AI

Really simple tests should make for fast accurate AI generated code

Once the testing framework is in place then the next obvious step is to be able to automate the loop required to ask AI to write a function and run it through the testing framework.

Then give the AI clear information on what unit tests have failed and hopefully have it correct the code.

This is where I hope to get soon since it's key to other ideas like having the AI separate concerns by taking a function and breaking it into smaller ones but keeping the same functionality.

Then the question comes - can Flow detect functions which are equivalent in the framework?