ChatGPT and Particle Functions

Hey everyone!

Not a comprehensive post by any means, but I wanted to put this out there in the hopes it might inspire some lightbulbs to go off.

Basically, we all know particles in Flame are not the most intuitive thing in the world. They’re logical, but the learning curve can be daunting (at least for me).

After exhausting the excellent but limited resources from the community, I decided to try and train ChatGPT in how to write custom functions.

I gave it some examples of different functions and explained what they did. I also copy and pasted the entire body of the “Using a Function as a Manipulator” from the user manual and told it to read.

I then started asking for simple functions. It seems inclined to write the functions as Python Hooks, but after simply asking it write the function as plain text, it caught on.

I’ve only been doing some simple stuff thus far, and it definitely needs some more help in understanding when to use “Power” or “Magnitude,” but so far it’s pretty cool and promising!


Very excited to see some examples…


According to GPT, everything it’s learning form what I’m doing is getting added to it’s “knowledge.”

So, if you guys try asking it for particle functions, it should be much more competent than when I originally tired!

1 Like

That’s what they say. I’ve tried to feed it info about some other stuff, expecting it to learn. But, alas, like many human beings chatGPT also has preconceptions and doesn’t actually learn, at least not right away.

I can’t get it to answer in anything other than python conventions :grinning:

This is what got it out of python for me.

"the format is not meant to be in python, but plain text that it entered into flame. The function would be a single line of text. Again, here is an example of a function

pos = pos + turbulence3(pos, 1) * power"

It reverted back to python once or twice, but I kept beating it over the head that it should NEVER be Python