And you won't uninstall DuoLingo because that would break your streak ;)

I doubt the permanent change in users due to user protests is noticeable. There may be a protest peak, then they keep silent for a moment and then people come back or new people start using the app ignoring past shitstorms.

It is also not reasonable. But suing people for music playing in the background of their home videos is not reasonable either, nor is wanting kindergartens to license the music they sing.

They likely want to limit it early on. For two decades, people have been able to self-publish music on the internet and no longer need the music industry. For some reason, this never posed a significant threat to their income or control, as the ways consumers discover music remained largely the same. The challenge was more about how self-publishing musicians could find an audience.

Now a new frontier is emerging, a whole new category of music. This could include not only AI-generated music created by humans but also AI systems that allow users to generate their own music ("prompt your next playlist"), collaborative approaches, and potentially fully automated systems. Since these are all novel, it's possible that they will first appear outside the control of the established industry, perhaps through a startup that sees no need to collaborate with the current recording industry. Such a startup could become the alternative gateway through which people discover new music, something that didn't happen for self-publishing musicians before.

If these large companies, including YouTube, Deezer, and others who wish to maintain the current state, establish that AI-generated music, identified by content ID systems that detect short snippets of other works, requires licensing, they may later attempt to sue independent sites that host AI music. These sites might not see the need to search for small pieces that, with enough imagination, could potentially come from existing songs.

When I criticize the approach, I am talking about music that is actually generated. If something cuts together existing music, of course that should be licensed. However, the generators I tested do not create works that are substantially similar to other pieces on a large scale. I suspect, though, that if you examine the similarities down to the level of individual seconds, you can find another song that has a very similar second, regardless of whether you recorded it yourself.

You're a Youtube manager. The recording industry says "use our new content id system and force AI musicians to buy licenses for our similar music, or we do not license our music to you anymore". What do you answer?

That doesn't work as entertaining as you think, but the concept is adversarial noise. You add a crafted noise to the song (if possible unhearable by humans) that detracts the classification from the song being AI. Of course that's a cat and mouse game and rarely as robust as you'd wish, if you look for example at glaze.

They may or may not succeed in making this a law, but the article suggests they certainly want to deplatform tracks that have stems similar to those in their portfolio. If YouTube and Spotify cooperate, independent artists will have to pay their tax for having music similar to anything in the vast portfolio of large recording companies. Music isn't that unique. If one has a large portfolio and looks at short segments, one always finds something similar and then can demand to be compensated for that similarity.

What about AI music played live? Google recently released a DJ tool capable of generating music in real-time. The DJ can control how the music changes over the next two seconds, with a 10-second context window. This isn't a tool for creating songs but rather for performing live music.

https://magenta.withgoogle.com...

If that isn't obvious: Capitalism also requires regulation. Not because it would be unstable, but because the stable state it achieves without regulation results in significant societal issues. Unfair starting conditions lead to unfair outcomes, and society must help provide equal opportunities and support those who cannot achieve enough to live well.

Capitalism is not the best system, but the only stable one.

If you suspect that everyone will at some point try to scam others for their own advantage if they get the chance to, then capitalism is not only honest about people being egoistic, but also provides a system that gains stability (and inflexibility) *because* everyone is trying to maximize their own profit.

But that's a stable system like the outcome of the prisoner's dilemma. Everyone acts selfish so the overall outcome is worse for everyone. The problem is, better systems need either strong regulations or collapse, because the first person who stops cooperating makes more profit by stopping to cooperate, as long as the others still cooperate, just like in the dilemma.

It won't take long for the first false positive to occur, just look at the mistakes made by YouTube's Content ID system.

This could also lead to a dangerous system where companies (let's be honest, it's not about the artists but about the music industry) trademark certain musical styles and sue anyone who uses them, regardless of whether they knew someone else had used that style before. That's bad news for independent artists, regardless of whether they use AI or not.

Bluesky works like a centralized network because it is factually centralized at the moment.

Mastodon doesn't really aim at centralization at all. Your hashtags are also not network-wide. If your post was never federated to an instance, a search on the instance will never find them. That means it is a network with clusters. Some instances are highly federated with each other (let's say a tech cluster and an political cluster, for example), but have fewer cross-followings between the clusters, such that each has some content that is not available on the other cluster. That doesn't mean it cannot arrive there, but that would need, for example, someone with cross-following who receives the post so it arrives at one instance, and when the person then boosts it, it is federated to all their followers, thus arriving at a large number of instances in the cluster.

One may say you're caught in a bubble, another person sees it as a safe space, and in the end, you avoid a lot of conflicts between people who see the network completely differently and would prefer the others not to be on the network. The crowd who's afraid of "tech bros" won't see too many tech-bro posts because they are not on an instance that is close to the "tech bro" cluster, so they won't have many fights among each other.

"And ChatGPT complains that saying 'thank you' costs them millions?"

Simple solution:
if(input == "thank you") {
    return "You're welcome";
} else {
    return process_query(input);
}

I think the factor 10 isn't that bad anyway. A Google search is a highly optimized database query. If they wouldn't beat text generation by an order of magnitude, they would do things wrong. Rendering this HTML page in your browser probably costed more than a Google query as well (on the server-side, of course your browser afterward rendered the result page so it adds up).

You use the assumption, that your current water usage is already optimal. The shower head that uses less water works, because you're currently wasting water and it reduces it to the amount that is required to clean your body. There is no use in the surplus water of the full stream, it just goes down the drain.

Here the source for the image TL;DR image: https://blog.kyleggiero.me/Ima...
Also a good read that puts things in relation.