|
Imaginary Friend posted:After watching some YouTube videos about robots and how there are chatgpt-powered ones being developed a question sprung up in my mind; when will we see the first murderbots? I mean, eventually, tech-savvy people will be able to buy a robot and dunk an open source, unrestricted AI into them. Will we see a robot war á la "I, Robot" soon? I think the thing with drones is that they are good at creating videos you can share. So if you are somebody like the governement of Ukrania, and want people on USA to mind about your war, so the USA congress send money, weapons and personal your way, you need to refresh the memories of the war often. "News" are things that are new, and if your war is on the 2 or 3 year, is not new anymore. Basically drones are great for fighting the information war. To create memes on the internet that are shareable and viral. Has for the war effort itself, who knows. Maybe. Drones seems to be used to hunt tanks, since most drones are x10 or x100 cheaper than tanks. That don't stop the use of tanks. Tanks are still a very popular fact of war. Since people still use tanks, don't seems drones have won. What a very technological warfare would look like?, since the first civilizations, war is done with a lot of central control, soldiers are not free do attack when they want and what they want. General choose to allow the enemy advance in a disadvageous position to attack from a advantageous position. Maybe get large group of enemy soldier to surrender, instead of fight. Focus fire saturate defenses, if all your rockets are fired at the same time, anti-rocket defenses are flooded and can't stop but a small fraction, but if the drone or soldier fire at will, the defense will have a easier time. So again, you want a central control, instead of free to go drones that do whatever they please. Murderbots are not much a matter of advanced technology, you can make something like a autoturret that automatically murder any target in his cone of vision. https://www.youtube.com/watch?v=SnTqK8JRi0k To make this walk, or fly, or drive.. would not require much, but would not be much practical. Except if your objetive is terrorism, but not for war.
|
#
?
Dec 24, 2023 22:05
|
|
|
|
| # ? May 23, 2024 21:07 |
|
|
Imaginary Friend posted:After watching some YouTube videos about robots and how there are chatgpt-powered ones being developed a question sprung up in my mind; when will we see the first murderbots? I mean, eventually, tech-savvy people will be able to buy a robot and dunk an open source, unrestricted AI into them. Will we see a robot war á la "I, Robot" soon? Look up autonomous drones to learn more of the subject. I'm not aware of any systems that kill people in an automated way but if it hasn't happened yet it certainly will in the near future. They have clear military uses and plenty of ethical problems.
|
|
#
?
Dec 24, 2023 22:06
|
|
|
Nervous posted:What about a non self aware murder swarm with faulty/breached IFF coding that's running amok on a civilian population center?
|
|
#
?
Dec 24, 2023 22:16
|
|
|
Count Roland posted:Look up autonomous drones to learn more of the subject. I'm not aware of any systems that kill people in an automated way but if it hasn't happened yet it certainly will in the near future. They have clear military uses and plenty of ethical problems. I thought Samsung had some controversial automated gun turrets setup along the north/south Korea border. Wikipedia tells me it is the SGR-A1.
|
|
#
?
Dec 24, 2023 23:36
|
|
|
It'd be pretty trivial to have the drone find and shoot targets by itself, you don't need any fancy AI for that. Just a basic neural net to classify the potential targets (like a consumer DJI drone can track a person). Then just have it go from target to target and shoot/drop a grenade on it.
|
|
#
?
Dec 25, 2023 04:13
|
|
|
mobby_6kl posted:It'd be pretty trivial to have the drone find and shoot targets by itself, you don't need any fancy AI for that. Just a basic neural net to classify the potential targets (like a consumer DJI drone can track a person). Then just have it go from target to target and shoot/drop a grenade on it. Anything with even the most basic neural net is being lauded as AI these days. You can bet if what you describe happens there will be hand wringing about murderous AI and a Skynet apocalypse just around the corner.
|
|
#
?
Dec 29, 2023 17:58
|
|
|
Seph posted:Anything with even the most basic neural net is being lauded as AI these days. AI does not mean super inteligence. Any machine that make smart decisions automatically can be labelled AI. Even if is pretty simple.
|
|
#
?
