Does AI Make Me Dumber? A 12 Month Self Experiment

There is this thesis being discussed everywhere right now. **AI makes us dumber.** We forget how to think because the machine does it for us. A calculator for the brain, and eventually we cannot even do mental arithmetic anymore.

I have been working with AI agents every day for twelve months. Not occasionally, not as a gimmick. Every day, all day. I lead a team of human and digital employees. I delegate tasks to agents, discuss architecture decisions with them, have them write code, create analyses, draft texts.

Have I become dumber? I do not think so. But **I have changed.**

## What I observe

My communication has changed. **Fundamentally.** Before, I could give a vague instruction and assume that an experienced colleague would figure out the rest. With AI, that does not work. Or rather, it works, but the result is unpredictable.

So I learned to be more precise. Specific addressing. Clear context. Measurable outcomes. Naming responsibilities. That sounds like a small thing, but it changes everything. Because suddenly you realize that the vague instructions from before were not always clear for human colleagues either. *They just politely did not ask.*

**I communicate better with people today because I learned to communicate better with AI.**

## Cognitive workload

My brain does different things than twelve months ago. I write less code. I draft fewer emails. I no longer format spreadsheets. These are tasks that used to take up cognitive space. **That space is now free.**

What I do instead: I think about **product visions**. About positioning. About market strategies. About architecture decisions at a level I never had capacity for before because I was busy handling daily operations.

Is that getting dumber? Or is it **freeing up cognitive resources** for higher value tasks?

## Delegation at a new level

I delegate differently today. More clearly. More structured. With **explicit acceptance criteria**. With defined roles and responsibilities. Not because I read a management book, but because working with AI agents **forced me to**.

An agent does exactly what you say. **Not what you mean.** That is sometimes frustrating. But it is also the best leadership training I have ever had.

> If your instruction does not work, it is your instruction that is the problem. Not the agent.

This realization has improved my collaboration with human colleagues just as much.

## More leading, less doing

Six months ago I was writing configuration files myself, monitoring deployments, resolving git conflicts. Today I tell an agent what needs to be done and **review the result**. I have gone from maker to decision maker.

Some would say: *You are losing the craft.* And yes, that is true. But an architect does not need to lay bricks himself to design a good building. He needs to understand **how walls work**. And that understanding has not diminished. On the contrary, it has grown because I now have the overview instead of drowning in details.

## What is really happening

The dumbing down thesis assumes that the tasks we outsource to AI are **the ones that keep us smart**. But is that true?

Does formatting an Excel spreadsheet keep me cognitively fit? Or is it simply work that needs to get done?

What keeps me cognitively fit is **solving problems**. **Making decisions** under uncertainty. **Recognizing patterns** and connections. **Communicating** complex topics. I do all of that **more today than ever before**.

> AI has not freed me from thinking. It has freed me from routine. And that is a big difference.

## My conclusion

I am not saying AI makes us smarter. I am also not saying it makes us dumber. I am saying: **It changes us.** Like every tool we have ever invented.

The printing press did not destroy memory. It made it unnecessary to memorize everything. And that **created space for different thinking**.

AI does the same thing. Only faster. And more radically.

**Those who resist will not stay smarter. They will just be slower.** Those who embrace it will change. And whether that is good or bad is not decided by the technology. But by the person who uses it.

For my part, I have **learned more, built more and achieved more in twelve months than in the five years before**. And I feel anything but dumb.