An article on Lifehacker this week explained a lot about how I initially became a computer professional. Its advice was to fly by the seat of your pants, try things without guidance or manuals, not be afraid to fail occasionally, and learn before you go to sleep.

So when I spent many nights in my late teens disassembling and reassembling obsolete IBM PC/XT clones to learn how they worked, I was unwittingly doing all of it right.

The unfortunate death of hacking legend Barnaby Jack at 35 drives this home. People like Barnaby Jack are good at what they do because they experiment, try things that shouldn’t work, and see what happens. Because every once in a while, something happens that isn’t supposed to happen, then you investigate and see why something that wasn’t supposed to work did indeed work. None of what they do is in the script.

And there is no Youtube video for that. Whether by intuition, suspicion, or an insider tip, Barnaby Jack suspected that embedded devices–the tiny computers that run everything that uses electricity these days–lack the basic security that we take for granted on personal computers these days. By exploring those suspicions, he discovered that, unfortunately, he was right. And while security on PCs is extremely important because you can lose sensitive data, lack of security on an insulin pump or an ICD means an attacker can kill someone.

So, yes, there’s value in “playing on that damn computer” when normal people are usually in bed.

I haven’t been doing as much of that lately myself, and I’ve been wondering what’s missing. I’m pretty sure that’s it.

Now, when it comes to fixing a lawn mower that’s lost its ability to self-propel, or installing an over-the-range microwave, you bet I pay Youtube a visit. My goal then isn’t to become a master at fixing lawnmowers or installing appliances, but rather to save a hundred bucks by doing something in 30 minutes that takes a pro 15.

But when it comes to becoming a professional, there’s no substitute for spare-time trying, failing, and figuring out why something failed.