Granted lots of mediocre engineers also use the “freeze the results” method for meaningless test coverage,

I’d be interested what you mean by this? Isn’t all unit tests just freezing the result? A method is an algorithm for certain inputs you expect certain outputs, you unit tests these inputs and matching outputs, and add coverage for edge cases because it’s cheap to do with unit tests and these “freeze the results” or rather lock them in so you know that piece of code always works as expected or it’s “frozen/locked in”

You could have it write unit tests as black box tests, where you only give it access to the function signature. Though even then, it still needs to understand what the test results should be, which will vary from case to case.

LLMs will generate the tests based on how the code works and not how it is supposed to work.

You can tell it to generate based on how it’s supposed to work you know

I think machine learning has a vast potential in this area, specifically things like running iterative tests in a laboratory, or parsing very large data sets. But a fuckin LLM is not the solution. It makes a nice translation layer, so I don’t need to speak and understand bleep bloop and can tell it what I want in plain language. But after that LLM seems useless to me outside of fancy search uses. It’s should be the initial processing layer to figure out what type of actual AI (ML) to utilize to accomplish the task. I just want an automator that I can direct in plain language, why is that not what’s happening? I know that I don’t know enough to have an opinion but I do anyway!