General_Effort @ General_Effort @lemmy.world

Posts

166

Comments

1,955

Joined

2 yr. ago

I really don't get how opinions on intellectual property and its "theft" turn 180 whenever AI is mentioned.

No. At least, the last version I saw wanted text messages to be scanned for "grooming".

People want goods and services, as well as jobs. Politicians need to make that happen, and so they listen to the people who know how to make that happen. Sometimes that goes wrong because eventually employers don't have quite the same goals as their employees. There is no good alternative, though.

One player that clearly had a lot of input is the (news) media. EG the press publishers want to license their old news articles for AI training. They can do that thanks to EU copyright law. That's free money. But news articles talk about living people, which means they contain personal data.

Despite competition from social media, the trad media, including press publishers, is still extremely influential. Politicians need their favor to get votes.

I don't see how Big Tech is getting much here. Of course, NGOs need the media's favor just as much as politicians. Pointing the finger at some nebulous forces from outside is certainly the safest choice, politically speaking.

Das ist ein echtes Problem. (Nicht nur Alte-Leute-Gemecker, wie es in der Überschrift klingt.)

mein Vater, obwohl “einfacher Arbeiter” hatte immer schon Spaß an Technik und war zu Atari Zeiten auch ein begnadeter Programmierer

Wow. Das muss ein Charakter gewesen sein.

Übrigens Respekt für den Bildungsaufstieg. Ich hatte Easy-Mode. Mich beeindruckt sowas immer.

There's that episode of The Orville where social media is the law.

https://orville.fandom.com/wiki/Majority_Rule

Maybe I’m talking more about enforcement than actual law,

Probably. Different countries in Europe have very different traditions there. I think the former socialist countries are still more relaxed. But the EU-line is rather dominated by countries like Germany.

Come to think of it. Switzerland officially takes a very lenient approach. It's legal to download media files for personal use. But as you can see here, that leaves research and business hanging.

I don't know any European country that has anything like the copyright clause in the US Constitution, or anything close to Fair Use. I don't see the argument.

There's a good chance that European AI companies like Mistral are breaking the law. We will have to see how it eventually goes in court. Recently, there was a decision in a Munich court against OpenAI. By that standard, even Apertus might be in trouble. But I doubt that decision will stand.

Plutonium-238’s half-life is 87.7 years, Americium-241 is 432.6 years. Which… is almost 5 times longer, so… not sure why that’s cringe?

What's cringe is the word "staggering". Natural radioactive isotopes have half-lives on the order of billions of years. All elements heavier than iron are created in supernovae. Billions of years have passed since the novae that created that heavy elements now on earth. Anything with shorter half-lives is no longer around. (More correctly, one should talk decay chains.)

What's staggering is that these isotopes are available at all. They are artificially created in nuclear reactors. Mass production of Pl-238 began only during WW2 for bombs. That's almost a half-life ago. The shorter half-life makes the availability of Pl-238 much more impressive.

I believe they’re referring to the fact that it’s not an element of major topic. This is the first time I’ve ever heard of it.

There are over 100 named elements. I don't think I could name half of them. Americium is relatively prominent because of it's use in smoke detectors. And while I'm at it: Americium is the element. Americium-241 is a specific isotope; a specific variant, chemically identical to other variants but with slightly different physical properties.

There are a number of isotopes suitable for RTGs. It's a matter of trade-offs. There's half-life, which is basically how fast the properties of the material change. There's also energy density and how bad the radiation is for the device. And always, there's cost. Fun fact, in Chernobyl they did try robots, but the electronics could not withstand the radiation. People don't withstand it either, but there's a lot of them.

For fastest inference, you want to fit the entire model in VRAM. Plus, you need a few GB extra for context.

Context means the text (+images, etc) it works on. That's the chat log, in the case of a chatbot, plus any texts you might want summarized/translated/ask questions about.

Models can be quantized, which is a kind of lossy compression. They get smaller but also dumber. As with JPGs, the quality loss is insignificant at first and absolutely worth it.

