Tutorial Request: S3-only usage

One of several issue tickets @satra requested to be opened for tracking purposes

@yarikoptic came up with this request originally - to have a notebook demonstrating how to use the S3 bucket directly to perform various operations usually delegated to the DANDI API

Would include:

- fetching blob download IDs/URLs from asset metadata files
- navigating Dandisets directly from the metadata files
- downloading a Dandiset version using something like `s5cmd` instead of the DANDI CLI

Happy to take this one myself as I've done all these operations before internally