Presumably there is an antialiasing low pass filter somewhere before JS gets to the data. I have a similar sample rate and it certainly didn't work at all for me.
The JSON types are string, number, boolean, null, object and array. So how could the suggested code possibly work? Do you want JSON.parse to do arbitrary code execution like Python's pickle?
You definitely need discipline to use C++ in embedded. There are exactly 2 features that come to mind, which makes it worth it for me: 1) replacing complex macros or duplicated code with simple templates, and 2) RAII for critical sections or other kinds of locks.
They explain why in their AI policy. It's an ethical stance. Of course they wouldn't notice if there aren't clear signs of LLM-ness, but that's not the main reason why they forbid it.
Stuff like that is a good argument for using structured logging, but even if you are just parsing text logs, surely you can make the parser be a bit more specific when retrieving the log level.
I would be surprised if there is no LLM-assisted code in there prior to this commit, this is just the first where the author chose to disclose it.