So I did this.

ls -al ~/.vscode/extensions/ms-python.vscode-pylance-2020.6.1/server/
total 34520
drwxr-xr-x   8 kghose  staff       256 Jul  1 07:59 .
drwxr-xr-x  12 kghose  staff       384 Jul  1 07:58 ..
drwxr-xr-x   7 kghose  staff       224 Jul  1 07:58 bundled-stubs
-rw-r--r--   1 kghose  staff  15714756 Jul  1 07:59 libonnxruntime.1.3.1.dylib
drwxr-xr-x   3 kghose  staff        96 Jul  1 07:58 native
-rw-r--r--   1 kghose  staff    164804 Jul  1 07:59 onnxruntime_binding.node
-rw-r--r--   1 kghose  staff   1757870 Jul  1 07:58 server.bundle.js
drwxr-xr-x   6 kghose  staff       192 Jul  1 07:58 typeshed-fallback

That onnxruntime lead me to this: https://github.com/microsoft/onnxruntime

ONNX Runtime is a cross-platform inferencing and training accelerator compatible with many popular ML/DNN frameworks, including PyTorch, TensorFlow/Keras, scikit-learn, and more. aka.ms/onnxruntime

Many users can benefit from ONNX Runtime, including those looking to:

Improve inference performance for a wide variety of ML models
Reduce time and cost of training large models
Train in Python but deploy into a C#/C++/Java app
Run on different hardware and operating systems
Support models created in several different frameworks

Not sure what to think of this all. It might be a very sophisticated language server, and they do point out that it is built on top of https://github.com/microsoft/pyright.

Anyhow. I might revert to the old server.