I'm sure the BBC did exactly that for ages, iirc because somebody who got their tweet embedded by them changed their display name to "big dogs cock" or something.

That sounds like a flaw in the embedding technology — surely the display name should be whatever it was when the tweet was created?

How do think embeds actually work? There is not snapshotting of data - it's simply a URL that renders at request - an HTML iframe.

Sure, but there's no reason Twitter/whoever couldn't decide to offer a parameter like `at=<timestamp>` - which is really not that different to, say, a link to a file on github at a particular commit.

Sure. I don't think Twitter has such an "look up as of effective date" enterprise feature though. Have you been able to access that data that way either through the UI or API?

No, I don't use the API (or twitter at all, for that matter). I was just reacting to your statement:

> There is not snapshotting of data - it's simply a URL that renders at request - an HTML iframe.

"Simply a URL", the resource to which it refers being rendered in an iframe, could easily be a snapshot - as it is on other sites. Hence to me this read like a non sequitur.

I didn't mean to suggest that it would necessarily be possible today - I doubt twitter keeps a full history of username changes, for example, and they'd probably need to introduce more restrictions around changing them if they were to do so. That can all be part of the 'flaws' in the embedding technology I referred to.

Would the copyright even belong to Twitter, as opposed to the author of the tweet?

Short works can fall into copyright in the U.S.

Heck, even taking a photo of the Mona Lisa is subject to copyright in the U.S under the theory there is creativity in the composition of the photo.