Disagree. The technology will never yield AGI as all it does is remix a huge field of data without even knowing what that data functionally says.
All it can do now and ever will do is destroy the environment by using oodles of energy, just so some fucker can generate a boring big titty goth pinup with weird hands and weirder feet. Feeding it exponentially more energy will do what? Reduce the amount of fingers and the foot weirdness? Great. That is so worth squandering our dwindling resources to.
I really don't know whats going about the Anti-AI people. But is getting pretty similar to any other negationism, anti-science, anti-progress... Completely irrational and radicalized.
Disagree. The technology will never yield AGI as all it does is remix a huge field of data without even knowing what that data functionally says.
We definitely don't need AGI for AI technologies to be useful. AI, particularly reinforcement learning, is great for teaching robots to do complex tasks for example. LLMs have shocking ability relative to other approaches (if limited compared to humans) to generalize to "nearby but different, enough" tasks. And once they're trained (and possibly quantized), they (LLMs and reinforcement learning policies) don't require that much more power to implement compared to traditional algorithms. So IMO, the question should be "is it worthwhile to spend the energy to train X thing?" Unfortunately, the capitalists have been the ones answering that question because they can do so at our expense.
For a person without access to big computing resources (me lol), there's also the fact that transfer learning is possible for both LLMs and reinforcement learning. Easiest way to explain transfer learning is this: imagine that I want to learn Engineering, Physics, Chemistry, and Computer Science. What should I learn first so that each subject is easy for me to pick up? My answer would be Math. So in AI speak, if we spend a ton of energy to train an AI to do math and then fine-tune agents to do Physics, Engineering, etc., we can avoid training all the agents from scratch. Fine-tuning can typically be done on "normal" computers with FOSS tools.
all it does is remix a huge field of data without even knowing what that data functionally says.
IMO that can be an incredibly useful approach for solving problems whose dynamics are too complex to reasonably model, with the understanding that the obtained solution is a crude approximation to the underlying dynamics.
IMO I'm waiting for the bubble to burst so that AI can be just another tool in my engineering toolkit instead of the capitalists' newest plaything.
Sorry about the essay, but I really think that AI tools have a huge potential to make life better for us all, but obviously a much greater potential for capitalists to destroy us all so long as we don't understand these tools and use them against the powerful.
This has been going on since big oil popularized the "carbon footprint". They want us arguing with each other about how useful crypto/AI/whatever are instead of agreeing about pigouvian energy taxes and socialized control of the (already monopolized) grid.
Pollock's paintings are art. a bunch of paint buckets falling on a canvas in an earthquake wouldn't make art, even if it resembled Pollock's paintings. there's no intent behind it. no artist.
Any work made to convey a concept and/or emotion can be art. I'd throw in "intent", having "deeper meaning", and the context of its creation to distinguish between an accounting spreadsheet and art.
The problem with AI "art" is it's produced by something that isn't sentient and is incapable of original thought. AI doesn't understand intent, context, emotion, or even the most basic concepts behind the prompt or the end result. Its "art" is merely a mashup of ideas stolen from countless works of actual, original art run through an esoteric logic network.
AI can serve as a tool to create art of course, but the further removed from the process a human is the less the end result can truly be considered "art".
As a thought experiment let's say an artist takes a photo of a sunset. Then the artist uses AI to generate a sunset and AI happens to generate the exact same photo. The artist then releases one of the two images with the title "this may or may not be made by AI". Is the released image art or not?
If you say the image isn't art, what if it's revealed that it's the photo the artist took? Does is magically turn into art because it's not made by AI? If not does it mean when people "make art" it's not art?
If you say the image is art, what if it's revealed it's made by AI? Does it magically stop being art or does it become less artistic after the fact? Where does value go?
The way I see it is that you're trying to gatekeep art by arbitrarily claiming AI art isn't real art. I think since we're the ones assigning a meaning to art, how it is created doesn't matter. After all if you're the artist taking the photo isn't the original art piece just the natural occurrence of the sun setting. Nobody created it, there is no artistic intention there, it simply exists and we consider it art.
there's something's highly suspect about someone not understanding the difference between art made by a human being and some output spit out by a dumb pixel mixer. huge red flag imo.
and yes, the value does go. because we care about origin and intent. that's the whole point.
if the original Mona Lisa were to be sold for millions of dollars, and then someone reveals that it was not the original Mona Lisa but a replica made last week by some dude... do you think the buyer would just go "eh it looks close enough"? no they would sue the fuck out of the seller and guess what, the painting would not be worth millions anymore. it's the same painting. the value is changed. ART IS NOT A PRODUCT.
there's something's highly suspect about someone not understanding the difference between art made by a human being and some output spit out by a dumb pixel mixer. huge red flag imo.
