Happy to see that you made it to Cemetech!
Someone sent me the Reddit post earlier today and I'm quite excited about it. Only the very early TI-83s had an Atmel ROM chip. The early TI-83 boards resemble a 1995 TI-82 or TI-81 more than they did any other TI-83, but those used completely different ROM chips.
The TI-81s from that era also had essentially the same board design:
That Atmel AT27C020 OTP-EPROM which is on this calculator was also found in early TI-86s, and the 512k variant of that same chip was found on the ROM modules of early TI-92s.
Given the ROM version and date code, it's pretty safe to say this is an earlier revision than the earliest known to date. Looking at the pictures you provided (thank you for those by the way), I don't see any difference between this board and revision As, or event other pre-As for that matter (not really surprising since they were all made within a couple weeks of each other). My best guess as to why there is such a drastic difference between early and late TI-83s is that they were in the process of integrating the CPU into the ASIC and for some reason, went through the trouble of designing a board to accommodate what they already had (maybe to meet some deadlines?). I've noticed that for TI, designing ASICs like the Toshiba T6C79 that eventually made it into TI-83s takes a lot of time and is very expensive, so by the time these calculators were produced, I'm certain they were already knees deep in that process.
Regarding the bug, I'm not sure why it would do that. I've heard of that method to store quotes and the sto arrow to a string, but I've never heard of it not working. Then again, I think the Equ>String and String>Equ tokens were added right around that time, so its possible that its that specific rom version (1.010) that doesn't like it.