Dec 29, 2023 21:17
|
|
|
Tei posted:AI does not mean super inteligence. Any machine that make smart decisions automatically can be labelled AI. Even if is pretty simple. Eh, no one was calling the lovely neural nets I worked on five years ago “AI” - it’s only since ChatGPT became mainstream that the term has expanded.
|
|
#
?
Dec 29, 2023 23:40
|
|
|
People refer to the monster intelligence in Doom as AI. The bar is very low.
|
|
#
?
Dec 29, 2023 23:48
|
|
|
Marketing departments in software have ruined lots of perfectly good words.
|
|
#
?
Dec 30, 2023 09:32
|
|
|
Mega Comrade posted:Marketing departments in software have ruined lots of perfectly good words. No I think the previous poster was right, we've had some weird drift from AI just meaning an automated process to make a decision, to "wizz bang fancy computer magic, give me funding". Something that's basically a flow chart is perfectly valid AI.
|
|
#
?
Dec 30, 2023 19:03
|
|
|
Bug Squash posted:No I think the previous poster was right, we've had some weird drift from AI just meaning an automated process to make a decision, to "wizz bang fancy computer magic, give me funding". Something that's basically a flow chart is perfectly valid AI. Same for AGI. Somehow instead of AGI just meaning a more general form of AI, it's now said to require learning, problem solving, reasoning, autonomy, multi-modality, emotional intelligence, ethics and even awareness.
|
|
#
?
Dec 30, 2023 19:10
|
|
|
SaTaMaS posted:Same for AGI. Somehow instead of AGI just meaning a more general form of AI, it's now said to require learning, problem solving, reasoning, autonomy, multi-modality, emotional intelligence, ethics and even awareness. I think when the lay person hears AI what they really are imagining is AGI with all the traits you described. The current hype train is capitalizing on that misconception to upsell capabilities that have existed for years.
|
|
#
?
Dec 30, 2023 20:26
|
|
|
Seph posted:I think when the lay person hears AI what they really are imagining is AGI with all the traits you described. The current hype train is capitalizing on that misconception to upsell capabilities that have existed for years. I think when the lay person hears AI they imagine literally anything in the range of "thinky computer"
|
|
#
?
Dec 30, 2023 21:53
|
|
|
reignonyourparade posted:I think when the lay person hears AI they imagine literally anything in the range of "thinky computer" you are trying to tell me that "thinking meat" exists? https://www.youtube.com/watch?v=7tScAyNaRdQ
|
|
#
?
Dec 31, 2023 01:26
|
|
|
reignonyourparade posted:I think when the lay person hears AI they imagine literally anything in the range of "thinky computer" Yeah, I think people way overestimate the degree to which your average person gives a poo poo. I’ve heard multiple people raise the objection that we shouldn’t call the current systems AI because your average person thinks that means a full general intelligence and that confusion is being exploited, but I can say from experience that when you talk with your average person about this their usual starting position is “thing I maybe heard about on the news once that I don’t care about and have no particular opinion on”. At Christmas a member of my family showed us a thing she had on her phone, which was a ChatGPT bot which conversed humorously in our local dialect. She passed it around and everyone had a laugh typing to it and reading it’s responses. As we talked about AI afterward, it was clear to me at least that none of these people, many of whom were having their first actual interaction with an AI product after hearing about it only from the news, thought they were interacting with any sort of human-level intelligence. Their conception of what was going on in that family member’s phone began and ended with “the phone does a funny thing”. Lay people don’t think AI is sentient or self-aware or generally intelligent because the term “AI” is misleading. They don’t think about AI at all.
|
|
#
?
