Sadly I have to admit I read it and didn't know it was satire. The way they described using LLMs to to achieve research results sounded plausible enough to me. As an attorney, I also fear a race to the bottom where legal research becomes simply editing the work of AI. Also, the idea of being subject to a test with little relationship to professional practice reminds me of the bar exam; which requires intimate knowledge of almost all areas of the law despite the fact that practicing attorneys generally are only familiar with their specializations (eg, it would be unwise to retain a civil rights attorney to represent you for even a traffic violation); and requires all the knowledge to be memorized, even though it would be unthinkable (and most likely malpractice) for any attorney to practice relying solely on memory.

I also suspect that there are attorneys out there who are starting to use LLMs to generate legal work with a method similar to what the author described (generate prompt, revise minimally). Lawyers are taught to write a certain way which LLMs mirror flawlessly; the surest way to identify LLM generated legal writing is if it happens to hallucinate citations. I used to regularly find hallucinated citations, and occasionally strange reasoning, but both have largely vanished as the LLMS improve.