Why isn’t there more?
Why have they been the same ones for years now?
Why no square, diamond or hex patterns?
Let’s tessellate.
Or a benchy shape for a benchy like the Cura infill.
1 Like
Where is the math AI when you need it?
1 Like
I thought the future would be cooler
Ask chatgpt to make a profile to do it. The future is now, buddy. AI to make AI.
ChatGPT sucks a at code generation.
Ask any software engineer.
Actually, it does a much better job than I do these day, so “sucks” is all relative. 
In all fairness, my best coding days are long behind me. What makes matters worse is that back in my day, when we had to compile code by hand, we were forced to memorize syntax. Modern tools like Microsoft Visual Studio’s IntelliSense practically write the code for you—it’s that smart. The downside is that it also dulls your coding prowess since your brain doesn’t have to do as much heavy lifting. Since I no longer code regularly, I’ve found that when I write something, a month later I forget how I arrived at my coding decisions, even if I heavily comment the code with generous “Note to self” statements.
I would never trust ChatGPT to write anything mission-critical. However, it writes code that is, for all intents and purposes, chabuduo (差不多). Although I am no longer a developer, I have never lost my systems analyst expertise. I can still write a code spec and hand it to a tech lead for implementation.
Now, enter ChatGPT. It’s dumb, it’s stupid, and worst of all, it does exactly what you tell it to do—whether it’s a bad idea or not. Not to be too unkind, but it’s a lot like outsourcing code development to a team in India. They are great coders and work for cheap, but they often have no understanding of the objective and struggle with Western jargon. Unless you spell out everything and go through an iterative process, the first several passes of code—and the rewrites—are junk.
That’s exactly how I use ChatGPT to write PowerShell scripts. I spell out the algorithm and, over many hours and countless tweaks, I get output that’s adequate for my task. A simple example is my frustration with Bambu Studio JSON files not syncing parameters across printer definitions. I told GPT what to look for and had it output that data into a CSV. Then I had it take the CSV, convert it back into a JSON file, and tested it in Orca. Once that was debugged, I expanded the task into a directory-wide scan, making global changes to all printer definitions.
Now, with Orca’s Dependencies Tab, that’s less of a need—but for a while, it kept me sane. Truth be told, I probably wouldn’t have taken on the task otherwise.
1 Like