Translation. I can't argue your point so I'm going to try characters assassination.
if the original Mona Lisa were to be sold for millions of dollars, and then someone reveals that it was not the original Mona Lisa but a replica made last week by some dude... do you think the buyer would just go "eh it looks close enough"? no they would sue the fuck out of the seller and guess what, the painting would not be worth millions anymore. it's the same painting. the value is changed. ART IS NOT A PRODUCT.
Pretty ironic to say art is not a product and then argue that its monetary value would decrease, which can happen only if you treat art as a product.
Imagine if instead of a physical painting Mona Lisa was a digital file and free on the internet, would people think Mona Lisa is less impressive as an art piece because anyone could own it? I think it's artistic value wouldn't decrease, only its value as a product would decrease because everyone could get it for free.
AI is a tool used by a human. The human using the tools has an intention, wants to create something with it.
It's exactly the same as painting digital art. But instead o moving the mouse around, or copying other images into a collage, you use the AI tool, which can be pretty complex to use to create something beautiful.
Do you know what generative art is? It existed before AI. Surely with your gatekeeping you think that's also no art.
I'm so sick of this. there are scenarios in which so-called "AI" can be used as a tool. for example, resampling. it's dodgy, but whatever, let's say the tech is perfected and it truly analyzes data to give a good result rather than stealing other art to match.
but a tool is something that does exactly what you intend for it to do. you can't say 100 dice are collectively "a tool that outputs 600" because you can sit there and roll them for as long as it takes for all of them to turn up sixes, technically. and if you do call it that, that's still a shitty tool, and you did nothing worth crediting to get 600. a robot can do it. and it does. and that makes it not art.
So do you not what generative art is. And you pretend to stablish catedra on art.
Generative art, that existed before even computers, is s form of art in which a algorithm created a form of art, and that algorithm can be repeated easily. Humans can replicate that algorithm, but computers can too, and generative art is mostly used with computers because obvious reasons. Those generative algorithms can be deterministic or non deterministic.
And all this before AI, way before.
AI on its essence is just a really complex and large generative algorithm, that some people do not understand and this are afraid of it, like people used to be afraid of eclipses.
Also, you seems not to know that photographs also take hundreds or thousands of pictures with just pressing a button and just select the good ones.
cameras do not make random images. you know exactly what you're getting with a photograph. the reason you take multiples is mostly for timing and lighting. also, rolling a hundred dice is not the same as painting something 100 times and picking the best one, nor is it like photographing it. the fact that you're even making this comparison is insane.
If you know how to use an AI you also know how it's working and what are you going to get, is not random. It's a complex generative algorithm where you put in the initial variables, nothing more.
People said exact same thing about CGI, and photography before. I wouldn't be surprised if somebody scream "IT'S NOT ART" at Michaelangelo or people carving walls of temples in ancient Egypt.
Technology is a cultural creation, not a magic box outside of its circumstances. "The problem isn't the technology, it's the creators, users, and perpetuators" is tautological.
And, importantly, the purpose of a system is what it does.
So you're using a different specific and niche technology (which directly benefits and exists because of) the technology that is the subject of critique, and acting like the subject technology behaves like yours?
"Google is doing a bad with z"
"z can't be bad, I use y and it doesn't have those problems that are already things that happened. In the past. Unchangeable by future actions."
No one wants to ban technology outright. What we're saying is that the big LLMs are actively harmful to us, humanity. This is not fear mongering. This is just what's happening. OpenAI, Google, Microsoft, and Meta are stealing from humanity at large and setting the planet on fire to do it. For years they told us stealing intellectual property on an individual level was a harmful form of theft. Now they're doing the same kind of theft bit its different now because it benefits them instead of us.
What we are arguing is that this is bad. Its especially extra bad because with the death of big search a piece of critical infrastructure to the internet as we know it is now just simply broken. The open source wonks you celebrate are working on fixing this. But just because someone criticizes big tech does not mean they criticize all tech. The truth is the FAANG companies plus OpenAI and Microsoft are killing our planet for it to only benefit their biggest shareholders
I did not believe in Intelectual Property before. I'm not going to start believing now.
The same I think that corporations having a hold on media is bad for humandkind I think that small artists should not have a "not usable by AI"hold on what they post. Sharing knowledge is good for humanity. Limitate who can have access or how they can use that knowledge or culture is bad.
The dead of internet have nothing to do with AI and all to do with leaving internet in hands of a couple big corporations.
As for emissions.. are insignificant relative to other sources of CO2 emissions. Do you happen to eat meat, travel abroad for tourism, watch sports, take you car to work, buy products made overseas? Those are much bigger sources of CO2.
Technology is a product of science. The facts science seeks to uncover are fundamental universal truths that aren't subject to human folly. Only how we use that knowledge is subject to human folly. I don't think open source or open weights models are a bad usage of that knowledge. Some of the things corporations do are bad or exploitative uses of that knowledge.
