Then I'm sure the developer of the app would be happy to screen for copyrighted content or malware, and attempt to remove it. It wouldn't even be hard, they could just poke a lightweight LLM at it and have the llm make the call, and maybe add a submission field to report illegal content.
That’s what we’re preventing here. The agreement is ISP’s/infrastructure gets the clear, and applications need to regulate the content they host. If they stop regulating content, ISP’s will
That's nonsense using a disingenuous framing. Application like Browser have existed long before and after ISP DMCA exemptions, applications do not have any responsibility to enact laws on someone else's computer/property.
Otherwise why not also remove safari since apple does not prevent copyright abuse there.
It's not disingenuous framing, it's just an attempt to understand the motivations of the people writing the laws. What would be disingenuous would be comparing an App Store or an Authorized developer account to a web browser.
In the US, Digital Millennium Copyright Act (DMCA) of 1998, Section 512
In the EU, E-Commerce Directive (2000/31/EC), Copyright in the Digital Single Market Directive (2019/790/EU, “DSM Directive”), Digital Services Act (DSA).
Additionally specific countries each have their own laws on the matter.
DMCA would require something like a torrent tracker to respond to specific takedown requests in the prescribed manner. It has zero relevance to a torrent client.
1. You have absolutely no clue how the BitTorrent protocol works.
2. You have never maintained a widely used app as a single developer.
3. The extent of your use of LLMs is either academic / hobby or very narrowly focused and not integrated into a global product.
Just these points make your "suggestion" about using "LLMs to detect stuff is extremely easy" laughable at best.
The reason why people want to install their own software is to have freedom over their devices. The copyrighted content removal has a mechanism for it, called DMCA. And this is not how it works. The application does not have any content or means to circumvent any measures.
> Just these points make your "suggestion" about using "LLMs to detect stuff is extremely easy" laughable at best.
I ran a company doing this for real time internet traffic and the tech worked great. My mistake was suggesting a specific solution; the reality is there is a dozens ways to go about it, and it doesn't matter to me how its solved. What does matter is the EU probably isn't going to work overtime to protect people illegally downloading music, and I don't fault Apple for wanting to limit how many people can do it
