Don't use AI to avoid thinking. Use it to think better.
TLDR: Use AI to become smarter instead of just more productive
Using LLMs for Agentic workflows and work delegation in general seems to be dominating online mind-share currently. However, I find that the most useful feature of using AI professionally and personally is not as a tool I outsource work or research to, but as a teacher in the Socratic mode. Don’t use AI to remove the need for you to think; use it instead as a partner to sharpen your ability to think. The better use for AI is to use it to learn and improve yourself rather than to diminish your mental capacity through over-delegation.
Traditional books and tutorials have their advantages, but by their nature they’re one size fits all. The author has to guess where you might be confused and explain accordingly. An LLM on the other hand can adapt in real time to your learning journey. You can have it explain something in three different ways, ask it for examples which are relevant to you or reverse roles and have it ask you questions till the concept finally clicks. Instead of just accepting an explanation as is, you ask the AI questions and clarifications and dig into the subject matter until you have fully absorbed the information.
Some time ago, I was trying to figure out how to use podman to run a couple of containers and getting nowhere despite reading some tutorials on the subject (which all seemed suspiciously similar). A few minutes with Google Gemini and I had a clear understanding of the way forward as well as a better grasp of the limitations I would need to work around. The breakthrough wasn’t that Gemini knew more than any one individual tutorial. Rather, it was that I could interrogate it whenever I was confused till I was clear on the subject.
Similarly, there was a particular database concept I was unclear on and for which I had never been able to find a proper explanation online. After some back and forth with an LLM, I finally understood what it was all about.
What I’m trying to say here is that we should be using AI as a launchpad rather than a crutch. When you close that chat window, you should step away more knowledgeable and capable than when you opened it. Use it like a responsive librarian or a sometimes fallible research assistant. Build your own understanding and don’t fall into the trap of becoming an AI ‘meat proxy’ - someone who blindly recites AI output and adds nothing but transmission error. Don’t abdicate your intellect.
Outsourcing work to an AI is fine. Outsourcing your understanding is fatal.