What you are describing is option types, which are an entirely valid and very useful construct that helps make programs more rather than less reliable. But you need proper language type system support and compile-time enforcement to make it work, and C does neither of those.
This is Sturgeon's Law: "ninety percent of everything is crap".
But the corollary to it is the fact 10% of everything isn't crap, and just as there always has been crap, there always has been excellence. Just less of it.
I experimented with turning out complete airport thriller novels this way, using earlier LLMs. It's not terribly hard to make this happen, the hard part wasn't gunning out prose, it was plot arcs and internal consistency, but even that was suprisingly easy to solve.
Of course I didn't do anything with the idea, for what I hope are obvious reasons.
Ironically, the article points out that the original authors publisher actually put out two DMCA notices to google last year, apparently with no effect.
I guess DMCA takedowns are only for the big fish fighting the good fight against car pirates.
Why aren't they suing Google in court? Now is the best time to do this politically, the site doesn't mention what state they live in but I'd doubt that you wouldn't be able to get a state AG to listen to you if you reached out.
edit to add: Google has ignored all safe harbor protections, they would lose this protection and be held liable for all damages. This seems like a pretty solid win for the author here if they're telling the truth.
There are a lot of really solid reasons someone might choose not to sue one of the world’s most powerful corporations over their core business practices, even being completely in the right. Pretty similar to reasons someone might not call the cops on some mafiosos that are blocking your driveway while fencing a truckload of stolen electronics.
I disagree, state AGs are taking any opportunity to attack big tech for easy political points. If you want justice, now is the best time to try and get it.
Filling a complaint with the civil division of your AG so they can take legal action vs suing a giant powerful corporation and hoping the AG’s civil division steps in are very different things. I’ve also known numerous people who’ve filed multiple complaints with the MA AG with easily provable and well documented cases of repeated, ongoing wage theft as restaurant workers, at a time where service industry labor abuses were a popular hot button issue, and they didn’t even get a callback. Not exactly what you’d call a slam dunk. Not having support from the AG would be one of the very solid reasons I was referring to.
then DMCA the entirety of google and alphabet, and tender class action for direct and contributory violation, with the option to back it off to the literary work in question when goog takes a seat at the table and takes it seriously
The entire purpose of the DMCA is that it’s a bludgeon that the rich and powerful can use to beat the poor and powerless. It was never meant for individual copyright owners to use against giant infringing corporations.
Speaking from experience, neither apple nor Google will enforce dmca takedown requests unless they come from the courts. Even when provided clear and direct evidence. Their position is "we do not arbitrate". I sort of get it but also the cost barrier to copyright infringement is headed to zero while the ability to protect your IP and enforce your copyright remains expensive. Wild times ahead.
The blind spot is the same one that's always there. Takedown notices from partners are fast tracked, takedown notices from Joe Bob require a court filing.
(It makes some degree of sense - I shouldn't be able to use a burner identity to get Google to take down (even temporarily) a million-subscriber channel. The big problem with the DMCA is the impossibility of proving that a grey-area filer is acting in bad faith, but that's in the wheelhouse of the courts, not the platforms.)
DMCA to take down AI slop? I feel that is twice the mistake. DMCA should not exist in the first place. Neither should AI slop. Here we have AI stealing from people. AI is a thief.
So you're against copyright protection, but also against AI using copyrighted work? I can understand both of those positions separately, but how can you combine them in the same statement?
Yes, I am. Copyright is legalized plunder of anyone who does not pay a protection fee for a "license" to not be plundered. Going after torrenters and people trying to regain functionality on their thermostats and 3D printers is legalized plunder.
there are two different things — one can be against the idea of copyright property, or just copyright laws as written into the legal system, while holding the expectation these should apply to everyone equally
Sure, there are plenty of things to dislike about DMCA or any other modern copyright law, but in this context the discussion was specifically about using DMCA to take down an unauthorized republishing of a copyrighted book online, which is a pretty normal application of copyright law. So I don't really understand how "AI is bad because it's trained on stolen work" but also "it's bad to take legal action against people distributing your work".
There seem to be people who believe that the DMCA is a morally compromised instrument.
There are people that believe that using a morally compromised instrument to do a moral end is always bad.
"AI is bad because it's trained on stolen work therefore we must never use AI, even if our ends are good" is such a belief that many people seem to have, and there seems to be some likelihood that the original poster might be such a person.
Therefore it seems to me reasonable to believe that a person who maybe believes that you must never use AI because it is trained on stolen work, could also believe that you must never use the DMCA because it is based on bad and corrupted law.
I myself do not exactly believe these things, although I consider they may have some arguments for them, albeit not arguments likely to persuade me in all instances, nonetheless I do not find any difficulty in believing someone could hold both opinions at the same time and I think, in fact, it is a reasonably consistent pair of opinions, especially given the apparent ability of people to believe all sorts of inconsistent things day to day.
