Introduction

This tab explores how to use AI tools to help create test platforms. The work here is experimental in nature. However, at this point some facts are evident.

The code output by AI IDEs isn't optimized for humans; it's optimized for the next pass by the AI. That's why code review time goes up - the code becomes less parsable by people.

We've seen this movie before. Compilers made assembly language less important because humans stopped working at that level. It seems AI tools will do the same to today's source code. 

Accepting that the valuable artifact is shifting—from code itself to architecture, functionalty, testing, and devops is key. If AI can reliably generate and modify implementations, those higher-level assets become the things worth managing and maintaining. And, if I may say as well, we need to architect security and resiliance in from the start, because gluing it in to existing architectures hasn't been working. See the Intel iAPX 432 as a cut at a notable hardware and software architecture.

Nothing however reduces the need for talented and dedicated people of all stripes. Every abstraction shift has changed where expertise matters, not whether it matters.