Unless something happens to change my mind that’s my last day of work before retirement.
This is my thinking, too. I see systems being able to automatically create a history tree from other formats. Most likely with some iteration, they will give a usable history tree 85% of the time. I don’t think history tree usability is black magic, it can probably be reduced to a set of rules, and testing algorithms. Say this is developed in-house… they can train systems by looking at history trees of tons of files.
Maybe AI can magically purge all the “Trash” file space that builds up in my drawing files.
autocreated tree are not necessarily the best in terms of performance or model quality.
working with surface and molds in NX for years and sometimes you have to play with feature tolerance (a concept SW keeps away from their users, with the exception of surface “sewing” into solids probably) : a “bad” mathematical model, but fine for most NCs.
car bumpers and front grills are a collection of garbage.,10yrs ago we were getting 300+ MB step files for a single part exported from a dozen of cad and cams before reaching our company…
as in recycled water bottles, CRT housings and old toys?
bnemec
3D modelling wise they are a patchwork of bad surfaces and bad conversions plus all the engineering issues as a bonus.
for a seat frame we had to split the mould design on two workstations at the time. and it was not even the biggest mould we made, but the catia step data was so big (and bad) we had no choice.
“Over just a few months, ChatGPT went from correctly answering a simple math problem 98% of the time to just 2%, study finds”
Even AI got dumb talking to human.
Entropy eventually wins
Increases… so, idiocracy? 