The case for cognitive friction.

Brain fog is a common symptom today. The democratization of technology has brought infinite information and an abundance of choices. Though they seem positive, they can become sources of cognitive overload.

When we need to remember something, we write it down. If we need to control the time of cooking, we set an alarm. These are examples of cognitive offloading, which help free up space in our minds for other thinking and responsibilities. Used this way, offloading is adaptive and often essential. The risk appears when the tool no longer supports our thinking but replaces it. Using AI agents for the same purpose can paradoxically turn offloading into cognitive overload. The problem is the repeated removal of the effort required to process, structure, and retain information.

As an example, when you go to therapy your psychologist rarely gives you direct solutions. You need to work your way through the problem, and the specialist is there to give tools that can help achieve that.

Give a man a fish, and you feed him for a day; teach a man to fish, and you feed him for a lifetime.

Now, if you decide you need an ad-hoc solution you may turn to AI, which will always, always give you an answer. It won't be a solution you worked for, sculpted piece by piece, but one handed to you, easily enough that you could ask for another, and another, and another.

This is when cognitive overload smoothly enters your life. The given amount of information is difficult to process and reflect on. It felt efficient to hand off the problem and receive a solution without the work. The memory muscle was never trained, so it atrophies. Moments later — fog. The answer is blurry, it's there somewhere, but difficult to articulate.

The same pattern appears wherever we use agentic tools to acquire knowledge. Because memory is a muscle, it only strengthens when we actively work it.

So, by making things easier, we risk making our brains lazier. Might making them more demanding, in carefully chosen places, enhance our thinking instead?

Making things deliberately difficult to use runs against the usual goal of a good User Experience, but the idea of a seamful design (in contrast to seamless) that embraces imperfection could have a positive impact on our learning, understanding, and memorizing.

Friction by design is a method of intentionally adding resistance into the experience. The goal of it is to create reflection and prevent users from acting carelessly. It is called good friction because it preserves cognitive effort.

If memory and thinking are trained through resistance, then the tools we build might benefit from occasionally working against ease on purpose. The point is not to make everything harder, but to introduce small, deliberate friction in the AI-workflow where we now reach for instant answers.

The fog doesn't have to be permanent. It only takes putting the muscle back to work, one deliberate rep at a time. A few modest experiments could test this:

• In writing: require the user to sketch their own outline or argument before receiving AI assistance.
• In learning: ask the student to recall or explain a concept before requesting a summary.
• In searching: have the user read the proposed sources and form a provisional view before requesting a synthesis.

As small interventions, to restore the effort that turns information into knowledge.