ability to think is the most important skill we have.

I expect that when, and if, the AI revolution arrives, people who have the ability to think are the ones who will not be left behind. Thinkers will be the ones who will thrive in these uncertain times.

And this is how to think:

I see four parts that are necessary for thought. You need to cultivate all four, because one or two alone may not be enough to form your best thinking.

These are:

Thinking in silence

AI, algorithmic timelines, and generally just the noise1 we live in, don’t give us the space to think. They hijack our attention and concentration.

This is our fault. Whenever we have the slightest moment of silence—and we call that boredom—we try to fill the void with whatever we can find.

But there is a reason you have your best thoughts under the shower, or as soon as your head hits the pillow. These might be the only moments you experience true silence and boredom.

When we manage to turn off the outside world, we are able to listen to our inner voice. That is thinking. That voice that speaks to you, at times maybe roughly, though that is for another essay, is what thinking is.2

By listening and talking to the inner voice, we can give it problems to solve. We can mentally go through the steps and let our minds untangle whatever we are currently working on. If we feed it with the correct pieces, and let it do its job without interruption, it’ll allow us to solve the puzzle.

This is hard. Thinking is an active skill (though there is a passive element to that, more soon) that burns a lot of energy. The brain alone consumes, on average, around 400 calories per day. To give you an idea: 30 minutes of running burns the same amount. (So feed your brain the nutrients, exercise, and rest it needs.)

I am a strong believer in cultivating silence to let our minds go wild and start forming thoughts. It is not easy, though, modern civilization likes to flood us with distractions. Therefore, I try to find moments throughout the day where I embrace silence. (And I am not talking about absolute silence like you have in outer space. You don’t need 0 decibels; rather, what you need is to not have inputs. White noise is completely fine and might even be beneficial to some. If I struggle with sounds, I listen to a mix of white noise and thunderstorms.)

But embracing silence is hard for me. I struggle with this because I have the tendency to fill the silence with... something. Anything. Not necessarily because I’m afraid of the silence, but because boredom is at times painful. Boredom is just so... boring.

I’m not used to it anymore, so I have to force myself to accept it. And only then can I sit in silence and let my mind work. And every time I give it the space it needs, I am surprised by what that squishy thing in my skull is capable of.

Thinking through inspiration

While the building block above shows how to create space for thinking, it’s inspiration, I believe, that sparks thought in the first place.

Not every thought is worth something. I doubt this is a surprise to you, but if we’d follow every thought we’d ever have life would be pretty fucking weird.

To succeed at thinking, we need to feed our minds the necessary material to refine what happens up there.

This happens through a process I (and I probably stole it) call cross-pollination.

Cross-pollination is when you take a whole bunch of Lego bricks from all kinds of different sources to build your own castle in your mind.

You achieve this by consuming broadly.

But not all consumption is equal.

There is a reason everyone talks about brain rot currently, because mindlessly scrolling through TikTok and watching people do whatever the algorithm gets them views, is not the type of consumption I am talking about.

Rather, we’re talking about content (and it can still happen on TikTok, the medium is NOT the problem) that challenges you.

For me, this content primarily exists as the written word. It is the reason why I read as much as I do. But I also find it in blog posts like these, or newsletters, or at times even on text-based social media like Bluesky or Threads (though let me be real, this is the exception, most content on there is mediocre).

You can find that content also in multimedia formats, be it podcasts, YouTube videos, or (good!) TikTok shorts. Or you find it as a little nugget in some random TV show or movie. Or while talking to other people, or observing nature.

What is important here is that you consume actively. Not necessarily to learn every time you look at something, but by spending focused time with the media.

And yes, for that, you need to put away your phone, turn off your gaming console, or whatever else you are currently doing. NO multitasking. We all know by now that multitasking doesn’t exist. Sit with the material, consume it, and let it feed your thoughts with new Lego bricks.

