Doesn't matter. Under the US Court of Appeals ruling, the agreement to the terms of GPLv3 would apply, at which point all associated software would be subject to GPLv3 terms. Even if one remade everything from scratch with an AI, one would remain subject to the terms of the license.
> LLMs won't add information to context, so if the output is larger than the input then it's slop
That doesn't align with my observations. A lot of times they are able to add information to context. Sure it's information I could have added myself, but they save me the time. They also do a great job of taking relatively terse context and expanding upon it so that it is more accessible to someone who lacks context. Brevity is often preferable, but that doesn't mean larger input is necessarily slop.
Yup. The comment about the LLM generated PRs is telling. The complain is the LLM generated PRs don't describe design intent. You know how to avoid that? Tell the LLM to provide intent, and if need be, give it the intent. A PR which doesn't capture intent should be categorically rejected and the parties responsible should expect to never get a PR through without it.
I think mistercheph is right to be concerned. This bill applies to all "operating system providers", defined thusly:
(g) “Operating system provider” means a person or entity that develops, licenses, or controls the operating system software on a computer, mobile device, or any other general purpose computing device.
Regarding penalities:
1798.503. (a) A person that violates this title shall be subject to an injunction and liable for a civil penalty of not more than two thousand five hundred dollars ($2,500) per affected child for each negligent violation or not more than seven thousand five hundred dollars ($7,500) per affected child for each intentional violation, which shall be assessed and recovered only in a civil action brought in the name of the people of the State of California by the Attorney General.
>This bill applies to all "operating system providers", ...
Not really.
>...for the purpose of providing a signal regarding the user’s age bracket to applications available in a covered application store.
So the OS has to provide an age signal to apps from a "covered application store" defined as:
e) (1) “Covered application store” means a publicly available internet website, software application, online service, or platform that distributes and facilitates the download of applications from third-party developers to users of a computer, a mobile device, or any other general purpose computing that can access a covered application store or can download an application.
(2) “Covered application store” does not mean an online service or platform that distributes extensions, plug-ins, add-ons, or other software applications that run exclusively within a separate host application.
It doesn't say "only if there's a covered application store present on the system". But maybe everyone in power will interpret this non-logically in exactly the right way that this doesn't become abusive.
Wouldn’t that classification apply to Linux package managers as well?
They are publicly available online services that distribute and facilitate the download of applications from third party developers to users of a general purpose computing device.
That may not be the intent, but it seems like it would still apply. Many of the “app stores” on Linux are just front ends for the package manager in some way.
I assume the people behind this don’t know things like apt or dnf exist, so it likely wasn’t considered.
There are a large number of things that the people behind this don't know about. That doesn't mean that they screwed up. It means that the law does not apply to those things.
Except that there are tremendous advantages to constant-time execution, not the least of which is protection from timing security attacks/information leakage (which admittedly were less of a concern back then). Sure you can get the one instruction executed for the <6 case faster, but the transistor budget for that isn't worth it, particularly if you pipeline the execution into stages. It makes optimization far more complex...
reply