Maybe for you it's "playing". For some people (and this used to be me) it was a question of how I built big dangerous machines and put them in a factory where stupid people on piecework would come back from the pub at lunchtime with some great idea about how to work faster by jamming the guards open.
My problem wasn't the smart people - they could by and large look after themselves. I once dangled a co-worker by his ankles _inside_ a double eccentric press whilst I barred it over on the slow motor, so that he could count the teeth on the dog clutch. This was an act guaranteed to give any safety inspector conniptions, but it was _relatively_ safe because we both had a pretty good idea of what we were doing and were concentrating on doing it.
Sometimes I got to work on machines in toolrooms, where they'd be used by the most skilled machinists you could hope to meet. Most of the time though, they were for low-skilled grunts to earn minimum wage with at McJobs and yet I _still_ had a legal responsibility to keep these people safe from any deviousness they could invent (and they sure were inventive, when it came to defeating guards).
I got out of the control gear business when a software bug in my code very nearly took an operator's arm off. I looked very carefully at how this problem arose, and my conclusion was that it was a systems error more than a coding problem. Given the stack of sub-contracting going on, and the poor communication and rushed timescales between the people building the ironwork and my software team, I couldn't guarantee anything I could really stand behind as a competently safe system. It was my responsibility to sign these things off as "built to best industry practices", even if I couldn't guarantee them to be entirely safe, and under those cost-squeezing measures there just wasn't a way I could do this.