Dec 31, 2023 11:24
|
|
|
repiv posted:well they can be if the model is over-fitted, as demonstrated by the new midjourney V6 model which seems to be especially prone to regurgitating near-perfect replicas of images from the training set for some reason Speaking of Mindjourney, as part of one of their ongoing court cases (Andersen v. Stability AI Ltd in the Northern District of CA), they filed a list of artists that they used to train their model, and it's pretty damning. Just at a glance for names I recognize, it includes pretty much every modern fantasy artist who's done a book cover since the 70's. Phil Foglio, Larry Dixon (Mercedes Lackey's husband), Larry Elmore (the OG fantasy cover master), Michael Whelan, Todd Lockwood, Julie Bell, Boris Vallejo, Frank Frazetta, Gerald Brom, Keith Parkinson; as well as pretty much every popular webcomic artist. Foglio's the one who brought it to my attention, as his response is 'Hey kids! Here's a fun game! They just dropped a list of All the artists that Midjourney admits to having scraped to train its A.I. engine. Are YOU on it? (We are) Do you have a good lawyer? (We do)'. Liquid Communism fucked around with this message at 12:50 on Jan 4, 2024 |
|
#
?
Jan 4, 2024 12:42
|
|
|
That program's gonna get filtered to hell in a handbasket in '24 and I'm looking forward to all the AI artbro meltdowns peppered throughout the year from it 
|
|
#
?
Jan 4, 2024 20:20
|
|
|
The problem the plaintiffs have is that its using their art in the first place. Filtering the front end doesn't do anything to change that, it just makes it harder to get styles reminiscent of those artists. Not impossible, though, since the data is still in there. MJ might crack down using filtering, but it's unlikely to do anything to change the outcomes of lawsuits. Neither do these chatlogs, since everyone has known from the start that they've been taking copyrighted works from the internet and tagging them. It's just PR guff from the plaintiffs and their supporters, a way to generate favourable press and get their message out. SCheeseman fucked around with this message at 01:56 on Jan 5, 2024 |
|
#
?
Jan 5, 2024 01:53
|
|
|
In the battle between AI and the creative arts, i expect the worlds of business, law, and politics to join forces and do what they've done since before i was born: marginalize and disappoint the actual creators
|
|
#
?
Jan 5, 2024 02:04
|
|
|
The fight should always been with, in a broad sense, capitalism. Copyright is a part of that framework, using it as a weapon against itself doesn't seem likely to work. I've been thinking taxing big corps for using AI generators would probably be a better approach? That way it doesn't matter where they got the data from, they get punished for using tools to replace humans regardless of the origins of their datasets.
|
|
#
?
Jan 5, 2024 02:24
|
|
|
SCheeseman posted:The problem the plaintiffs have is that its using their art in the first place. Filtering the front end doesn't do anything to change that, it just makes it harder to get styles reminiscent of those artists. Not impossible, though, since the data is still in there. Not really a PR thing, it's a document produced by StableDiffusion as part of an ongoing lawsuit. It's being widely communicated because artists and authors tend to point it out to their peers when they see their works being plagiarized or infringed on en masse.
|
|
#
?
Jan 5, 2024 05:55
|
|
|
But it doesn't really change anything. At what point did we not know they were putting copyrighted works in a machine, labeled with words relevant to the image? That's how the whole thing works. The infringement is a given, the question that needs answering is whether they're allowed to do it.
|
|
#
?
Jan 5, 2024 06:31
|
|
|
SCheeseman posted:But it doesn't really change anything. At what point did we not know they were putting copyrighted works in a machine, labeled with words relevant to the image? That's how the whole thing works. If they're allowed to do it consistently by many courts with many different judges ruling similarly, it becomes fairly definitionally not infringement. It's very possible that the usage of the material is not infringing, as the models themselves could be argued are a transformative work, even if that transformative work could spit out an image which so closely resembles the source material the new image becomes infringing. If this holds in court I could see a pretty straightforward argument to say that these services can still be legally offered if good enough safeguards are put on the output. It's not like compressing a file where the original content is actually still all there. None of the above is meant to say however that I think this usage by the ML companies is ethical, to be very clear, just that it could be legal currently. Which would be a failure of current copyright structure to do its job in the face of new technology, which wouldn't be the first time. Does anyone have any legal summaries of the issue that they like which are more straightforward than reading the court documents in ongoing cases? It'd be great to read a few different lawyers argue their position on whether or not the training aspect and model generation portions are legal under current law. I'd especially like to see legal opinions from those who don't believe the law matches what they ethically think. (as in a lawyer who is an ML supporter arguing that as the law stands now it's illegal and the opposite) Edit: This was much of what drove my thinking as well. SCheeseman posted:I see the precedent set by Google Books, which was arguably even more clear-cut in it's use of copyrighted works without permission, as an indicator of where this generative AI stuff is going to land in court. IANAL though Kagrenak fucked around with this message at 07:08 on Jan 5, 2024 |
|
#
?
