AI ThoughtsJoonas Tanner

If You're Not Building, You're Not Learning

Access to the newest AI tools is useful. Hands-on building is what turns access into skill.

A person leaving a gray comfort-zone office for a bright futuristic path of building, experimenting, failing, learning, and repeating.

When my dad retired from the military after 30 years, he went back to high school in his fifties and completed his degree alongside my sister.

He liked to joke that he slept with his history book under his pillow so the knowledge would transfer into his brain overnight.

It was funny because everyone knew learning did not work that way.

Yet I think it is easy to approach AI a little like that. Get access to the newest tools, attend a few presentations, use them occasionally, and assume the skill will somehow accumulate.

It does not. If you are not building, you are not learning.

This weekend I burned through the weekly quotas on both my $200 Codex and Claude plans. I do not consider that an achievement. Usage is not a score. What mattered was what happened while I was using them: building, testing, debugging, changing direction, trying new workflows, and finding the edges of what the tools can actually do.

I am learning a lot in my new role at work too, but large companies naturally have security requirements, established architecture, production systems, and delivery processes that cannot change overnight.

My personal projects give me a different kind of laboratory. I can try a new model or tool the day it appears. I can replace an approach that stopped working. I can experiment across product, design, engineering, testing, visuals, and release without waiting for an organization to adopt the same thing.

That freedom helps me stay close to the cutting edge.

This is not an argument that everyone should spend their nights and weekends coding. People have different responsibilities, energy, and time. Building can happen at work, in a prototype, in a hackathon, in an open-source project, or in something small you make for yourself.

The point is that hands-on practice matters.

AI is changing software engineering quickly. Keeping the same workflow you used for the last ten years and simply having access to modern tools is unlikely to keep your skills current.

Access is not experience. A subscription is not a skill set.

You have to build.

Joonas


← Back to all posts