Start with the thing you do forty times

Today I want to talk about the most boring task on your desk.

Not the interesting problem. The one you have done a hundred times and could describe in your sleep. Something like, repathing external references across a folder of drawings or renaming the same layers on every sheet.

That is where to start. Here is why.

You already know what correct looks like

The hardest part of building a tool is not writing it. It is knowing when it is wrong.

On a task you have done a hundred times, you can look at the output and see immediately that the third stall is six inches off. You do not reason about it. You just see it.

That instinct is the whole game. AI will produce something that looks entirely plausible and can be quietly wrong, and the only defense is somebody who knows the work well enough to catch it. On the boring task, that somebody is you.

Pick an unfamiliar problem and you lose that edge. Now you are checking two things at once: whether the tool works, and whether you understand the problem.

The boredom is the point

Spend an afternoon on something you do twice a year and you have traded an afternoon for twenty minutes. Spend it on something you do weekly and it pays for itself inside a month.

Boring tasks are boring because they are frequent. That is exactly what makes them worth automating, and exactly why we look past them.

So the question is not "what would be impressive to build?" It is:

What did I do this week that I have done a hundred times before?

Build something rough. Use it on real work. Fix what annoys you.

One more thing

When you get something working, tell somebody.

I wrote about Civil 3D from 2004 to 2016 for one reason: the fastest way for an industry to get good at a tool is for the people using it to say what they learned. Not the vendors. The users. That was true of Land Desktop, it was true of Civil 3D, and it is true now.

You do not need a blog. Tell the person at the next desk.

Be Better - Build something.