Or they'll get hired, get taught really basic shit (chmod/chown) on the job, get a cisco/AWS certificate then leave 6 month later to join a better paying job.
In my first job we loaned Hadoop clusters that we either managed ourselves or let the client manage the IT. One of our client decided to use a contractor instead (I might miss details, it was almost 10 year ago) .
The day after configuring the cluster and giving him the metaphorical keys, we saw low-level alert on one computer in that cluster, some monitoring daemon couldn't be reached but nothing really concerning (could still be pinged, no issues with the virtualisation). We still tried to look at it, but could't ssh on the computer (which explains why our daemons couldn't be reached). We contacted the client, said basically: 'we can remove the monitoring and our access keys if you want, but please tell us before you do this. Do you want it done on the other computers?'.
He came back the day after that saying basically 'what?' then talking about a shard that couldn't be reached (he was nice about it).
Turns out, 'chmod 777 /' is the dumbest way of breaking your workstation I've ever seen. We all hear about 'rm -rf /' but let's be real, no one has ever done it, not outside of school at least. Chmod 777 / because you couldn't manage to understand a Java stacktrace has 100% been done in a professional environment: I had to fix it.
Luckily you can copy the permissions of filesystem A onto filesystem B (can't remember how, but it was easy) so the fix didn't involve any reinstallation and Hadoop wizardry.