They have infinite resources. They're making gestures to dissuade normies. I suspect this will get them most of the result they want. They're also wasting time, effort and resources of adblock programmers (and that is a far more limited resource).
Sure but as long as there is a least one dedicated bearded dude hidden in a dark underground room behind his screen, they will be defeated. No matter how much they spent on the new technology. What I mean is that devs might burn out, they will still be replaced by others. And we get such people faster than youtube is able to burn them out
nah, if they embed the ad into the video stream (they were testing this for some users!), the only adblocking option will be to blank out the screen and wait through the ad (or download the video in advance and edit the ad out automatically), both of which would make it a lot more annoying to adblock than currently.
nope, the ad time varies unlike a sponsor segment, and also youtube would not let you skip through an ad while streaming it, whereas sponsors you can, hence the download and edit out with LLM or whatever algorithm works best
yes, but if youtube only serves you the real video chunks after your client plays through the ad chunks (all in the same media stream to the client), theres gonna be some waiting involved, not like adblocking today where it is instant.
you can skip through sponsor segments, but these are ads from youtube, not from the creator, and youtube will not let you conveniently skip through the ads. if implemented correctly, youtube could ensure that the ad is fully played, which would need downloading and automatic editing to counter.
How would they determine if the ad is played without trusting the client? I guess they could screw with the buffer, but that would really piss of people with poor internet, and most people would prefer an ad-length black screen to whatever attention wrenching dark pattern manipulative brown noise wants to infect your mind today.
That is an interesting question. From what I know, youtube has every video in chunks that they serve to the client, and so server side ad injection is just serving some ad chunks before the video. I think you're right with the buffer thing, it seems to me like the only way to make sure the client can't skip it would be to make the buffer shorter, impacting some people (although seems like only really people with internet thats fast enough for streaming some seconds, but not other seconds, which is an odd catagory)
Ultimately it would be a tradeoff for youtube, but the fact that they put the effort into doing mass testing of the idea at all shows that clearly there are some good incentives, and it may eventually be implemented.
Don't worry, I have been using the powerful technology of "the mute button" and "doing other stuff" since before cable TV existed. We always have alternatives.
This is actually one of very few valid cases for an LLM, to help sponsorblock determine ad segments by analyzing the word choice and speech patterns in segments of the video.
"Stop worrying what other people think" is bad advice anyway.
It's not like that's a possibility for people with anxiety.
Give concrete advice: self affirmation, reflective writing, journaling positive feedback....
It's also bad advice in general? The moment you stop caring what others think you become a terrible member of society. You'll go around smelly, haggard, rude, in your pajamas and absolutely fail in life because you "don't care what other people think"
Someone needs to tell google that AI powered search is not working right now, and that they better wait a few years to try massively implementing that in a successful way.
Other AI fields are working really good. But search engine "instant AI answers" for general use are not in a phase when they should be as widely used as google (or microsoft) is trying to use them right now.
In what sense does a small community working with open weight (note: rarely if ever open source) llm have any mitigating impact on the rampant carbon emissions for the sake of bullshit generators?
It's times like these that online advertisements need to get creative to get ahead in this never ending adblocking arms race, just like the very subtle advertisement in the car chase scene in the Academy Award nominated film, "Barbie", now available on Blu-ray and select streaming services.
I know right? The other day I was drinking a coke and wondering about side effects of weight loss drugs such as Ozempic, and it occurred to me that advertising could be a lot more creative and subtle.
In my country this kind of AI is being used to more efficiently find tax fraud and to create chatbots for users to understand taxes, that due to the much more reliable and limited training set does not allucinate and can provide clear sources for the information given.
I'm one of the dozen people that bought premium to not have to deal with it. I'm just patiently waiting for alternatives to become more viable so I can jump ship entirely. YouTube is the last remaining Google service I still use.
The root problem is capitalism though, if it wasn't AI it would be some other idiotic scheme like cryptocurrency that would be wasting energy instead. The problem is with the system as opposed to technology.
If you think that sounds like "Žižekian nonsense", then you obviously don't understand what Žižek argues, because he clearly doesn't say anything silly like "human ideology" (or "Žižekianism", for that matter). The article you posted also does wonders completely breaking down Žižek as an abonimable human being - while not truly engaging with his ideas. It is pretty worthless, takes things deliberately out of context, and, after rigorously defining him as a persona non grata, invests no proper effort to do what actual communists like Marx and Lenin did - acknowledge that even enemies like that can give contributions to understanding, and things to learn from and work at doing so.
Does he sometimes spew bullshit? Absolutely. Does he believe in "human ideology" or spout anticommunism on a worse level than The Black Book of Communism, as the article wants to imply? Only if you deliberately misread and misinterpret him.
