I don’t know what’s more bizarre:
- So many days named after random and really dumb things like, well, odometers. Or…
- So many software engineers publicly insisting that their job is going to remain relevant into the future, because they can do something LLMs can’t.
Yes, I say that to be inflammatory. It’s a conversation starter. (Perhaps a fire starter too… let’s find out. π
I don’t want people to suffer, lose jobs, discover that what they’ve spent 10, 20, 30 years of their lives obsessing over is suddenly not particularly useful. That’s painful. And pain sucks.
But just because I don’t want something to happen (i.e., because something really sucks) doesn’t mean it’s not going to happen.
Personally, I believe that platitudes and logical fallacies are more damaging to society than recognition and acceptance. I’m purposefully focusing on software devs here, but similar points could be made about GIS professionals and cartographers as well. I address these in my forthcoming book.
Nonetheless, even if my “devs are doomed” perspective proves wrong, which it very well may, it needs to be posted. Particularly because this past week I’ve seen numerous articles, posts, and diatribes online by GIS devs and mid-level GIS bloggers about how traditional dev and GIS human skillsets will be needed and more important in the future. I even had to read a completely daft section in a blog about how our brains will atrophy if we don’t keep coding. (A comment that appears akin to what a mathematician might have blogged about in the 1960s when the first scientific calculators were being produced β your brain will atrophy if you use that thing, keep doing everything longhand!)
I think such arguments are not only wrong; they are dangerous. If you’re a GIS or code-first cartographer who wants to stay gainfully employed and feed your family in the near future, you need to start hitting the proverbial books (i.e., online lessons and ChatGPT, or maybe a textbook? ;-). What we’ve been doing our entire careers was important and valuable. It’s much less so now. It sucks, but it is the situation. A new technology has come along. The workforce no longer needs human beasts of burden to do coding or create a map manually via GUI or CLI.
Obviously, for a couple more years, until LLM workflows are established, human devs and GUI/CLI-based cartographers will still be useful in the workforce. But coding and engineering maps… as a skillset? Skillsets only exist because they are useful. Skillsets go extinct based on market demand β not how much we cherish them.
Which is why I shudder a bit when cartographers chant to themselves at conferences that “Cartography matters”. Admittedly, it’s fun, but history shows that chants and parades to fervent believers spell the end of an institution, not a high point. (Just ask Viktor Orban.)
With LLMs, coding knowledge has no practical goodness of fit for human evolutionary purposes. It’s akin to knowing cursive or, perhaps more aptly, how to take notes in shorthand.
You can learn how to do all of these things β even today. But in today’s job market you’ll never get paid well to do it.
It’s absolutely startling how quickly LLMs have turned the market for coders and web map creators on its head. It terrifies me in fact. I began playing with prompt cartography for fun and quickly saw the writing on the wall. In the last two years, it has advanced more than I could have ever imagined. But it’s also allowed me to create maps in my spare time that blow my mind. I can’t imagine what a full-time practicing cartographer could actually do using prompt cartography full-time today.
Anyway, this is what ran through my head today, as I waited roughly four or five minutes to create the following map.
Two prompts:
- one to create the dataset.
- one to create the map.
In fact, writing this blog took longer.
Ah, I digress again! To conclude…
Is today’s map a masterpiece? Absolutely not. It’s a four-minute, one-off puff piece, in journalistic speak.
Is it AI slop? Nope, It’s not that either. It’s better than a lot of graduate student project work I’ve seen, for sure. (And completed in four minutes, not a semester.)
And really… is that not good enough to prove that prompt cartography is going to change everything we ever knew about how to make a map with computers?
Because coming full circle, I’m creating a map about… odometers?!
What the #@$%?!
Who would be able to spend the time to make such a map in any other world? Proof my brain hasn’t atrophied from LLMs coding for me; I’ve been liberated to be more creative than ever before. Personally, I find it absolutely amazing!
Thanks for reading today’s mini-rant. (There was an essay-length version, but I took it out so devs and cartographers don’t show up at my door with pitchforks.)
Oh, and of course, one major caveat I need to mention: I’m wrong β a lot. π
But I just want to counter the current narrative I’m hearing that seems, from my perspective at least, to be a 100% logical fallacy. I feel like I’m watching lemmings follow B-level bloggers over a cliff, and I can’t stomach not saying something.
GIS Dataset Doctor GPT
Hi. I’d like to create a map for National Odometer Day. I’m trying to come up with some tongue in cheek maps that aren’t overly complex but still informative.


View map here.
View Map Here
View Map and Play Some Miniature Golf 1980s Style Here
View map and play game here
View the map
View and Play the Map
View Map Here