Nope, not my job at all.
I was just interested in the tech behind the AI world generation, so been diving into it a bit.
I’m not saying it’s always perfect, and/or that it does or does not require any fixing/improvements. Of course there are possible improvements.
I’m just explaining (to the best of my understanding) how the tech works, and how that might explain some artefacts we might see. If you want to call that defensive, that’s your prerogative.
About quality control; They have an entire globe. You don’t seriously expect them to go over every square mile of it manually do you? If you find issues with the world, make a report to Zendesk so that they can improve it in the future.