|
Snowglobe of Doom posted:Netflix's latest true crime documentary used fake AI-generated photos of the killer to show what a funloving person she was: https://futurism.com/the-byte/netflix-what-jennifer-did-true-crime-ai-photos You can fix that poo poo if you know what you're doing, and it only takes a couple minutes longer. The ability to pres butan, receive image, has made most people super lazy. "gently caress it, good enough, nobody will notice" is now the norm
|
#
?
Apr 19, 2024 07:00
|
|
|
|
| # ? Jun 3, 2024 08:00 |
|
|
lmao making AI pictures for a loving documentary, un-loving-realKinkyJohn posted:You can fix that poo poo if you know what you're doing, and it only takes a couple minutes longer. The ability to pres butan, receive image, has made most people super lazy. "gently caress it, good enough, nobody will notice" is now the norm lol who gives a poo poo about that, the point is that people in the entertainment industry are now using AI to make up facts for documentaries
|
|
#
?
Apr 19, 2024 07:09
|
|
|
Snowglobe of Doom posted:Netflix's latest true crime documentary used fake AI-generated photos of the killer to show what a funloving person she was: https://futurism.com/the-byte/netflix-what-jennifer-did-true-crime-ai-photos Jfc, I was always uneasy about these trial by extremely emotive documentary things, but at least trusted they were probably responsibly researched by actual journalists Like it's not just true crime trash, they churn these out on actual geopolitics
|
|
#
?
Apr 19, 2024 07:50
|
|
|
Duck and Cover posted:That didn't take long at all. I think my favorite one was “using ai to fake things in a documentary is good because it opens a dialogue about whether or not using ai to fake things in a documentary is good”
|
|
#
?
Apr 19, 2024 12:22
|
|
|
Duck and Cover posted:That didn't take long at all. Unsurprisingly, that statement was also produced by pressing a button
|
|
#
?
Apr 19, 2024 12:49
|
|
|
Strategic Tea posted:Jfc, I was always uneasy about these trial by extremely emotive documentary things, but at least trusted they were probably responsibly researched by actual journalists I've honestly never been totally comfortable with the entire concept of documentaries in general and I hate that we've kind of all just accepted "specific argument" docs and "pure emotional manipulation" docs (lots of overlap there) as legit subgenres (even when made by amateurs!) without controversy. Snowglobe of Doom posted:Netflix's latest true crime documentary used fake AI-generated photos of the killer to show what a funloving person she was: https://futurism.com/the-byte/netflix-what-jennifer-did-true-crime-ai-photos I have to believe these folks would know better than to photoshop such pictures from scratch, why the hell would they think it'd be any different to do it with AI?
|
|
#
?
Apr 19, 2024 13:06
|
|
|
I edit documentaries and I can say I have no idea how that could've happened without insane, insane gently caress ups higher up than the level of the editor/director who would've put the image in the cut at first. You have production people and lawyers for the production company AND lawyers for the platform/broadcaster whose sole role on the project is to tally up every second of every image or piece of music used that wasn't shot by us/composed for the film, identify the rights holder of every one and clear it with a contract (and often payment, if it's a stock footage/historical archive or music library etc.). It's not a simple process, depending on the film. Multiple people whose job that is had to ask and be told that those images were fake. Incredible. The only other explanation is the people in the edit (editor/director) for some reason thought faking pictures was better to show bosses in a rough cut than putting in subpar real pics or even leaving a black screen reading "STILLS OF KILLER TO COME (VIA SECOND COUSIN)" or whatever researchers have trawled up, and then somehow those images ended up in the final cut without being substituted out. Hard to believe it as an accident - as I say, rough cuts often have gaps you need to find images for if you're doing a historical story or whatever - but even harder to believe multiple lawyers, execs etc knowingly approved the fakes.
|
|
#
?
Apr 19, 2024 14:28
|
|
|
Yeah, makes me wonder if they had too few photos on the topic in general and decided to just say 'gently caress it, AI us up some'. I'd be willing to bet there were other images in the documentary that just haven't been caught out yet because they're just good enough to escape detection.
|
|
#
?
Apr 19, 2024 14:45
|
|
|
they'll reclassify it as one of those dramatized, partially animated docs and pretend like everything is cool.
|
|
#
?
Apr 19, 2024 15:07
|
|
|
Dokapon Findom posted:Unsurprisingly, that statement was also produced by pressing a button quote:"Oh, look at you, Captain Obvious, with your superhuman ability to detect AI-generated content. What's next, predicting rain when the sky's gray? Bravo, Sherlock! 🙄"
|
|
#
?
Apr 19, 2024 18:15
|
|
|
|
|
#
?
Apr 19, 2024 22:03
|
|
|
https://twitter.com/Kotaku/status/1781006804576801089
|
|
#
?
Apr 21, 2024 15:05
|
|
|
alright folks, markets had a good run, but it's time to put them in hospice care
|
|
#
?
Apr 21, 2024 15:41
|
|
|
Ahh there it is, that proper real cyberpunk dystopia poo poo. Workin in the ai porn mines and being paid in cosmetic skins.
|
|
#
?
