I think the full description of “stateless, serverless functions” adds a bit of clarity. My read of this is that whatever code is running doesn’t maintain state

I think the full description of “stateless, serverless functions” adds a bit of clarity. My read of this is that whatever code is running doesn’t maintain state between requests, and doesn’t touch the underlying operating system. Which seems pretty standard for highly managed environments anyways. It’s been years since I’ve written backend API code that touched the underlying system, or left objects in heap between requests. The flexibility of knowing that any machine can instantly run the code for your API gives a lot of flexibility to rapidly scale up an API. Nothing is “serverless” to everyone. Especially when you run the data center. But being “serverless” and even sitting above the “language runtime” gives API developers a lot of freedom to focus on business logic.