The full argument.

The same sentence, twenty years later

We are on the precipice of a new era for many engineering companies. The decisions made today, will change how we have traditionally done things. There are methods and mechanisms that need to be re-evaluated and redefined.

Like I said before, I wrote that in March 2006, about firms deciding whether to leave Land Desktop for Civil 3D. I could publish it today without changing a word. Only the tool is different.

That earlier transition taught me something I did not expect. The firms that moved early were not smarter. They were just further along the learning curve when it stopped being optional. The ones who waited did not avoid the change. They did it later, with less practice, under more pressure.

So when I say I am not neutral about this, that is where it comes from. I have watched a whole profession make this decision once already.

What actually changed for me

In under a year, the tools on this site went from nothing to a working catalog spanning seven platforms and four programming languages.

I want to be careful here, because the obvious conclusion is the wrong one. I did not become a software developer in a year. The range is not me.

The range is what becomes possible when somebody who knows the domain cold can describe precisely what he wants. I know what a curb return is. I know why striping every one of them by hand is a waste of an afternoon. I know what wrong output looks like the moment I see it. That knowledge is the scarce part, and it is the part you already have.

The objections worth taking seriously

"It makes things up." It does. It will write code that looks entirely reasonable and is wrong. This is a real property of the tool, not a rumor, and anybody who tells you otherwise is selling something. What it means is that you verify. Not once, at the end, but continuously, the way you would check a junior engineer's work before your name went near it.

"I am not a programmer." Neither was I, in the sense you mean. You do not have to be able to write the code. You have to be able to say what is wrong with it, precisely, in your own vocabulary. That is a skill you have been building for your entire career.

"It will replace us." I do not believe it will, and not out of optimism. The work that a licensed professional does, deciding, judging, taking responsibility, signing, is not the work these tools do. What they do is the forty-times work. The repathing, the striping, the inventory, the reclassifying. Handing that over is not a threat to the profession. It is the whole point of having tools at all.

Where I do not trust it

It does not know your standards. It does not know your client's standards, your jurisdiction, or the reason your firm does that one thing differently. It will guess confidently, and the guess will look like an answer.

It cannot take responsibility. The stamp is yours. That does not change, and I would not want it to.

So every tool in this catalog was tested by hand, in the actual software, on real drawings, before I trusted it with anything that mattered. Not one of them was accepted because it looked right. If that sounds like a lot of checking, it is, and it is still faster than the afternoon it replaced.

Start with the thing you do forty times

If you take one thing from this page, take this. Do not start with an ambitious idea. Start with the most boring, most repetitive thing on your desk, the task you could describe in your sleep because you have done it a hundred times.

You already know exactly what correct looks like for that task, which makes you the ideal person to catch it being wrong. Build something rough. Use it. Fix what annoys you. A rough tool in your hands will teach you more in a week than any amount of reading about this.

Then tell somebody what you learned. That part matters more than the tool. The reason I wrote 528 blog posts about Civil 3D is the same reason this site exists: the faster we share what works, the faster the whole industry gets good at it.