Yeah, look, I did read the article, and the article, unlike the person who might very well have done that in their work, did not do that. All I see is the same flipping of materialist analysis into an ideological dogma, that becomes ahistoric, trying to repeat instead of following material developments towards communism. From a quick look at your links, there's even a lot I agree with, especially in criticising the French intellectuals. It still reads like a polemic removed from reality, that values its own farts more than understanding and working towards change, but it has value. And the article you linked in the beginning does nothing, but try to opportunistically recruit people away from one ideologue (which Zizek can definitiely be called) to another idealist "team" that tries to redirect proletarian material interests and analysis. You seem to think it's a contest of who can quote "great people" the best and who can be the most orthodox, which treats it all like a religion instead of a material movement to change the world and mode of production.
In the end, I fear, we will be on other sides of the river, each seeing "their idealist perversions" across from "our materialist analysis", but I at least won't cross the river for your side any time soon.
Nice burn, even brought in the "libertarian", at least be consistent, if I am a Zizekian heretic, I'm not an individualist libertarian who's afraid of authority, I am of course a liberal anticommunist reactionary who won't acknowledge the achievements of "really existing socialism". You strike me as someone who would have written a hit piece on Marx for profiting from British imperialism and his capitalist buddy Engels, citing the letter and his drinking habits to make clear that he is an immature mind, then join some utopian socialist fringe group.
Nah, human ideology is much broader than a single economic system. The fact that people who live under capitalism can't understand this just shows the power of indoctrination.
What you're saying is that you're not self aware enough to realize that you have an ideology. Everyone has a world view that they develop to understand how the world works, and every world view necessarily represents a simplification of reality. Forming abstractions is how our minds deal with complexity.
Do you think people should be treated with respect? Do you think there should be consideration for your condition so you are not exempt from certain events, activities, opportunities?
These are matters of ideology. If you say yes to it, it is ideological in the same way when you say no to it. There is no inherent objective truth to these value questions.
Same for the economy. It doesn't matter if you think that growth should be the main objective, or that equal opportunity should be the focus or sustainability or other things. You will have to make a value judgement and the sum of these values represent your ideology.
Observations may be objective, but the values are always subjective. Two different people can look at the same set of facts and come to entirely different conclusions of what constitutes desirable actions based on their world view.
You entirely missed the point of what I said. Two different people can agree on an objective fact that a table is a table, but disagree on whether it's a good looking table.
Right, but the technology has the system’s philosophy baked into it. All inventions encourage a certain way of seeing the world. It’s not a coincidence that agriculture yields land ownership, mass production yields wage labor, or in this case fuzzy plagiarism machines yield a transhuman death cult.
Youtube's crackdown on adblocker usage has created a huge Streisand effect. There's a lot of people all over the world that would've never known about adblockers if youtube just didn't say anything and didn't do exactly what they did.
Now, all the major adblockers and ad-blocking browsers have stepped their game up and made it so people can still block ads on youtube.
There's userscripts you can get from greasyfork...or is it greaseyfork that allow you to bypass all of youtube's bullshit.
100% of the videos that actually belong on youtube are almost always demonetized anyway.
If you're not constantly getting demonetized on every video you put out, despite bending over backwards to follow the rules, you're doing something wrong.
Not to mention that Vanced-type apps and content mirrors are still going strong, and proper alternative platforms like Grayjay and Nebula are getting more attractive.
I wonder how soon self-hosted video distribution will be feasible. Does ActivityPub support that yet?
I think Peertube fits that description: "PeerTube is a self-hosted ActivityPub-federated video streaming platform using P2P directly in your web browser."
Not OP but in my personal experience I've talked to countless people that had no idea watching YouTube without ads is possible (without paying) and their minds were blown. With Google calling out attention to it like this, I'm sure many will get curious. Most people don't use their electronics beyond basic functionality..
And tbh you're gonna be hard pressed to get real data of how many ppl are suddenly finding out about it.
Personally I think AI systems will kill us dead simply by having no idea what to do, dodgy old coots thinking machines are magic and know everything when in reality machines can barely approximate what we tell them to do and base their information on this terrible approximation.
Machines will do exactly what you tell them to do and is the cause of many software bugs. That’s kind of the problem, no matter how elegant the algorithm, fuzzy goes in, fuzzy comes out. It was clear this very basic principle was not even considered when Google started telling people to eat rocks and glue. You can’t patch special cases out when they are so poorly understood.
The relative number here might be more useful as long as it's understood that Google already has significant emissions. It's also sufficient to convey that they're headed in the wrong direction relative to their goal of net zero. A number like 14.3 million tCO₂e isn't as clear IMO.
Memes
Aktiv
Dieses Magazin erhält keine Updates (letzte Aktivität vor 0 Tage(n)).