Hello, today i receive email about someone’s evaluation of using Rhino to do reverse engineering, he used Chat GPT, which i never used before.
I think it is quite accurate ( described what i processed in rhino) and it is worth to think about how to improve rhino.
#1 is incorrect info, MeshToNurbs makes one joined mesh and it will preserve sharp features, as it makes one surface for each facet - i.e. it’s identical in shape to the original mesh.
Beware of what “sounds like” good info from ChatGPT - it often sounds really good, but is inaccurate.
And when it comes to code it will frequently suggest libraries and API that do not exist.
It’s also sort of missed the point…
Quad Re-mesh is the best approach (well certainly a strong contender for that title) initially at least to get the mesh into a workable resolution before trying any of the other approaches.
Google AI search results are also in my experience more often incorrect than not…
such as when I just googled ‘how to start Rhino 8 in safe mode’
No it’s not worth anything. There is simply not enough information about anything to do with Rhino or even CAD in general floating around the Internet for this to be reliable about anything.
According to people who actually know programming and aren’t making up fake demos, the best LLMs today can fairly reliably handle…the most basic of generic make-work programming tasks around starting to build some sort of generic “web app” handed off to fresh hires out of school just to give them some experience. It’s worthless for Office automation stuff I try to work on, and you would think there’s plenty of that material available to train on.
You are right, MeshToNurbs changes each individual face to nurbs, but it will take lots of memory and I don’t use it.
Yes, I know nothing about ChatGPT.
I know nothing about ChatGPT and uncertain about its accuracy, some of its description is close to what I am recently doing reverse engineering with Rhino. For example, ToSUBD to define crease/ sharp edge/ fillet edge is difficult to perform accuracy in the resurfacing of Iron Spider project, ChatGPT result sent from a workmate gives me a bingo.
It’s only “valuable” in the same sense that the way to get engagement on Twitter/X is to say something obviously wrong. Post some AI slop, get corrected, claim there was therefore some value in the AI slop.
It’s not fun or interesting and I’m not going to knowingly help anyone who does it. Just learn to ask proper questions or hire me to help you. If you don’t know enough to figure out how to fix its mistakes yourself you shouldn’t be bothering everyone else to help you for free in your quest to shortcut everything.
Sir please don’t get me wrong, there is just something ChatGPT mentioning similar to what I used rhino and thats it, all of you disagree ChatGPT and I didn’t agree what ChatGPT‘s, I just want to know how you guys think that ChatGPT.
Do you think I asked too much questions here without solving problems by myself? I don’t think so. I solved most problems by myself, those questions I asked before are mostly something for improvements to use Rhino in better efficiency, and I think raising these kind of question is a positive skill exchange. Other people who see these post will get benefit as well. Positive.
