First Draft of Multivariate OLS and AI Security

Today, I completed a first draft of a tool for performing Multivariate Ordinary Least Squares regressions. It's similar to my previous tool except that this one does regressions against multiple input variables. I'm quite pleased with what I've done, it's still a bit rough around the edges but I'm quite proud that I managed to get to grips with the matrix algebra necessary to do it. There's something quite satisfying about working your way through the algorithms as you build them and think them through.

I've also been reading this very interesting article on how LLM based coding assistants represent a potential security risk.

It's often discussed that LLM's could produce potentially compromised code, but this goes further and identifies the coding agents as a potential security risk. Coding assistants with their ability to run commands with a developer's elevated privileges do seem like a potential security risk.