SAO is the worst fucking offender. The first half of the first season is top tier, but then the fairy shit is almost unwatchable. And then they keep going back to the awful fucking fairy world in every sequel. It almost felt like it was done out of spite.

Ah it’s just lemmy.ml with the soft ban. The main “dev” instance.

I will say that I would be a lot more willing to donate if the soft ban on Ukraine news was lifted.

Hexbears are just trolls. They generally aren’t coming into threads to have real discussions (and when they make attempts at it they often seem confused about their own stated ideology) - the unstated goal is to be disruptive to deny “shit libs” a platform.

I mean, there is a hard coded word filter which doesn’t allow you to use the scientifically correct term for a female dog. Which is pretty wild.

Hexbear is far more obnoxious than lemmygrad imo. At least lemmygrad doesn’t turn every thread into a middle school “#random” message board by filling it with cringe jpeg emojis.

It’s ok, you can say hexbear. They know they are annoying and everyone hates them.

deleted by creator

The first one was botted as well near the end. I know because I wrote a tool to do it in about an hour using my private bot farm I had been building for years. I never gave it to anyone else because it was sloppy and integrated with the bot creation tools I had, and that would have revealed all of the tricks I used to create and protect the bot horde. But I used it myself, and I assume I was far from the only one with the same idea.

The interesting thing is that reddit basically condoned it. They are usually pretty decent at detecting bots if you don’t take measures to make them look human, but with place (at least the first time) they seemed to intentionally have that functionality disabled. I kind of assumed that they doing it specifically as a bit detection scheme, but they never cleaned house afterwards like I expected.

Laughs in

struct LooooongCounter {
    long highPart;
    long lowPart;
};

Op, we have decided to go with a different candidate

Netcode. I make my own layer 2 headers sometimes.

laughs in reinterpret_cast

Is it even a programming language if you can’t casually introduce type overflows?

deleted by creator

Nah, it’s pushing inference to prod. Any idiot can make an ipynb to train a model, but packaging everything into an app ecosystem is where you actually need a lot of non-ML software engineering know-how.

Savage

Yeah I never understood to pearl clutching about downvotes. It just reeks of “everyone who drives faster than me is a lunatic, everyone who drives slower than me is a granny” mentality.

Fuck you, I disagree.

Is that better?

Undercook the fish? Bot, believe it or not.

Because it’s easy. Someone is just testing some basic tools, to which they can add countermeasures later.