One of the most fortunate things I experienced in my career was a few years in the operations/hosting side of software. Working there actually helped me understand how many things are necessary to align so that a simple application works as intended to serve a number of users 24/7. Due to an increasing number of abstractions (mainly cloud/saas providers), I’m confident that this skill has been deteriorating in the IT space, and it’s the reason these dev-adjacent speakers love these doomsday scenarios.

They can imagine their code doing a million crazy things, but they hardly think about the incredible amount of things that need to exist and operate at 100% before a single line of code can be run on a VPS.

How many of these guys have had to tell a customer something silly like ”we lost connectivity to the DC because a farmer decided to do some digging and cut fibre lines connecting the DC to the internet”? If they knew that this was in the realm of possibilities, they wouldn’t be so confident about a program being able to somehow run amok and simultaneously feed itself all the resources and components it needs to run, as you mentioned.