Jan 5, 2024 06:52
|
|
|
I see the precedent set by Google Books, which was arguably even more clear-cut in it's use of copyrighted works without permission, as an indicator of where this generative AI stuff is going to land in court. IANAL though
|
|
#
?
Jan 5, 2024 07:04
|
|
|
Liquid Communism posted:Not really a PR thing, it's a document produced by StableDiffusion as part of an ongoing lawsuit. It's being widely communicated because artists and authors tend to point it out to their peers when they see their works being plagiarized or infringed on en masse. That was a document produced by Midjourney as part of an ongoing lawsuit, not Stability AI. Midjourney and Stability AI (Stable Diffusion creators) are two different companies. I had to track down the amended lawsuit because that wasn't adding up. Stability AI partnered with another company Spawning AI to create a DO NOT TRAIN opt out list at http://haveibeentrained.com/ over a year ago and have been following that in all the models they released after. This document is the first I've seen of anything that Midjourney has used outside of Laion 5B. The fact that Image Generative AI Models are trained on images with text pairs is not surprising or new, that's the entire point. They don't have to be trained on copyrighted material and there are a few models out there now with completely clear "bills of health" so to say, entirely public domain. These lawsuits against these specific companies make headlines but even if these companies were to shut down shop today, the knowledge and means to continue would still exist. I guess I need to be the one to point out that you are allowed to infringe copyright if your use is Fair Use. That's what Fair Use means. That your use serves the Public Good more than upholding a monopoly does. That's what copyright is, a monopoly. It's been wild watching staunch leftists do a complete 180 and suddenly take up arms to protect capitalism. Wild. SCheeseman posted:The fight should always been with, in a broad sense, capitalism. Copyright is a part of that framework, using it as a weapon against itself doesn't seem likely to work. I like the idea of an Automation Tax, especially as part of a sweeping comprehensive tax plan. Tax the billionaires until there are no more billionaires. $2,000 a month UBI for everyone will actually solve most peoples problems and that is how you can pay for it. Kagrenak posted:It's very possible that the usage of the material is not infringing, as the models themselves could be argued are a transformative work, even if that transformative work could spit out an image which so closely resembles the source material the new image becomes infringing. If this holds in court I could see a pretty straightforward argument to say that these services can still be legally offered if good enough safeguards are put on the output. It's not like compressing a file where the original content is actually still all there. I'm not a lawyer, I'm just following this very closely because this means an incredible amount to me being able to create again after breaking down physically. To that end, I've been looking at additional tools and one of them actually allows you to make a heatmap out of the image based on the specific keywords used. Diffusion Attentive Attribution Maps or DAAM Scripts. I made a long detailed post about this in the GBS AI Art Thread all the way forever ago back in the ancient days of... June: https://forums.somethingawful.com/showthread.php?noseen=1&threadid=4000251&pagenumber=315&perpage=40#post532452282 Two main images were made: "gnarled tree" and "gnarled tree drawn by leonardo donatello michangelo raphael". You can clearly see the difference in using the four artists names on the output of the image.     Here's the thing that almost everyone is going to skip over. The four artists name HAD NO BEARING ON THE ORIGINAL IMAGE.   I'll summarize this incredibly important part that will likely be skipped over by most people: it doesn't matter what the model was trained on, only what you specifically attempt to make each time. This is A, Big, Deal. This means that even if an artists work was used in training an Image Generation AI Model like Stable Diffusion, it doesn't mean their work influences all of the resulting outputs, only when the bits of specific art from that artist are called is it actually used in the generation process. Clearly this is a transformative process.
|
|
#
?
