Ephera

If they couldn't find/trust additional admins, I don't see how they could've handed it off entirely...

The meme is a reference to a popular German YouTube channel, called "Held der Steine" (basically translates as "Hero of Bricks").

On the channel, a guy shows off building block sets which you can then also buy in his accompanying shop. Up until a few years ago, it was almost exclusively Lego sets. Then Lego sent him a mail that his channel logo, which contained a Lego-like brick shape, violated the Lego trademark.
At a later point, they also demanded videos of him to be taken down where he had colloquially referred to building block sets from other manufacturers as "lego", as we often do in German.

And yeah, this did not go down well, so now the guy mostly shows off building block sets from other manufacturers and frequently highlights how insanely expensive Lego is in comparison.

Non-gaming anecdote: Colleagues wanted to build a Rust application for different platforms. (Save for scripting languages, Rust has some of the nicest tooling around that.)

Building for Windows:
cross build --release --target=x86_64-pc-windows-gnu

Building for Linux:
cross build --release --target=x86_64-unknown-linux-gnu

Building for macOS:
Uh, you need some signing key or something like that? I believe, they had also concluded that you'd need to use a Mac to do the build, rather than being able to cross-compile from wherever.
In the end, they decided not to support macOS...

Well, the unfortunate part about Metal is that it's incompatible with the rest of the world, too. They could've integrated Vulkan and chose to do something slightly different instead, because that's the way the Apple crumbles, I guess.

There is MoltenVK, which is a compatibility layer to be able to run Vulkan games on macOS. Maybe they'll integrate that. But well, it wouldn't be on-brand, and it certainly still doesn't make it easier for gamedevs looking to support macOS.

The thing with DIAAS is that it's hardly relevant and I feel like it's played up by misinformation from the meat industry.

Let's say you only eat red lentils for your proteins, which according to that DIAAS calculator has only 59% of the SAA compared to the amino acid distribution that your body needs. Then the solution is simply to eat twice as many red lentils to get to 118% SAA. Your body needs a certain amount of each amino acid, but if you give it more, it can work with that perfectly fine.

DIAAS is only relevant, if you eat close to the minimum amount of protein that your body needs in general, which is hard to do. For example, in the US, the Recommended Dietary Allowance is at 0.8 grams protein per kilogram of body weight. Which is a one-size-fits-all number they chose to cover the necessary intake even for athletic and pregnant folks. The majority of people need less protein than that. And yet, according to this site the average American eats 1.6 times as much.

I always find this question bewildering. There are so many vegan protein sources, some of which are really close to meat in taste/texture or which taste great on their own.
With insects, we'd need to invest lots of work to come up with recipes, to build farms and we'd ultimately need to grow plants to feed them, too, meaning they would generally be more expensive.

Is it just the assumption that because it's a dead animal, that this makes it 'better' somehow? Otherwise, I don't understand why we're even considering insects.

I don't have much experience with IPv6 yet either, but as I understand, the primary benefit is that you can get rid of a lot of the crappiness of IPv4, which you might just deem 'normal' at this point, like NAT and DHCP. It does happen quite a bit, for example, that we'd like a unique identifier for a host, but with IPv4, you need to store a separate UUID to accomplish that.

You could say, they are

...sheepish. *Insert CSI: Miami intro*

A few years ago, I would have fully agreed with you, but having tried my hand at (hobbyist) gamedev broke those rose-tinted glasses for me. It's just extremely hard to curate gameplay mechanics.

The only real way to know whether a mechanic works in your game, whether it's fun, is to implement it. That means you'll be programming for weeks and at the end of it, you might end up deciding that it actually isn't fun, so you get to rip it back out.
This is also a somewhat linear process. If you think of another mechanic at a later point, you're not going to re-evaluate all previous mechanics to see whether a different combination would've been more fun. Instead, you just decide whether this new mechanic adds fun to your mechanic-soup or distracts from it.

Point is, even as a hobbyist and idealist, with theoretically infinite time, I quickly learned to swallow my pride and appreciate when something just adds fun, whether it perfectly fits in or not. You're just not going to create the perfect game. And a game that's a sum of inconsistent, fun parts is still more fun than a coherent game that doesn't exist.

Of course, this does not mean, you should include mechanics even though they're overused. That seems to rather be a result from long development cycles, where games decide to include the mechanic when it's not yet overused, e.g. when a popular game featured that mechanic, but once the game comes out, then a whole bunch of other games have come out before, which had also decided to include that same mechanic.

I think, part of it is also that it's a rather isolated feature which is fun on its own. You don't need multiple systems working together to make parrying fun. Instead, you just react in the right moment and there's your endorphins. Pretty much the hardest part about implementing it, is to make enemy attacks readable, which you likely need for dodge rolls, too. And then especially for AAA titles, which can't afford to experiment much, such an isolated feature is just a no-brainer to include.

OCaml has a camel with two bumps. So, that's gotta be the Perl dromedary camel...

Any normal UI framework will unload UI elements when they're not shown. Yes, that means a CPU/memory spike is normal. But on a modern PC, that spike should be much lower than even 1%, which is why you can't typically see it.