Annoying for dynamic DNS and DNS ACME challenges where you want a server to manage its own records and nothing else

For instance, Walmart doesn't want to pay their largest competitor.

And no, a 174th team doesn't solve it. Communication and collaboration across teams is key

This is a huge knowledge drain. You're constantly spending time getting new engineers up to speed and it takes years to relearn all the nuances the last person knew.

You're in a constant cycle of re-learning the hard way instead of proactively applying experience.

For instance, our Patroni clusters were much more performant and stable than Azure Single Server Postgres (what a terrible product...).

These super duper magic global services seem to be the cause of most outages since the blast radius is so huge.

On the other hand, the proposition of a magic, infinitely scaleable service endpoint is nice from a developer perspective.

Don't remember the details of #2, just that they were a "Microsoft partner" of some sort which was beneficial to integrating with the Microsoft apps the product depended on and appearing as an app in the marketplace. The company built software that ingested IM/chat data from corporations (Teams and I think something older)

In the U.S., some of it was tariffs, though.

e.x. ChatGPT-5.4 Thinking, Claude Opus 4.6

Older/smaller models were far worse

Src https://www.protondb.com/app/8500