LLM-Induced Brainrot
Agentic engineering has made it easy to outsource our thinking to LLMs—and that quietly rots your brain. A personal regimen for reversing the decay: chess, hard thinking, and maths…
Sami Hindi
Now that agentic engineering is the norm, you need to redirect your brain toward other activities. Otherwise, this rapid cognitive decay will have devastating consequences. Not just on the code you write, but on you as a person.
I've noticed this first with myself. I stopped trying to think altogether. I outsourced my thinking and behavior to LLMs.
I call it LLM-induced Brainrot.
You can undo its effects, but the longer you wait the harder it gets. It also doesn't help that most of us feel some kind of superiority when we talk about how actively we're integrating this new AI technology into our lives, not realizing that we're not setting any effective boundaries.
And eventually, you'll stop. Stop with reasoning, stop solving difficult problems yourself—and in the process of doing so, literally rotting your brain. It's dangerous to think LLMs are entirely new and that none of the concepts from the 'old world' apply. Most of them still very much do.
But how do you reverse the brainrot? There are a lot of ways to go about this, but here is the regimen I'm using to get my brain accustomed to this new way of utilizing technology:
- Play chess. This might sound silly, but re-activating your brain in a computation-intensive game like Chess helps you get adjusted to difficult tasks again.
- Think extremely hard about the problem you're trying to solve. Now that implementing ideas becomes cheap, ideas matter the most. Forget the talk about "Execution is more important than the idea" for a second. You need to effectively spend 80% of your time brainstorming, reasoning, and polishing your idea before you even attempt to build it. And this may not sound "attractive", but this is what a technological revolution does in its essence. It automates part of your life, so that you can focus on the other, now more important, aspects. And the more you do it, the more attractive it becomes (mostly).
- Do Maths. This goes back to the first point of playing chess (stimulating your brain with difficult problems, etc.), but it also goes beyond that. We are in a period of time where we are research-bottlenecked. That may have not been true ~20 years ago, but it is definitely the case today. This means that we need humans in the loop, thinking extremely hard about what research we should focus on, and direct all our efforts towards solving that problem. Maths helps humanity discover new things, especially in the age where we can train very specific models to do the research for us.
This will not be easy, and I myself am still very much in the process of undoing the effects that I've allowed LLMs to give me. Keyword: allowed. It's not Claude's fault my brain is rotting, it's mine.
Don't shove responsibility onto machines, even if it does end up being their fault. It takes away your power to do something about it.
PS: Written by myself, no AI used.