AI is changing software engineering, but I do not think the first skill it asks from us is prompting, model selection, or knowing the latest tool.

I think the first skill is simpler: try it yourself.

I Do Not Trust Distance

I have noticed that it is very easy to have an opinion about AI without using it much.

The opinions are familiar. The output is shallow. The code is unreliable. The context is not enough. The cost does not make sense. The hype is bigger than the reality.

Some of that is true. I have seen weak output too. I have seen it miss obvious details and produce work that still needs careful review.

But I also think there is something lazy in judging these tools from a distance.

If I want a real opinion, I need to burn tokens. I need to try the tools on actual work. I need to see what happens when I give them a vague task, a specific task, a bad prompt, a well-structured prompt, a real bug, or a messy draft.

That is where the useful understanding starts.

Trying Changes the Quality of Judgment

Once I started treating AI tools as something to test instead of something to debate, my questions changed.

I stopped asking whether AI is good or bad in some abstract sense. That is not a very useful engineering question.

The better questions are much more practical:

  • Where is it actually faster?
  • Where does it create cleanup work?
  • What kind of context makes the output meaningfully better?
  • Which tasks still depend heavily on my own judgment?
  • When does it help me move, and when does it just make noise?

Those answers do not come from reading other people’s opinions. They come from repeated use.

This Is Just Engineering

In that sense, I do not think AI is special.

When a new database, framework, or platform appears, I do not form a strong opinion only from screenshots and commentary. I try it. I look at the failure modes. I see what breaks under real constraints. I learn where it fits.

AI should be treated the same way.

Trying is not passive curiosity. It is a real engineering skill. It is the willingness to test, compare, inspect, and revise instead of speaking confidently before touching the thing.

Takeaway

The first skill in the AI era is to try it on.

Burn your tokens. Use the tools on real tasks. Build your own judgment from direct experience.

You do not need to be uncritical. In fact, the point is the opposite. Real criticism gets sharper after real use.

But if you are an engineer and you are still standing outside the tools, your opinion is probably too cheap.