[Request for Feedback] Gen-AI Assistant from Dynamo Team

Numbers 1 and 2 both already exist in the form of node auto-complete. Double click on an input or output to see the likely next node.

More here: What is Predictive Scripting with Node Autocomplete? - Dynamo BIM

ok,I forget it,cause I seldom use it and find it annoying when it come out.

Have you tried the ML Autocomplete version? Or just the original Node Type Match version? The former is getting pretty powerful now :slight_smile:

Just tried both mode.So it collect data of usage of nodes that customer uses,and give you advice which node most people would choose next.Yes it’s helpful.It’s a source for choice when you don’t know what to do next.

I am so excited for this feature. I would be great to expand onto the autocomplete but in the form of prompts:

  • Welcome to Dynamo! What can we help you automate today?:
    • Select and Filter Elements
    • Get data from Excel
    • Obtain linked model information
    • Get grouped elements in the active model
    • Work with Views and/or View Templates
  • It looks like you would like to update elements in the active model, would you like to update the parameter values of the elements output in the node [node name]?

Channeling Clippy there haha!

Need to find a similarly catchy name for it in the context of Dynamo…

Blippi is taken by a Childrens show, Clippy had it’s heyday already, so we are left with, umm, Trippy? :roll_eyes: :laughing:

If I understand correctly, you would like the Gen-AI assistant to

  • know the context on Dynamo integration whether it is DynamoRevit or Dynamo Civil3D or others, so the assistant would be able to suggest workflow only applicable in that environment
  • based on the context user is working on, provide option to auto complete the some mundane workflow for the user so you do not have to start from stretch

I had suggested Dynie.
OIG3

Different agents for different tasks.

Explain graph

Organise graph

Update Python nodes

Convert Python node to zero touch.

Add multi thread to Python Node.

Dippy surely :slight_smile:

image

For me I just see recommendation for next node and understanding group of task is best deal

Hi,

Questions that came to my mind while reading:

  • Can it automatically integrate ‘different libraries (Shapely, Numpy, Pandas…)’ when writing Python code or how to deal with it. Or is it only Ironpython based?

  • Will there be a source user for LLM or is your dataset already available?