Working on automating a thing before you’re already doing the thing regularly is almost always a bad idea (though it can be fun).

And it’s always a bad idea if you’re working with a team.

I was reminded of this the other day when I noticed this guy had a screenshot of his blog post for the post’s social image, so you could read the beginning of his post when he shared it on Twitter:

Brilliant! He must be using some kind of plug-in that runs a headless Chromium browser that loads and takes a screenshot and—I asked him, excitedly…

His response:

Those of us who like to play with automation (“Leave matters of the robots… to the robots…) are always jumping the gun with automation, over-engineering something before we’ve even fully understood the problem it’s solving by encountering it over and over.

We spend all this time automating, perhaps with justifications to ourselves like “this thing won’t be worth doing at all if it doesn’t happen automatically.” This is usually just Yak shaving, often to avoid doing the thing in the first place.

Sometimes we’re prematurely automating just for fun, because we like playing with computers and robots. That’s fine, but if it’s something we’ve put between ourselves and doing the thing we care about, it’s probably just a common Yak.

This becomes particularly unhelpful when we’re working with a team—when others need to use the products of our automation. Most of them a) won’t get it/care, and b) when it inevitably doesn’t do what they expect or doesn’t cover an edge case, they’ll just do it manually anyway.

Now you’re not just the premature automation person, you’re the angry premature automation person.

This happens all the time. So many software applications exist to accomplish a task that would have been done better manually with a spreadsheet, or by email, or on paper…

And all this leaves aside that most stuff isn’t actually worth automating… Could a robot do it? Sure. But is the time and context switch that it will take to automate worth it, particularly since the automation is inherently inflexible?

Probably not: