LLMs turn your job into mostly code review, a task everyone famously loves to do and is good at

@jcoglan sorry for the literal screenshots, but check this out.

Wrote a library, then got asked to "put the problem into ChatGPT and see how close it gets to yours".

Well, it took me quite a long time to get the over-confident, defensive PoS to admit it f'ed up OCR* on a table.

(*kinda. I still don't know what exactly went wrong along the way, but since it barfs out and executes python code to get there, it might as well have screwed up some indices under the hood to mix up the table cells.)

@jcoglan ah well I think I see it now.. column index off by one

Having LLMs do tax reports and medical decisions will be SWEEET

@jcoglan @MisterHW isnt there legal ai allready one from cloude. I love where this is going. Non deterministic output to deterministic systems thats so sweet. I love how we'll start debugging it. We should have psychaietrists for ai's