Reverse engineering chips from die photos

[PurpleMelbourne]

Looking over some almost 4k res photos of Motorola chips, i first noticed the pictures are kind of blury, but also very regular...

So i was thinking it'd be a great avenue to train an AI on 6800 series before 68000 series as the similarities would be very beneficial for the AI to make good guesses.

If there were preceding compone ts to learn from then even better.

Personally I've never dealt with AI's, but this sounds like a perfect job for them. The 060 is no where near high enough resolution to do more than admire its orderly beauty. But an AI which has learnt all the patterns might recognise 80% of the circuit, making it easier to develop TG68 going forward.

Afterwards we might have a lego library of structures which can be manipulated a lot faster than with HDL alone.