1) Convenient means of verifying data integrity to a repo creator without having to re-download the file
2) A public dataset should be easily provable to be identical to the source (especially where the source has a readily available MD5 / SHA256 hash associated w/ the data)
Stretch goal for point #2- during upload, allow the uploader to specify a URL linking to the MD5 / SHA256 checksum of the source data (presumably from the same domain name) that is embedded in a page, as well as a CSS selector to grab the hash from the page's HTML. Upon upload, bit.io could then issue an API call to https://archive.org to snapshot the page and link the snapshot in the repo along with the extracted checksum. It's an advanced workflow but would provide a future-proof way to verify that the original dataset came from the source of the MD5 hash.