For one, it adds better handling of Java-style enums, for another, it supports stuff like flags, and the best part is that it's totally compatible with existing enum code.
I suspect the insight you're looking for- if it might get nasty- would be from the legal side of things. If so, I assume this is a difficult question, and the legal answer to what is appropriate differs based on things that don't necessarily have to do with software development:
The trick is to be a shitty coder. Then you never have to speak to either.
Keep your head down on the job. Lots of people do. Depending on the department you're in, the people might be majority conservative, but won't say it out loud. The pajeets also hate the troons, by the way. Your projects will probably be mind-numbingly boring unless you are both really smart and gullible enough to take resposibility for something. Also, if you are a white man in engineering, you aren't making it past "senior" unless you are literally worth hundreds of millions of dollars to the company, and even getting there will be a bit of a slog unless you learn to play internal politics well. The same goes for whities of all genders outside engineering. Your job is getting promoted, and getting promoted means playing internal politics.
Truly some of the most insufferable people on the planet
A few jobs ago, there was this autistic Chinese guy with a PhD who literally could not speak English. He didn't have a heavy accent or bad grammar, or even both of those. He straight up could barely speak English. I have no idea how he functioned in life. He would get some request from our director, who also sucked at communication, barely understand it, then disappear for a few weeks before producing some fork of Jupyter Notebook or something that didn't solve the issue at hand. He also worked from 6AM-2PM for some reason. Didn't last for more than a few months, obviously.
Aye, I've said it once and I'll say it again: 99.99% of ML guys are shit at programming. I used to work with a lot of brilliant guys in this field, but so many of them were technically helpless. Their code is barely readable and it's like their ability to process information shuts off the second the python interpreter spits out an error. To their credit, they know their way around numbers better than most people in the world.
It's difficult to mesh software engineering approaches with computer science approaches sometimes. I think both have value, but they definitely are optimized differently.
You also have to optimise your insults accordingly
If gen AI was actually good this would be pretty much the perfect use case for it.
That's because you don't know better. Mathematicians mock ML people for being dumb at math and going "hurr durr gradient descent" too. They seem to know just barely enough math to get through ML interviews, then it's back to "StAcK mOrE lAyErS" and "MoAr DaTa" like a petulant child.
It's probably just sample selection bias. Some of these guys I'm referring to are genuinely brilliant mathematicians and leading experts in their respective fields, but they can't write a python PoC to save their lives. As you might expect, this issue was most prevalent among my fellow theorist colleagues.
Haven't written any file systems but I'm pretty sure you're just going to have to suck it up and use FUSE. Kernel modules are probably going to end up being both more difficult to work with and more painful to use as a user.
A little bit related to my console hacking work. FUSE is good.