Inference can be split between GPU and CPU, substituting VRAM with normal RAM. Makes it slower, but you'll probably will still feel that it's smooth.

Basically, it's all trade-offs between quality, context size, and speed.

Badly. This was released almost 3 months ago and completely failed to make a splash, just like the other European models you have never heard of.

There's a lot of denial about this, but you just can't make such models competitive in Europe. The copyright industry is too strong. This combines with a general culture of data ownership and control. Case in point, the copyright industry is especially strong in Denmark, and they are also champions of chat control.

Yes. And also:

Its half-life is a staggering 432 years, five times longer than plutonium-238.

Cringe...

AI slop?

My suspicion is always the copyright industry. That's where you got a gazillion dollar financial interest in controlling the flow of information. Indeed, the lobby organization "Thorn" is a Hollywood creation.

It's noteable how the language around copyright infringement has changed. The industry used to call it "theft". Now they use the language of sexual assault and talk about "consent". It's very sexist, too. Poor, helpless women become victims of so-called "digital violence". Kinda fits with that recent prosecution of a Redditor.

GEMA's social media game is certainly top. Nice to see the money being put to good use.

If all seats are filled,

Sure. Everyone's Taylor Swift. Let's just assume that.

The actual truth is that if you do not play GEMA music, you have to provide evidence of that to GEMA. Young musicians who foolishly reason that they don't have anything to do with GEMA will be dragged through court.

Fair Use exists only in the US. I believe it is part of the reason why the US became so culturally dominant. It certainly is why the internet is US dominated. European copyright laws are stifling.

GEMA was created by the Nazis to take over pop culture. There's a certain logic there.

Of course, under the Nazis you could be sent to camp if you belonged to the wrong subculture. So there is a difference. The rebels of the time listened to jazz, to swing. "Negro music" was the social media of that age, corrupting the youth.

Ich fänd es ja schön, wenn sie nicht einfach über alle Nutzer berechnen würden, sondern mein Beitrag anteilig auf meine Streams aufteilen. Dann geht mein Geld auch an wen ich höre. Und wenn ich weniger höre kriegen die halt mehr pro Abruf… Scheint aber nicht üblich zu sein so zu rechnen und damit geht halt das meiste Geld an die großen und populären Künstler die ich nicht höre.

Ich glaube, die Logik ist klar. Die berühmten Musiker haben für sich schon mehr Einfluss. Hinter den Großen und Halbgroßen stehen die Major Labels, die maximale Verhandlungsmacht haben. Die handeln natürlich Verträge in ihrem Sinne aus. Bei der GEMA hat man intern dasselbe Problem.

Vielleicht gibt es Nischenanbieter, die sich auf deinen Geschmack spezialisiert haben, unabhängig von den Großen.

Die GEMA hat anscheinend mal eine Veröffentlichung zu Streaming zusammengestellt mit vielen schönen Grafiken.

Man kann ein paar Sachen ablesen. Etwa 70% des Umsatzes wird an Rechteinhaber ausbezahlt. Das heißt, wenn die Server umsonst wären, wenn 100% ausbezahlt würden, dann wären das kaum 50% mehr.

Wenn Spotify also jetzt ~EUR 3 pro 1000 Streams zahlt, wären das bei 100% Auszahlung knapp EUR 4,50. Da frage ich mich, wie andere EUR 12 oder sogar 18 zahlen können, ohne dass das Abo teurer ist. Möglichkeit 1: Die Vielhörer sind alle bei Spotify. Das heißt, weniger Geld pro Stream. Möglichkeit 2: Spotify macht mehr gratis/werbefinanziert, was dann weniger Einnahmen pro Stream bringt. Möglichkeit 3: Die anderen Dienste verbrennen Venture-Kapital.

Mit dem Angriff der Altersverifikation wird das alles in Ordnung kommen.

That is not dead which can eternal lie, ...

(And I do mean lie.)

https://en.wikipedia.org/wiki/Anti-circumvention