Jan 5, 2024 08:06
|
|
|
KwegiboHB posted:I guess I need to be the one to point out that you are allowed to infringe copyright if your use is Fair Use. That's what Fair Use means. That your use serves the Public Good more than upholding a monopoly does. That's what copyright is, a monopoly. Kinda like how the whole Open Source movements is build on top of the Copyright laws. Because is the current legal framework. So you must build anything based on the foundations you have. You can't choose your society*, but you can build whatever based on our society already work. Here for artist, our society does not have the tools to support creative people. Many artist starve, or have to have a second job so they pay for the creative process. The only tool we provide is Copyright, so artist have to use the Copyright laws. I am curious how the legal system will react to this. Maybe is there something else in the law that can be used to stop this? I wish things like midjourney where made ethical, because I want to use tools like that, but I don't want to turbofuck artists, I want to support their lives so they create new cool stuff 
|
|
#
?
Jan 5, 2024 09:10
|
|
|
KwegiboHB posted:
This take was just as dumb when crypto bros were using it. Copyright is a crap system but it's all artists have to protect their work and scrape a living.
|
|
#
?
Jan 5, 2024 09:25
|
|
|
Tei posted:Kinda like how the whole Open Source movements is build on top of the Copyright laws. Because is the current legal framework. So you must build anything based on the foundations you have. You can't choose your society*, but you can build whatever based on our society already work. https://huggingface.co/Mitsua/mitsua-diffusion-one Here is the Stable Diffusion style model trained entirely on public domain and CC0 images. You can download that model and use it in any standard Stable Diffusion front end like AUTOMATIC1111 or ComfyUI. Knock yourself out without losing sleep over "ethics in AI training datasets". I'll say again, and again, and again, and again and even more if I have to. Universal Basic Income isn't just the answer, it's the solution. An Answer will temporarily fix a problem while still keeping the problem, a Solution will mean there's no more problem. Tax the ever living hell out of the rich and give yourself your money back. Watch your problems disappear.
|
|
#
?
Jan 5, 2024 09:27
|
|
|
Great and if all those companies making profit off copyrighted material in their models switch to public domain sources all those law suits can be dropped.
|
|
#
?
Jan 5, 2024 09:31
|
|
|
Mega Comrade posted:This take was just as dumb when crypto bros were using it. The reason you can only "scrape" a living is because billionaires have already stolen all of your money. Tax the poo poo out of them. Steal it back. Mega Comrade posted:Great and if all those companies making profit off copyrighted material in their models switch to public domain sources all those law suits can be dropped. That's a great sentiment except I've read these lawsuits and the law and there's greater than even odds that what these companies are doing is legal. What then?
|
|
#
?
Jan 5, 2024 09:36
|
|
|
I'm not sure if it's been discussed but the NYT lawsuit is pretty interesting and seems stronger than a lot of others. They have over 100 examples in it of verbatim NYT articles being spat out and also examples of hallucinations that the model then claims is from the NYT, so you have copyright and damage of reputation. Maybe Google books will cover this but as some of them are such a large amount of the articles it also might not. https://nytco-assets.nytimes.com/2023/12/Lawsuit-Document-dkt-1-68-Ex-J.pdf KwegiboHB posted:That's a great sentiment except I've read these lawsuits and the law and there's greater than even odds that what these companies are doing is legal. What then? Then nothing, they make millions, artists get screwed over again, you get to play with your toys. KwegiboHB posted:The reason you can only "scrape" a living is because billionaires have already stolen all of your money. Tax the poo poo out of them. Steal it back. When that starts happening you can dismantle copyright all you like and I won't complain but you don't remove a system, regardless of how poor until you have a replacement. Mega Comrade fucked around with this message at 10:22 on Jan 5, 2024 |
|
#
?
Jan 5, 2024 09:41
|
|
|
Mega Comrade posted:Then nothing, they make millions, artists get screwed over again, you get to play with your toys. But everyone wants the copyright win, it's priority number one. A win is a killer blow, they'll have no data to train on! Ha Ha Ha! Well, except for all the IP they already own, the IP they have the ability to pay for and a monopoly that will allow them to grossly profit from it all, which lets them buy more IP. SCheeseman fucked around with this message at 10:56 on Jan 5, 2024 |
|
#
?