How do you suggest Qontour be held accountable for this blatant theft of Koenig's work and name if the DMCA is a mistake?
And don't pass the blame off onto "AI" from the people who said "let's make a web site that totally steals this book we like". AI is a tool of thieves, founded upon thievery. Qontour is an agency made up of thieves who are using AI to perform their thievery.
Gala Aranaga, Founder & CEO of Qontour, is a thief.
Jason Chandler, Founder & Creative Director of Qontour, is a thief.
Atif Fazil, Technical Director of Qontour, is a thief.
Pemi Ogunkeye, Webflow Developer at Qontour, is a thief.
Daniela Aranaga, Head of Content & Marketing at Qontour, is a thief.
Ahmed Qayyum, Solutions Architect at Qontour, is a thief.
Bukunmi Ogunmodede, Webflow Developer at Qontour, is a thief.
Hassaan Rasul, Senior UX Designer at Qontour, is a thief.
They used ChatGPT, a copyrightwashing tool developed in a massive act of thievery by the employees of OpenAI, all of whom are thieves. OpenAI was founded by Elon Musk, Sam Altman, Ilya Sutskever, Greg Brockman, Trevor Blackwell, all of whom are thieves.
I'm very much in favour of blocking children from social media - it's an absolutely vile cesspit of cognitive addiction, bullying and social (and potential sexual) abuse. But none of it requires a mass-surveillance network to be put in place.
Just for one example; it would be trivial for Apple and Google to put age estimation on my phone, verify it on opening the web browser and provide a zero-knowledge proof of age to websites in a way that does not reveal my identity. All the infrastructure is already there, and it's relatively trivial to turn it on. The downside is that this will only work for people who are older than about 25 because of the uncertainty of face-to-age recognition, but it would be a start.
Another way to do it is for my bank, who know my age already, providing a similar credential that I can feed into the zero-knowledge proof engine on the phone.
This was all done properly for the covid tracking apps, at a time when the phone providers actually wanted to do tracking with anonymity - this is a similar problem, and it's easily cracked by technical means.
And you don't even need zero knowledge proofs if you perform on-device content detection - turn it on for kids, keep it off for adults. Modern phones have more than enough TPU capacity to do this.
But none of the actual implementations I've seen are truly anonymizing, and they all rely on trusting some really dodgy companies with your identity and browsing habits. Yes, the more respectable ones have security and privacy policies that are audited, but will they always? The cynical answer is "no", because history shows that someone will always do something sooner or later if (a) it makes money, and (b) they can get away with it.
Everything I see suggests that the desire for mass surveillance is the driver, and the "protect the children" front on this is a strategem by the people who are really driving this from behind the curtain. There are huge amounts of money to be made by capturing verifiable, blackmailable, personal data, and this is a magic money fountain for those who will be able to mine it.
In more sensible times, we'd run an ad campaign highlighting the dangers and informational campaigns for parents on what to do to prevent your children getting access to social media.
Perfect is impossible, but if its stigmatised then the network effects stop being so punitive to children who have reasonable parents.
it's the 10-80-10 rule: 10% of kids will still access social media, 10% will never... but 80% can be swayed.
At last. It's time the whole would gets on board with open standards that are truly open, and there is explosive devopment going on in the world of new approaches to media production and distribution that this can only aid.
It's net-head vs. Bell-heads all over again, and one of the biggest reasons for the success of the IETF standards was the no-cost availability of all their standards.
We need a SciHub for standards docs. Since APIs have been ruled incopyrightable in the US, building a library that implements the standard shouldn’t itself be illegal.
APIs have not been ruled uncopyrightable. The Supreme Court found that Google's use of the Java API's was fair use; fair use is specifically a defense against infringement of a copyrighted work.
There may be court cases in the future that determine what the boundaries on API reimplementation are that distinguish fair use from infringement. A future Supreme Court may well overturn Oracle v. Google. APIs are specific forms of unique expression, and the same functionality can be made available through different APIs. (See for example, OpenGL vs. Direct3D.) Typically these are the criteria used to determine what is eligible for copyright, and ruling APIs uncopyrightable absent a statutory carve-out exemption may well put the copyrightability of currently protected forms of expression in jeopardy.
But as things stand, the Oracle v. Google decision has only made the API-copyrightability decision more ambiguous, it has not settled the matter in favor of making APIs uncopyrightable.
You're right, fair use was the ruling. Thank you for the correction. In that case, would the implementation of a standard qualify? Feels like it would?
What might be better is to, in addition to failing hard when the certificate expires, web browsers were to give a 'soft' click-through user warning if the certificate on the site - while still within its validity period - has less than say 7 days to go before expiry.
That's probably long enough for most companies to be alerted to the problem in time and to get their act together to fix the problem.
This is precisely why you do it in water - the water-skin contact is effectively perfect, as is the water-transducer interface, and the body of water is easily characterizable; in effect you are scanning one large object that consists of a body of water that just happens to have a human body in it, and then extracting the body from that scan.