It seems that within the past few years, there have been a great increase in the number of articles on low-level GPU internals/programming, which was previously a very niche subject. However, once you understand that a GPU is conceptually a processor specialised for graphics operations, with its own instruction set and RAM, then everything seems quite straightforward, and you can actually start thinking about how to implement one.
The specification is a behavioral target, rather than a circuit diagram.
Did anyone else get a slight feeling of LLM-ism from that sentence? Not a complaint, because I've noticed AI is definitely leading to some great boosts (the somewhat related https://news.ycombinator.com/item?id=49717638 is on the front page as I write this.)
If you're going to make some kind of argument about the benefits of copyright for preserving software jobs, keep in mind this is a driver for hardware apple sells. Everyone would be better off if drivers were mandatory open source.
I write software for a living and think we should abolish software copyright, and furthermore, source availability should be a basic consumer protection requirement. Basically the free software definition should be law.
Of course I think it would change things. You think I haven't considered that? But I also think society would be a better place if we worked to produce new software that people desire instead of gatekeeping what already exists, and if the law banned this one-sided gatekeeping. The goal of all people should be to build lasting wealth for the next generation to inherit, not to hoard what is naturally infinite.
Plumbers don't extract subscription fees from your use of the toilet. Electricians don't get a fee for your use of a light fixture. They don't get a monopoly on fixing your things because you bought it through them. The whole idea of IP is absurd. It should be the opposite: anyone should always be free to go to someone else if they become unhappy with us, taking the work they bought from us with them. The law should bar any attempts to stop them/build roadblocks.
Happily, I see LLMs have great potential to make everything effectively FOSS now anyway, so all we really need is regulations against anti-competitive behavior like requiring signed firmware.
> I also think society would be a better place if we worked to produce new software that people desire
That's the entire point of copyright -- to incentivize this work by conveying ownership to those who produce such software. Without ownership, others would be free to take it and give you nothing. The world doesn't work for free. I doubt you do, either.
That's the ostensible point, and I think does the job extremely poorly, and is often if not usually counterproductive. Without payment, there'd be nothing to take, since as you say, the world doesn't work for free.
Software either has some functional purpose, so people will pay to see its creation (e.g. device drivers; no one will buy your hardware if there's no software for it to do anything), or it exists as e.g. art, in which case patronage avoids the damage of copyright. Either way, if you want to incentivize creation, then you should want a system where payment is for creation, and where further creation (e.g. paying a different developer to modify it) is freely allowed. Copyright is a system where people pay to rent, so once it's created you can rest on your laurels forever unless someone undertakes the inefficiency of recreating the same thing for no reason before they can add to it.
> if you want to incentivize creation, then you should want a system where payment is for creation, and where further creation (e.g. paying a different developer to modify it) is freely allowed.
Why is that? I would think that the possibility of exclusivity over derivative works is an even greater incentive to creation.
And yet we have proof right in front of us that e.g. Apple will not port their GPU drivers to Linux, and people had to reverse engineer it, so clearly the evidence says it does not create that incentive. In fact, copyright law for software is quite frequently used to manipulate and control appliance owners or end-users, even when the software itself is not really "the point" of the thing (e.g. putting ads and spyware in cars or locking out the ability of the owner to use their heated seats).
> And yet we have proof right in front of us that e.g. Apple will not port their GPU drivers to Linux, and people had to reverse engineer it, so clearly the evidence says it does not create that incentive
On the contrary, Apple engineers--paid professionals--wrote the driver to make MacOS work with the integrated GPU.
I don't know why you think not writing a driver for Linux is proof of anything, other than that Apple does not care one whit about Linux. They have no interest in that. They build integrated computer systems (hardware + software); to build them for Linux is, to them, an utter waste of resources.
> copyright law for software is quite frequently used to manipulate and control appliance owners or end-users, even when the software itself is not really "the point" of the thing
That's right. Builders of things get control over what they build. That control is designed to maximize profit, which is a strong motivator for building novel and useful things. If you don't like what they're building, buy something else, or build something better! That's how capitalism is supposed to work.
Okay, but nobody can pay a random software engineer to go take the Apple driver's code and use it to make a working Linux driver that they can use on the hardware they purchased. So Apple's monopoly on their source code is not incentivizing any creation (they had to write it anyway to make the hardware useful), and is in fact disincentivizing it (now if someone wants e.g. a Linux or Windows driver, they need to undertake the previously monumental task of also reverse engineering the hardware, which is much more difficult than the original software creation).
And product tying (requiring that if you want to buy Apple's hardware you must also use their software) is an anti-competitive practice that harms society, which is why I oppose it. The world is a better place when you can plug your Dell monitor into your Apple computer and it just works. It's also a better place when you can buy your operating system from Canonical if you'd like to run on your Apple computer. It means we all have more choice. Even if Apple doesn't want to have any part in making that work, we should say they have to release the necessary information to let you pay someone else to do it if you'd like, or at least let people re-use whatever they do get from Apple to make it work (e.g. look at what Apple's binary code does). As I said, it's a basic consumer protection requirement for the 21st century where everything has software now.
Society gains literally nothing from letting them have a monopoly on software necessary for their hardware to run, so why give it to them?
> The world is a better place when you can plug your Dell monitor into your Apple computer and it just works. It's also a better place when you can buy your operating system from Canonical if you'd like to run on your Apple computer. It means we all have more choice.
