This makes me so mad, Google is not the first i've seen doing this (AWS, Upwork, Twilio) - and the cheek to say precisely nothing about why. If you're a buyer of cloud services right, i bet old-school on-prem is looking as tempting as an actress to a teenager rn.

Every regulated firm running on GCP is going to spend Monday explaining to their board how their resilience plan accounts for a hyperscaler that operates this opaquely. The compliance paperwork is the easy bit - the honest answer is we trusted that a hyperscaler would behave like a utility and they didn't.

So yes - they owe a statement. The whole point of paying hyperscaler prices is the assumption you won't wake up suspended with no explanation.

Had a similar one. They switched off Lambda and SNS because of a potential credential leak — none had actually leaked — and I was without service for 48 hours. Same flavour as the post: the provider's heuristic was probably right to fire, but you only find out which of your things were load-bearing once they're gone.

Opus 1M context window and lighting fast response time is hard to compete with, even if you run a local A100 the local models are just not as good as tool calling, long running tasks and non-hallucinations

It was hard for an Apple ][ to compete with an IBM mainframe at enterprise data processing, but the power of personal ownership & commodity economics was disruptive enough that 30 years later 99%+ of enterprise data processing was taking place on descendants of the original personal computers.