Why Thinking, Fast and Slow Is Harder for Me to Read After AI

I started reading Daniel Kahneman’s “Thinking, Fast and Slow” about a year ago, but I had to put it away because I wanted to finish other books on my list first.

Everyone who recommended it to me told me how good the book was, and after reading just the first chapter, I really understood why.

Then recently, when I returned to it, I had a strange feeling that I couldn’t focus and that my brain badly wanted to skim through the pages.

I tried and tried to progress further, but couldn’t help myself. I made it about a hundred pages further.

I don’t consider myself a bookworm, but it was unusual, especially with a book I liked so much at the beginning.

Somehow, everything in it felt artificial and mechanical, and I couldn’t figure out why for a long time.

Then recently, while tweaking my AI agent to avoid common slop patterns in its responses, it clicked. The writing style in the book felt like one of the most common AI slop patterns: “It’s not X, it’s Y.”

That’s when I came across the term “contrast framing”. More generally, I’m talking about explaining ideas through contrasts between opposing concepts.

Kahneman uses this a lot:

• System 1 vs System 2: fast, intuitive thinking vs slow, deliberate thinking
• Experiencing self vs remembering self
• Inside view vs outside view

There is nothing wrong with this. In fact, it is one of the reasons the book became so accessible. But I have to say, today this style feels different.

Modern LLMs absolutely love it because it is an easy way to create a sense of clarity. You see tons of generated text following the same patterns:

"X is fast and intuitive. Y is slow and deliberate."

or:

"One system reacts. The other reflects."

Again and again.

After reading enough generated text, my brain started detecting these patterns everywhere. Now, when I see one, I automatically try to extract the idea as quickly as possible and skip ahead.

Kahneman’s style has many of these features.

He often introduces an idea, gives an experiment, explains the experiment, restates the lesson, gives the phenomenon a name, and then summarizes what you should remember.

And it’s really great for learning. But modern AI also produces huge amounts of this structure, often without a good reason.

The form itself has started to feel cheap.

Of course, I wouldn’t say:

"Kahneman writes like ChatGPT."

It’s more that LLMs are good at reproducing this kind of writing, and then exaggerated its most recognizable patterns.

For me, this creates a weird retrospective effect.

Older writing becomes harder to enjoy because of AI, and somehow it feels less valuable when you see the same patterns a thousand times a day in every other post.

We have used a lot of phrases and patterns in our writing for a long time to make ideas clearer and easier to understand, but LLMs have used them so frequently that they start carrying less value.

Generative AI might eventually change how we read writing that existed long before generative AI.

The book itself is great, though.

The combination of its experiments, autobiographical context, intellectual history, and the way Kahneman connects ideas is unique, and generated content is nowhere near it.

It’s just that the internet is now so saturated with AI slop that my brain automatically reacts to the same patterns.

And I suppose this is going to become very common: people will see some older writing and think, “OK, this looks like slop” because AI uses the same style.

And there’s an irony in “Thinking, Fast and Slow” specifically. My reaction is like a Kahneman-style phenomenon itself.

System 1 recognizes the pattern and produces the feeling: “AI slop.” And that feeling arrives before I can even consider whether the paragraph was written in 2011, 1980, or yesterday.