One very important thing, however, is this: don’t only consume things with which you agree or that you already believe. All this does is feed your idiocy (and we are all idiots) and enforce negative cycles.

Consume stuff you hate. Consume what the enemy created, whoever that enemy is (and then ask yourself, why do you have enemies?). Consume things that are uncomfortable because they might show you truths you want to hide from. Consume broadly and widely, and outside of your comfort zone, because it gives you perspective and shows you things you may not have known.

I am not saying you need to adopt these views. Not if you fundamentally disagree with them, and especially not if they are just plain wrong. Bigots are bigots (and I believe they are bigots because they do not consume what their “enemies” create). But this at least shows you what not to think about.

This is crucial, too. This is anti-thinking, another part of having “good” thoughts. But how do you know what to anti-think if you don’t know what is out there?

Thinking through writing

Yes, I am biased. But I believe that writing is the other necessary skill to succeed in our current times.

Because what writing allows is to sort and distil the thinking you do, break it down into pieces and recombine it with other stuff.

As long as the thinking just stays in your mind, I’d argue that it is worthless. This is especially true for ideas. Everyone has ideas. The world certainly does not lack ideas.

Ideas are not worth anything if they do not lead to future steps.

The first step is to write it down. Because writing is the one other magic trick humans possess.

And before you tell me that Generative AI is taking this from us: LLMs do not write.

What they do might look like writing, it might feel like writing, but it is not writing. Instead, GenAI outputs text, syntactically flawless text, yes, but devoid of any substance. The machine just breaks down writing into a mathematical formula3, robbing writing of all that makes it magical. (And a lot of us lack the necessary taste to understand that this writing is simply not good. Grammatically correct ≠ good.)

So you need to write yourself. And as the screenshot at the beginning of this essay ironically shows, even writing down your problem as an AI prompt clarifies your thought.

There are two ways to solve problems through writing, and I alternate between the two of them: they are writing slowly, and writing fast.

Writing slowly

I’d argue that to write slowly you have to write by hand. Be it on a piece of paper, or like I do, on one of these fancy e-ink devices.

But through writing by hand, you are forced to slow down, simply because your hand cannot catch up to the speed of your thinking. And this allows you to “de-jumble” the mess in your head before you put it down on paper.

This blog post was first brainstormed on the equivalent of two sheets of A4 paper, and what came out was basically a completely finished post that just needed a bit of polishing (to transform bullet points into proper prose, for example).

I write most of my blog posts this way. I also write my journal by hand every morning, and most of my notes are handwritten, too.

Again, this is to make sense of what is in my head, by giving me the space (and the silence, there are no inputs when I do this) to think through things.

(In theory, you could also use an old-school typewriter. Because if you type too fast on that thing, you jam the keys. This is a great analogy because if you write too fast by hand, you jam your brain.)

Writing fast

Another practice I follow is what I call the brain dump. This has to happen on a computer, either by typing if you are a fast typist or maybe by recording a voice note.

The value of the brain dump is by “emptying” your mind. The goal is not to form perfectly finished nuggets of thought but instead to unload all that is in your mind, all that is taking up your mental bandwidth.

Often, what comes out of a brain dump session is not truly valuable if looked at through a vacuum. It is important that you don’t filter and instead write everything down that comes up, unedited and raw.

When you look at this brain dump, you’ll realize that most of it is trash. That is ok, that is the point of the exercise. You want to get the trash out of your head.

But with a certain distance (I never read the brain dumps the day I wrote them), you may find certain specks of gold. Here and there, you see a nugget that, if you disassemble it, might lead to something. And then I’d suggest you take that nugget and go through it by writing by hand.

Thinking by not (actively) thinking

Because thinking is problem-solving, in theory, the result of thought is a solved problem.

Sometimes you can’t solve the problem when you actively think about it. You just can’t find the solution, no matter how much time you spend on it.

In these cases, stop. Take some distance. Let it rest, do something completely different, and ignore it for a few hours or days.