Jan 5, 2024 10:51
|
|
|
I'm not missing at all. I'm fully aware copyright winning isn't going to magically make life better for artists. But I also don't see why i should be rooting for Microsoft.
|
|
#
?
Jan 5, 2024 11:25
|
|
|
KwegiboHB posted:The reason you can only "scrape" a living is because billionaires have already stolen all of your money. Tax the poo poo out of them. Steal it back. Then we have 3 to 5 years of new content before permanent stagnation until running generative AI becomes too costly to maintain for the value it provides. People will further slow down on coming into the visual art industry if doing so only profits others and doesn't have a sustainable path to being a livable career. Recall the whole driving force behind the existence of copyright was authors losing the ability to make a living on their work because infringing printers could produce knockoffs fast enough to beat them to market in more distant areas. Capitalism is a terrible system, but as it's not going away any time soon forcing companies to play within the scope and spirit of the laws is the minimum necessary level of regulation. SCheeseman posted:There's other paths. You're hinging your hopes on copyright saving the day all while there's been an upswell of labor organization in the US. You're missing. in spite of it coming up repeatedly, that even if the lawsuit goes against a fair use ruling, that doesn't save anyone's jobs. The tools are still gonna be made, but with licensed datasets. The core problem of loss of careers and jobs remains. Nobody gives a poo poo if they train on their own IP because generative AI is, as noted exhaustively, not actually all that creative. Especially without a broad training set to take ideas outside of an existing IP from. The costs to run it at macro scale are not insignificant, and without being able to feed it the entire corpus of the internet for training data it isn't going to be the huge success they need to justify that expense long-term.
|
|
#
?
Jan 5, 2024 11:27
|
|
|
Liquid Communism posted:Nobody gives a poo poo if they train on their own IP because generative AI is, as noted exhaustively, not actually all that creative. Especially without a broad training set to take ideas outside of an existing IP from. The costs to run it at macro scale are not insignificant, and without being able to feed it the entire corpus of the internet for training data it isn't going to be the huge success they need to justify that expense long-term. It's nuts to me that they don't give a poo poo. The technology is the threat to peoples livelihoods by the mere fact that it exists, whether any particular application of it uses copyrighted works or not does little to change the outcome of people losing their jobs because AI fills a content creation role that doesn't require creativity, which is a lot of jobs. A good image generator may not need as broad a training set as you think it does to serve this purpose, Adobe Firefly is indicative of this. Regarding power usage, they're using a technological pipeline created to make graphics go brrr to run this stuff, if it eventually becomes uneconomical they'll find inefficiencies to optimize so it becomes economical again. Though these days, you can run a heavily optimized LLM on Raspberry Pis, I wouldn't doubt there's plenty of overhead with training too. You're placing your bets based on assumptions that don't even hold up today. SCheeseman fucked around with this message at 12:22 on Jan 5, 2024 |
|
#
?
