Exploring the present of healthcare data manipulation

• Parquet

• Parquet → JSON → Tcl

• Parquet → Tcl visualizer

• By building these tools in Folk we can use spatial tools to explore our data:

  • (view the code for this demo)

  IMG_0704.mov

Machine learning detection with Folk

• Python script backgrounded that can search for keyboards (document inline, add photos)
• Ultimately working toward visualizing machine learning interfaces

Working with LLMs in Folk

• Want to explore a spatial flow for a weighted DAG for multiple answers to a query

• Should support continuous printing to save answers/points in the conversation you prefer

• Should be able to bucket them spatially into groups for easier management of concepts

• Preliminary doodle of the user experience flow of spatial LLM chatting:

  

Future Work

• Get Orin running Folk
• Convert ML compute into CUDA shaders, should see a massive leap in performance
• Sketch out data card how swapping for building out inference models
• Sketch out a visual pipeline for CV (base image → B&W → Threshold → )