> ... confused about why I would talk to the human that prompted the AI, rather than the AI that did the work.

One reason is because you can't talk to the AI that did the work, not unless the human who prompted it is able to hand you a file with the complete session and/or context data. (And I'm not at all certain whether Anthropic would allow users to obtain that kind of thing, since it would probably be extremely useful to people trying to distill their models. It's possible they might allow user A to say "Hey, allow user B" (or anyone with the link) "to access my session with GUID ba5eba11-1234-5678-abcd-0decafc0ffee", but I doubt they'd ever allow context data to be downloaded and passed around).

Without the context, you're talking to a different instance of the AI that doesn't have the memory of the instance that ported the code, so the answers you get won't be nearly as useful.

This is actually what I did to understand what happened during the port. I did analysis weeks after the game was ready.

While doing the analysis, I came to know about this 108 byte delta. I didn't quite understand the explanation Claude gave me. I added more details about this in the post above. It could be the difference in the assembler used, the environment used (Amiga 500 vs M5 Macbook), or how the binary was saved from memory 33 years ago