Jan 5, 2024 12:17
|
|
|
Liquid Communism posted:Then we have 3 to 5 years of new content before permanent stagnation until running generative AI becomes too costly to maintain for the value it provides. People will further slow down on coming into the visual art industry if doing so only profits others and doesn't have a sustainable path to being a livable career. Recall the whole driving force behind the existence of copyright was authors losing the ability to make a living on their work because infringing printers could produce knockoffs fast enough to beat them to market in more distant areas. Where are you getting any of these ideas? What does permanent stagnation mean? Are we going to run out of Art? I wasn't aware Art was anything less than Infinite in scope and breadth... Everything I've seen about Generative AI has only lead me to believe that we are currently in the Dial-Up phase and the real Broadband explosion is still yet to come. Power needs for Generation or Inference are going down as more is learned about the process and actual requirements, especially with new hardware being planned to make better use of Matrix Multiplication. Already barely a year in and we approach real-time image generation. This says nothing about about Text Generation, or Video, or Audio, or even Singing. To think this is going to slow down now instead of continue to snowball seems ludicrous to me and again I'd like to know why you think this will happen. https://www.bls.gov/ooh/arts-and-design/home.htm U.S. Bureau of Labor Statistics estimates most artists jobs around the average of all growth rates at 3%. Except poor Floral Designers  I don't know where to get better numbers or statistics on overall jobs or expected growth/loss rates so if you have a better source please let me know.As for Copyright, the "driving force" for the creation of Copyright was the https://en.wikipedia.org/wiki/Licensing_of_the_Press_Act_1662 expiring and the https://en.wikipedia.org/wiki/Worshipful_Company_of_Stationers_and_Newspaper_Makers no longer having monopoly rights to censor and charge for their published works. This censorship has always been deeply unpopular and prevented the law from being renewed multiple times until the Stationers changed tacts and presented their arguments as "helping" the Authors. "Learned men will be wholly discouraged from Propagating the most useful Parts of Knowledge and Literature", a popular lie that continues to be told today. The Stationers knew that all the Authors could really do is sign over their rights to the Stationers giving them their monopoly control back. This lead to the https://en.wikipedia.org/wiki/Statute_of_Anne or the first "Copyright" law. An important note on this was the strictly limited time controls which are a far cry from todays. 14 years plus one additional 14 years if the Author was still alive. To give some perspective, that would put Nintendos Super Mario Brothers in the Public Domain now. Further perspective, only last year was the first Synched Audio Movie put into the Public Domain, 1927's https://en.wikipedia.org/wiki/The_Jazz_Singer. The one upside to our current Near Forever Copyright laws is the Fair Use exemption. Copyright is already hard enough to understand and explain that I don't blame people for not knowing that it's both OK and Good for you to Infringe upon someone elses Copyright if your Use actually "Promotes the Progress of Science and useful Arts". This is legal! You HAVE to infringe upon someone elses copyright to invoke Fair Use in the first place! If your case advances the Public Good you shouldn't hold yourself and the rest of us back. I just reposted my example from June that Generative AI is a truly Transformative Process and not direct copying. If these companies you're focusing on already ARE "playing within the scope and spirit of the laws", then, much as you may hate to hear it, the "minimum necessary level of regulation" has already been met. It's important to understand who and what you are fighting for in this debate. Do you understand why you are arguing for more monopoly control and less public domain rights? Especially since fighting for Universal Basic Income instead would directly fix the problems? Wild I say... Wild. I particularly enjoy the freedom of my press, both figurative and literal. I'm going to continue to take steps to use and secure it. While I'm at it, what metric are you using to judge creativity in AI outputs? I would love to read these "exhaustive notes". Mega Comrade posted:I'm not missing at all. I'm fully aware copyright winning isn't going to magically make life better for artists. But I also don't see why i should be rooting for Microsoft. So don't root for Microsoft, instead talk about things like antitrust laws. If you're worried about something like unrestrained data scraping and bots being trained on things wholesale then make sure you link words in your sentences that say things you actually do want, like Microsoft and antitrust lawsuits and forcibly being broken up into smaller companies. This increases the chances the bot would "learn" those words belong together and output them at a later date! While the effect on a bot would be relatively tiny compared to the entire internets worth of data, it does have the bonus effect of stirring useful discussion amongst the humans that read it. Plus it feels better to talk about the things you actually want to happen rather than talk about the fears of things you don't want to happen. That's why I mention Universal Basic Income, because I actually want that to happen.
|
|
#
?
Jan 5, 2024 18:42
|
|
|
What are you on about? Are you seriously saying we should write in such a way to try and influence further models? I know your understanding of this technology has been shown to be lacking but your trolling at this point surely?
|
|
#
?
Jan 5, 2024 18:59
|
|
|
Mega Comrade posted:Are you seriously saying we should write in such a way to try and influence further models? Well it's more of trying to influence other human beings. That used to be called just... Talking.
|
|
#
?
Jan 5, 2024 19:06
|
|
|
|
| # ? May 23, 2024 21:07 |
|
|
Mega Comrade posted:What are you on about? For context this is the person who thinks he has a sentient consciousness chained up inside his computer that he can talk to
|
|
#
?
Jan 5, 2024 19:06
|
|