I don't think it has much to do with how "complex or not" it is, but rather how common it is.

It can completely fail on very simple things that are just a bit obscure, so it has too little training data.

And it can do very complex things if there's enough training data on those things.

I've heard this comment about OpenXML (the xml format of the office documents) before, and i'm a bit on the fence about it.

It's of course indeed ridiculously complex, but so is office. Microsoft both adds a shit ton of functionality to their documents, and keeps an impressive amount of backwards compatibility.

In the past i heard complaints about part of the OpenXML spec that also allows older binary data in there for backwards compatibility reasons, which of course means for OSS implementations that they don't just have to implement this spec, but also the older spec that came before to be truly compatible with everything a modern office version can open.

But on the other hand, if i look at it from the side of Microsoft, they opened up their format, they've got a gazillion functionalities, should they remove functionality to appease the open source developers? If so which? Should they stop being backwards compatible with documents of decades ago to appease the open source developers? If so how long should they support? Are you going to tell their customers?

Office is an immense program with an immense amount of legacy features, backwards compatibility, ....

It's incredibly complex by nature. And might they have made the format more complex to dissuade competition? Could be. However, in this instance Occam's razor pushes me more to "write a huge program over a timespan of many decades, with thousands upon thousands of programmers working on it, and you'll indeed most likely end up with something very complex...."

That's the claim, it has apparently been trained using a fraction of the compute power of the GPT models and achieves similar results.

If you need far less computing power to train the models, far less gpus are needed, and that hurts nvidia

I know it's a shitpost, but i hate that people interpret carcinisation as a crab being the ultimate piece of evolution.

When learning evolution like algorithms in computer science, one of the first things you learn is strategies to not get stuck in locally optimal solutions (solutions that seem the best when you look at other nearby solutions, but are worse than other solutions if you allow your algorithm to look further away).

Crabs seem like that, it's just an easy defensive evolution that then stagnates in a form that kind of works. Seeing how many crabs we eat, and how few crabs eat us, it's obvious that crabs aren't the actual pinnacle of evolution, just some locally optimal solution that evolution tends to get stuck in :p.

Whomever wrote this article is just misleading everyone.

First of all, they did this for other kinds of similar instruction sets before, so this is nothing special. Second of all, they measure the speedup compared to a basic implementation that doesn't use any optimizations.

They did the same in the past for AVX-2, which is 67x faster in the test where avx-512 got the 94x speed increase. So it's not 94x faster now, it's 1.4x faster than the previous iteration using the older AVX-2 instruction set. It's barely twice as fast as the implementation using SSE3 (40x faster than the slow version), an instruction set from 20 years ago....

So yeah, it's awesome that they did the same awesome work for AVX-512, but the 94x boost is just plain bullshit.... it's really sad that great work then gets worded in such a misleading way to form clickbait, rather than getting a proper informative article.....

At the very least it's a copy paste from a popular reddit post, and it indeed reads like a shitpost that combines every single trope...

I recently read an article about a doctor who was making a case that the issue is not that those 1 in 5 are "neurodivergent", but our current society is causing harm. When he sees ADHD symptoms his first "treatments" are proper nutrition, making sure they feel like they're doing meaningful things in life, enough exercise, etc...

I'm also sometimes starting to wonder if for a part we're not just medicating people to "thrive" in a society that's inhuman, rather than make society work for as many people as possible.

But it's of course a very complex & grey area, and let's be honest, something as vague as ADHD probably encompasses a lot of different causes. And it'll probably take decades of research before we actually manage to split up all the things that are today lumped together into the separate things with each their own propert treatment.

I love how you wrote all this, and are completely missing the mark. Nintendo is filing a lawsuit claiming that the palworld devs violated their patents, not their copyrights.

Anything palworld 'copied' from pokémon is either japanese lore, or from older games. This is not a copyright suit. If a copyright suit were possible, Nintendo would have brought it waaaay earlier. I'm wondering which patents Nintendo has that were supposedly violated.

I love how there's this entire discussion here about copyright etc... while that's not even what this is about.

Isn't it also just because it's old and people get bored of it? People crave new things, and even if it's just as good as in the beginning, it'll get lower ratings because it's not new anymore.

I remember quite some years ago i was like "i'm finally going to watch southpark". And people were already complaining about how the latest seasons were worse than the first seasons. Watched a ton of seasons in a short period, and honestly can't say the later seasons felt any worse than the first ones when you're not bored of the series yet. Now so many years later when i watch some more southpark, it's not as fun as when i started watching it since the "it's new and exciting" feeling is long gone.

As someone else already posted, the 3GB was incorrect, it was a router reporting incorrect traffic.

But that doesn't seem to stop everyone here from continuing to post how the thing that didn't happen in the first place is ridiculous...

Isn't HDR support on linux just a nightmare in general? I guess Steam is just waiting for linux to get its act together on this decades old feature rather than join in the madness it currently is.