No way to tell as it depends on all kinds of factors, like workload and IO requirements. For example you could spin up a very cheap spinning rust Ceph cluster for archival use cases or you could have a very fast NVMe/PMEM cluster for compute workloads.
I am managing all kinds of installations.
And before somebody ask no, not piracy related lmao.
(Replying with my lemmy account as my other account apparently isn’t federating with programming.dev)
There is no way to prove it didn’t just scrape 10 other summaries and reworded them slightly. And given the nature of such language models and limited context length it’s actually more likely, than it understanding and summarizing an entire book.