Is that really a screw up? Companies pay for certain tolerances on their products. I think it's more likely someone acknowledged your anecdote was possible and decided the cost/benefit analysis worked out to that being most profitable. Most things don't have to fit like Lego (while some things have to fit even better than Lego).
Yes, it is a screw up. If something is in tolerance, it goes into the hole it was designed to go into, that's what a tolerance is. What I am describing is something that is out of tolerance for which corrective action needs to be taken to get it into tolerance. This is not an uncommon occurrence, it is the norm. There are times when you deliberately put in room for fine adjustment, but you don't do fine adjustment by bashing on things with a hammer (or if you do then you've made some other mistake).
In the case of the thing that needs to be retightened periodically, I'm referring to a case where a cost/benefit analysis would show that regular maintenance is more costly in the long run. Imagine a $1 Million dollar machine that is supposed to cost $50k per year to run over its lifetime of 10 years, but due to poor design winds up costing $100k per year to run. No one is going to pay an extra $1 Million to build a new machine to save $500k, but that's still $500k lost due to a screw up.
There isn't some secret army of engineers performing optimizations behind my back. Engineers design things to work, but we are fallible human beings who suffer confusion, exhaustion, distraction and at times just pure incompetence. Depending on the complexity of a system, we might have fellow engineers review our plans to look for issues we missed, but those reviewers too are fallible so things nevertheless get through.