You may have experienced this before. You struggled for hours to come up with a solution, kept failing, and ultimately gave up.

But then, in the unlikeliest of situations, you had the epiphany you waited for. The complete solution to your problem suddenly came up in your mind as if planted there by some alien life form when you were not paying attention.

This is thinking by not thinking. It is passive. It happens without you forcing it, in the subconscious, while you do other things. I don’t know why it happens. I don’t understand what processes run in our subconscious mind in the background, I only know that I’ve experienced this before.

As a writer, the way I use it is to never hit publish on bigger pieces (like this one) the day I wrote them. I often let them sit and ripen in the back of my mind. When I sit down with them again, I often perceive things I hadn’t before.

The same goes when I struggle to fix a problem at work. Giving myself the space to not think about it is apparently what I need to solve the toughest of problems.

So sometimes, don’t think. Some people seem to be really good at this.

Tools for Thinking

I’m planning to expand this section into a separate post in the future, but here are a bunch of tools and tricks I rely on to help my thinking.

While the above steps are the basis needed to think in the first place, the tools below are what help me have “better” thoughts.

There are many more tools in my toolset, but these are the ones I (try to) rely on the most. I’ll expand this into a separate post down the line, so subscribe to the newsletter or RSS feed to get notified when it goes live!

Thinking is the most crucial skill we need to develop amidst our current, uncertain times. It will help us make sense of the mess of the world, and especially of the mess in our minds.

By becoming better thinkers, I’d argue we become better humans. And by becoming better humans, we’ll be able to make the world a better place.

None of this is easy. It requires a vast amount of effort from us, not only to take the time to think or improve our thinking, but also to reject what interferes with it.

It is probably why a lot of people will not do this. Instead, they might complain, shout at the clouds or simply give up. It is, after all, easier to feel defeatist than to struggle.

Those of us who hone this skill (and thinking is ultimately a skill) will learn a superpower that brings us ahead of the majority.

It’ll make us superhuman, and I strongly believe this.

So, go and practice thinking.

Some notes on AI: I bashed Generative AI, LLMs and algorithmic timelines a lot in this post.

The reason is I strongly believe we should not be offloading the skills that make us human to machines.

But I still use ChatGPT on a near-daily basis. The difference is that I (now) use it to complement my thinking. I use it for research (and then fact-check, because it still hallucinates a lot), I use it as a learning tool, or to see things from different angles by actively asking it to do so. It often fails, but sometimes it helps me.

Generative AI is a tool we need to learn how to use. I keep comparing LLMs to a friend who has a photographic memory and remembers everything. But he is also just plain stupid. He makes shit up. He doesn’t know what he is talking about, but just parrots what he learned by heart. (And memorizing ≠ understanding.) Sometimes he parrots something really intelligent, but that is more a coincidence than anything else. We just give this randomness more weight than we should, as we find it “magical”.

And as for algorithmic timelines: they are mostly shit. Their only worth is if you use them as a marketing tool.

  1. This is not a new development, however. The Stoic philosopher Seneca, back in 62-64 AD, already complained about how noisy Ancient Rome was back then (On Quiet and Study). And he didn’t even have Instagram, TikTok or ChatGPT. ↩

  2. Some people do not have an inner voice. I cannot imagine what that would be like, as mine never shuts up. But I’d love to hear from you. ↩

  3. The way LLMs “write” is by calculating what word is most likely to follow the preceding one. But since it was trained on gazillions of data (so-called tokens), it’s rather good at emulating the way humans write. But two things: since we’ve just argued that writing is thinking, and thinking is a human practice, we cannot call what the machine outputs as writing. These machines don’t understand meaning, they excel in (statistical) patterns. And second, the creators of these machines want us to believe that there is more magic in that output than there is. If they can sell us the idea that the machine has created something original by thinking, we’ll have more faith in these tools and thus will throw money in their direction. And they need a shit ton of money. ↩