Uh what the heck is that supposed to have helped you figure out what to do?
Its only a short visuell example of the structure u can get with Flowchart ā¦
it doenst matter what problem here got solved or which component was searched.
U have 100 Textfiles each with X,Y,Z Coordiantes ā import ā Construct Point ā Construct curves ā get length or somethink else ā¦
Iām not sure if this flowchart approach is that useful in practice. The given example is very simple and still its already incomplete. It ignores data management and parameters (e.g.split by which delimiter, localization, dealing with large data-sets, curve parameters etc). It also assumes many things which might not yield what you really want in practice. Again itās important to understand the basics and therefore you cannot shortcut this by relying on AI. In my opinion, the best use-case in this context is to gap your lack of understanding, by asking question like: what does component x?Can you give me an example? How can I extract points from text? etc.
And last but not least, I remember that getting fluent in Grasshopper was not difficult. It became a problem when solving real world problems under time pressure.
2 Likes
true
I have to disagree with you here. Iāve been using ChatGPT successfully to write python scripts to automate a range of tasks and Iāve found it to be very good. I canāt code in Python (I have some very basic JAVA) but I know Rhino well. In the space of a couple of days Iāve (it has?) prepared a number of scripts for automating a number of tasks, ranging from making Legends with swatches from selected objects, curve lengths or hatch areas, or calculating specific values for user selected objects and then using these values to create and name new layers. I work with the GPT interface to build up the script from the initial simple task and testing it as I go and then introduce more complexity.
I have produced scripts that meet my needs in a fraction of the time it would have taken me to learn the scripting myself - and along the way, Iāve learnt more about the structure of Python and my prompts to CGPT have gotten better, with faster outcomes too.
1 Like
I mean professional developers not trying to scam people struggle to make these things do more than the most absolute bare-bones generic tasks that a zero-experience computer science grad could do that absolutely exists a thousand times over in the training sources, but whateverā¦
Doesnāt this strike you as kind of bullshit though? Thereās no way that CGPT has a useful knowledge of the RhinoCommon API or the purpose of the objects that itās manipulating. So what exactly is the workflow here, and how is it possibly better than just reading the documentation and doing it yourself?
Edit: I just tried it. Asked it to make a script that selected a curve and then created a layer with the length of that curve as a name. Shockingly it was able to stick to VBScript syntax. Less shockingly, the script didnāt work and contained calls to functions that donāt exist.
Hi, @robert_monyer.
Maybe Iām late to this particular thread, but after extensively using AI the past 3 years for training, teaching, and in professional projects, I donāt think you want to use ChatGPT for translating between German and English Rhino commands. Iād use Google Translate to get your ideas into English, then interface with ChatGPT in English only, especially for Rhino-related commands and scripting/coding. That said, Iād highly recommend you understand some Rhino basics using our Rhino Level 1 training learning path (Rhino Training Guides (English) [McNeel Wiki]), along with also having some fun with creating a customized RhinoGPT that points to sources of information you curate as most important. As you train your own GPT, youāll be able to control formatting, possibly language (eventually), and results will be more predictable & robust.
If you can give us a specific use-case / scenario, that might be an interesting way to help us understand how to push you forwardāenjoyably.
Happy to keep the conversation going.
1 Like