The thing is, you're suggesting imposing that choice on the unwilling. What if I, as a creator, don't want to see the software I put my blood, sweat, and tears into running on inferior (or less capable) hardware, or vice versa? Whether that is "better" for the world isn't obvious to me.
> Society gains literally nothing from letting them have a monopoly on software necessary for their hardware to run, so why give it to them?
I think it comes down to whether you believe hardware and software are always severable things, or whether you believe they comprise an integrated product. If you think like a hacker, they are the former. If you think like a product person, they are the latter. Woz was a hacker. Jobs was a product guy.
What if I, the owner of Standard Oil, don't want my oil in inferior cars? Or if I, the owner of AT&T, don't want inferior phones on my network? Too bad. The correct course of action to take with companies that try that is to dissolve them.
Programmable hardware and software are obviously always severable. This is not up for debate, but is a basic fact. You can always sell the product you want. That doesn't mean you have the right to control how the owner uses it. If they want to install Linux on it, that's their prerogative. If they want to shoot it out of a cannon or detonate it with an M-80, also their prerogative. If someone wants to sell an emulator for your hardware that runs your software, also fine (and explicitly ruled legal).
> What if I, the owner of Standard Oil, don't want my oil in inferior cars?
That makes no sense, either as a hypothetical or a historical reference.
> Programmable hardware and software are obviously always severable. This is not up for debate, but is a basic fact.
That's a fact you made up that sounds "truthy". It's an opinion. You're entitled to your opinion, but you're not entitled to make up your own facts. And yeah, it’s up for debate. You know how? Because people have been debating this for decades.
Think about all the gazillion pieces of hardware out there that have dedicated control software that works closely in tandem with the hardware. Practically nobody is clamoring for the absolute freedom to replace the hardware in the stability control systems for for their vehicles or the autopilot systems in airplanes.
> You can always sell the product you want. That doesn't mean you have the right to control how the owner uses it.
You must be unfamiliar with this thing called "the law." You don't "own" the software component of an object when it's purchased. What you do "own" is a limited transferable license to use it, which conveys with the physical object. Same with all forms of physical prerecorded media.
Perhaps you are arguing for how something should be. And that's fine, but that's different from what is.
others would be free to take it and give you nothing
And that's fine, because I'm being paid to provide a service, actual labour (which might admittedly involve prompting an LLM), but it's still doing something new. There's tons of people being paid to work on OSS, and much of which would still exist by volunteer effort alone.
> There's tons of people being paid to work on OSS, and much of which would still exist by volunteer effort alone.
Those people are largely working on OSS as a hobby or for clout, i.e., to kickstart their career; or, they're working on a loss leader that will (if history is any indication) eventually lead to a license rug pull when they find out GPL/BSD/MIT licenes aren't a path to riches.
I'll concede that hobby projects that have led to Linux, programming languages, etc. do provide valuable connective tissue to the world, but most of the stuff that appeals to the end user isn't OSS.
> I'm being paid to provide a service
If you're like most of us, you're likely being paid to provide a service that depends on copyright as a foundation. Ask your company's lawyer their opinion on what the company would do if your work wasn't protected by it, or if someone broke into your company's source code repos and stole it.
The Internet Archive’s Wayback Machine has been hit by waves of high-volume automated traffic
Thank you for not immediately blaming it on "AI bots". I suspect there's some entity manufacturing consent for strong identity/age verification/sanctioned-browser-OS "walled garden" Internet, and these random DDoSes are part of that.
I knew something was up when a few alternative YouTube front-ends I use suddenly put up the 'nubis and complained about the high volumes of traffic they were getting flooded with; of course someone actually going after that data would be aiming their "AI bots" at YouTube directly instead of trying to suck it through a tiny little-known proxy-site, so it really strained the credibility of the argument.
...and the only way to stop[1] that is by effectively DRM'ing everything, which is a level of dystopia that I don't think even Stallman ever anticipated, nor do I want to happen.
[1] Analog hole and other workarounds aside, naturally.
CF classifies anyone not using a popular browser with Javascript enabled as a "bot"
This is my biggest complaint about CF. They are implicitly supporting user-agent discrimination in favour of Big Browser, instead of discriminating on actual behaviour.
...and of course there are already companies running tons of VMs with "officially sanctioned" browser + OS stacks, that can get past all these "protections", for a fee.
"AI bots" is the newest boogeyman they came up with to take away freedom.
Yes, and it's not quite the same as a normal "INT 01h". It causes a debug exception, which may enter ICE mode if it is enabled (undocumented bit in DR7, or PMCR on Pentium), otherwise it invokes interrupt 1, but without checking the privilege level on the IDT entry, or the interrupt redirection bitmap in V86 mode.
IIRC, older versions of the Linux kernel had a security bug because they didn't expect this to happen.
ICE mode was sort of a precursor to SMM, but both also coexisted for a time with slightly different behaviour. It was introduced in the 286, where instead of ICEBP there was "STOREALL" (opcode 0F04). F1 on that processor was a prefix instead, with the same function as UMOV on 386+. If you use them together - something Intel probably didn't intend - you can dump the internal CPU state to memory on a regular non-bond-out chip.
The specification is a behavioral target, rather than a circuit diagram.
Did anyone else get a slight feeling of LLM-ism from that sentence? Not a complaint, because I've noticed AI is definitely leading to some great boosts (the somewhat related https://news.ycombinator.com/item?id=49717638 is on the front page as I write this.)
reply