> What do we actually lose going from cloud back to ground?

A lot of money related to stuff you currently don't have to worry about.

I remember how shit worked before AWS. People don't remember how costly and time consuming this stuff used to be. We had close to 50 people in our local ops team back in the day when I was working with Nokia 13 years ago. They had to deal with data center outages, expensive storage solutions failing, network links between data centers, offices, firewalls, self hosted Jira running out of memory, and a lot of other crap that I don't spend a lot of time about worrying with a cloud based setup. Just a short list of stuff that repeatedly was an issue. Nice when it worked. But nowhere near five nines of uptime.

That ops team alone cost probably a few million per year in salaries alone. I knew some people in that team. Good solid people but it always seemed like a thankless and stressful job to me. Basically constant firefighting while getting people barking at you to just get stuff working. Later a lot of that stuff moved into AWS and things became a lot easier and the need for that team largely went away. The first few teams doing that caused a bit of controversy internally until management realized that those teams were saving money. Then that quickly turned around. And it wasn't like AWS was cheap. I worked in one of those teams. That entire ops team was replaced by 2-3 clued in devops people that were able to move a lot faster. Subsequent layoff rounds in Nokia hit internal IT and ops teams hard early on in the years leading up to the demise of the phone business.

Yeah, people have such short memories for this stuff. When we ran our own servers a couple of jobs ago, we had a rota of people who'd be on call for events like failing disks. I don't want to ever do that again.

In general, I'm much happier with the current status of "it all works" or "it's ALL broken and its someone else's job to fix it as fast as possible"!

Not saying its perfect but neither was on-prem/colocation