I like this. Is there a way to clean up the logs this uses to hide these traces? I didn't see any such functionality as I skimmed the README.

That map image reminds me how big this game was by Game Boy standards. Impressive!

Since there's not a section for x, that must mean that x doesn't suck. /s

Closure has the shortest section and is therefore, obviously, the best.

I don't know if this "problem" can ever truly be "solved" as there's not really a way to prove there to be a difference between something truly "feeling" emotions compared to faking it convincingly. I have to wonder at what point the difference becomes moot, a sort of Chinese room for emotions. I tend to lean towards solipsism when it comes to this kind of stuff, though. Does the difference actually matter?

I could argue that even people who genuinely feel emotions have, unconsciously, learned to feel them through their interactions with other people. Who can prove to me that emotions are something fundamental to humans and not acquired through culture? In a sense, everyone may as well be faking it.

So you're positing that emotions are a cultural virus.

How did you get a picture of my waifu!?

I agree 100%. However, people like us are the minority somehow.

Maybe we're the minority. Still, wube (factorio developers) seem to be living comfortably, they don't seem to be hurting for cash. There is enough money in niche communities if your product is good

Getting there is more challenging than it used to be. Plenty of non-AAA hits seem to come with tales of development almost breaking them, or running out of Kickstarter funds. Including two of my favourites: Pillars of Eternity and Divinity Original Sin.

I'm sure there's plenty that didn't make it too.

Survivor bias. How many great indie games never got the exposure and died on steam making basically nothing. The fact that some people can do alright is not evidence of a healthy market.

People who enjoyed video games used to be a weird minority as well. If a niche is sane it can grow to become the majority way of doing things.

s/Were/Are/

There, I fixed it for you.

I'd say that Google did the numerical analyses of the format, then proved the memory bandwidth improvement and behavior in regards to large production machine learning models with their TPUs. So they derisked the numerical format. That and how simple it is to implement given you already support IEEE 754 single precision (just fewer bits for significant), and lower overhead to convert to and from floats (relative to fp16) makes this format a no brainier for Intel.

Compared to enterprises requesting hardware backdoors... err "out of band management" (Intel AMT), adding an instruction or two is relatively tame.

Since Intel had major customers, they have been influencing the CPU roadmap in a major way (IE getting the features they want).

Not surprising though, new features shows up where there's money to make.

[2016]

I mention this because some, if not all, of the code links are dead.

I find this so bizarre. I did some quick back of the envelope math and I've found that I've been programming between 10k and 20k hours over the course of a 9-year professional career. That seems simultaneously low and high depending on how you look at it. I don't know that I can draw any concrete conclusions or correlations between time spent and proficiency once you get to such a large amount of time.