Exactly, because it's not true.

Machine Learning (or rather, Deep Learning and Neural Networks) do not "compress the data". They analyse data. They don't store any original art used in the training (otherwise, the size of these models would be in the thousands of terabytes. Instead we see them being a few gigabytes).

Furthermore, these models do not replicate the art it has been trained on. Every single piece of art generated by AI, is something entirely new. Something that has never been seen before. You can debate if it takes skill, but you can't debate that it's something new.

This video is an excellent source of information regarding this topic. It's created by a professional artist who has embraced AI generated art as a source of inspiration and to speeding up their own work.

Even furthermore, courts have indeed shown previously that Google IS allowed to data mine a bunch of data, and use this. Google has their "Google Books", which is a record of an enormous amount of books, which has been done via data mining - of course, there's a difference between the Google Books project and AI art models, due to the end result (one is a collection of existing stuff, and the other is one that can create new stuff). But the focus here was on the data mining.

One thing that a lot of people don't seem to know: You do not own a style. You cannot copyright a style. There have been a lot of artists that complain because "it's possible for people to just mimic my work". But yes, that is true, but it has always been true - simply because you do not own "your" style. People have always been able to go to another person and say "please make some art, in the style of this person". You have copyright for individual piece of art, but not the general style that you use to create said art.

Here comes my own personal opinion:

Tools using AI are the future. People are not going to lose their jobs because an AI makes them obsolete - people are going to lose their jobs if they refuse to use AI to improve their workload.

Take software development. These models can generate code from the bottom to an insane degree of detail. You no longer have to spend time on all the boring stuff, actually writing the code, you can focus on the problemsolving. The same goes for art: with AI tools, you get to skip the boring monotonous part of your workload, and you can focus on the parts that actually mean something.

Might be. The phone I had was a 6S, and the fingerprint scanner worked at best 50% of the time.

All I can say with certainty is that I followed the setup instructions to a T. I also ended up doing the setup process many times over, because it was incredible how bad the fingerprint scanner worked. But it was always bad.

But yeah, it caused me great frustration. So I'm really happy about FaceID. That, and the gesture based navigation, was the reasons I switched to an iPhone X as soon as it came out.

But yeah, if it wasn't clear before. My iPhone had a better fingerprint scanner than my laptop. That I will give Apple. While it was still trash (in my opinion), it still worked better than my current laptop does.

FaceID is the single best thing that Apple has ever implemented into their phones.

<rant>

That is my take. I hate, absolutely fucking hate, TouchID. Almost never worked. Only gave me frustration. It was so annoying.

With FaceID, I take out my phone, I look at it and swipe up, and then I can use it. It always just works. Anyone who ever complains that "but I have to look at it", well yes, in order to use your phone, you do have to look at it. This goes for both Touch- and FaceID. At least with FaceID, that's the only thing you need to do (and it actually works).

It's so bad, that I'll never buy a phone that does not have FaceID. My iPad doesn't have it, and it's annoying. More often than not, I just use the password, so I don't have to sit through several attemps at using TouchID that will inevitably fail multiple times, thus forcing me to use the password anyways.

Same thing goes for laptops. My current one has a fingerprint scanner, and it works like fucking 10% of the time. Sure, it's cool when it works, but it's so rare.

I really despise fingerprint scanners.

</rant>

Something that is never mentioned for these posts, is the value of money, or PPP. The value of 1 USD varies from country to country. In the EU, you might be able to get a tiny 2 cm piece of chocolate, while in Africa you can actually buy proper food. This applies for defence spending as well.

Furthermore, there is a lot of stuff that is not included in the e.g. Chinese defense budget, that IS however included in the US one (but stuff that China nonetheless spends money on). This should be corrected before one can make a proper comparison.

If one applies a modifier accounting for the PPP difference, then China is suddenly at $465 to $514 billion.

If one applies a modifier for both PPP and wage differences, then the equivalent budget would be around $882 to $988 billion.

Short video that goes over this, if one might be curious: https://www.youtube.com/watch?v=o39SFpfr6E8

But well, do note. It's China. So numbers are only guesstimates.