Apr 21, 2024 16:02
|
|
|
Reading the article it seems as though they're basically renting GPU time on people's computers. How valuable is AI generated porn anyways? I mean, the more you make the more worthless it is and people can just make it themselves with a mediocre compuiter. Reading further, they're lending the GPU time to other companies like CivitAI Tarkus fucked around with this message at 16:21 on Apr 21, 2024 |
|
#
?
Apr 21, 2024 16:15
|
|
|
Tarkus posted:How valuable is AI generated porn anyways? I mean, the more you make the more worthless it is and people can just make it themselves with a mediocre compuiter. Thus the thing that will hopefully kill AI, once CEOs realize it isn’t the free money printing machine after everyone says “wait why pay for your ai poo poo when I can just make my own ai poo poo with a prompt?”
|
|
#
?
Apr 21, 2024 16:28
|
|
|
Wait - pay for AI? 
|
|
#
?
Apr 21, 2024 16:39
|
|
|
An art director talks about a film studio hiring AI techbros as matte background artists and how that went to poo poo  
|
|
#
?
Apr 22, 2024 09:22
|
|
|
https://twitter.com/molly0xFFF/status/1780601652049043628 AI isn't useless. But is it worth it?
|
|
#
?
Apr 22, 2024 13:58
|
|
|
People are going about this llm thing the opposite way around; replace all the c-suite with llm's.
|
|
#
?
Apr 22, 2024 15:20
|
|
|
ultrafilter posted:https://twitter.com/molly0xFFF/status/1780601652049043628 That's a really good article, expresses pretty much how I feel about the tech. LLM's as they are now are tremendously flawed. They are not expert systems, they are not creative geniuses, they do not think and they are not a good source of raw information. However, once you understand the limitations they can become very useful. I use them almost every day with the knowledge that they are very flawed and I work around that. That said, they're not 100 billion dollars useful like microsoft says they are, nor are they trillions of dollars useful like NVidia claims they are. And frankly, while I'm no expert by any means, I've been studying AI on and off since I was a kid and I've been working on my own little AI systems for the past couple of years, I'm just not seeing where these very smart people are getting anything even close to AGI from what we're seeing. They're throwing alarm bells and trying to warn everyone about skynet but frankly, I'm just not seeing the intellect that they are. LLM's are like an interactive jpeg of human knowledge, the deeper you look, the more artifacts you get. It's a form of intelligence but it's not 'smart' by any means. In all honesty, even though I like the tech, I suspect that there's going to be a reckoning in the next year or so. People are going to realise both that the big promises of AI aren't going to materialize nor are the doomer scenarios and people are going to basically reject AI as some flash in the pan. Then again maybe these big tech guys know something I don't, who knows. I mean, they are dumping tons of money into adjacent AI tech like humanoid robots but there too, the stuff I'm seeing is largely the same stuff we've been seeing for the last 10 years, incredible amounts of work and it's impressive but I'm not seeing it actually working in a practical sense.
|
|
#
?
Apr 22, 2024 15:37
|
|
|
people got really excited about the early internet and made predictions that we would soon enter the cyber-future and all skateboard and wear sunglasses indoors and techno music would follow us everywhere, based on spinning skull animated gifs and pixilated porn that took 3 hours to download others said “this is a flash in the pan and will go away” both were wrong and the reality is much, much dumber than we could have possibly imagined AI will have a similar trajectory
|
|
#
?
Apr 22, 2024 16:04
|
|
|
staberind posted:People are going about this llm thing the opposite way around; replace all the c-suite with llm's. Tarkus posted:C-suites as they are now are tremendously flawed. They are not expert systems, they are not creative geniuses, they do not think and they are not a good source of raw information. However, once you understand the limitations they can become very useful. I use them almost every day with the knowledge that they are very flawed and I work around that. 
|
|
#
?
Apr 22, 2024 16:08
|
|
|
Snowglobe of Doom posted:An art director talks about a film studio hiring AI techbros as matte background artists and how that went to poo poo Why would they get upset at criticism of an image they didn’t even create in the first place? Like if I typed “Mona Lisa” into google and my friend sitting next to me calls that chick ugly. Why should I take that personally?
|
|
#
?
Apr 22, 2024 16:31
|
|
|
Gutcruncher posted:Why would they get upset at criticism of an image they didn’t even create in the first place? Like if I typed “Mona Lisa” into google and my friend sitting next to me calls that chick ugly. Why should I take that personally? They're angry because there is no way for them to control the output, they've been caught out. The current models can produce some pretty cool looking stuff but it's extremely hard to execute any kind of actual vision. So while you can fix minor errors, insert and change things, it's extremely difficult to do something like change the perspective or change the style to execute what somebody wants. The only people I've seen that are able to do that with any success with AI are, well, artists.
|
|
#
?
Apr 22, 2024 16:40
|
|
|
Gutcruncher posted:Why would they get upset at criticism of an image they didn’t even create in the first place? Like if I typed “Mona Lisa” into google and my friend sitting next to me calls that chick ugly. Why should I take that personally? The tool's limitations are their limitations and they hate being reminded of that
|
|
#
?
Apr 22, 2024 16:40
|
|
|
Gutcruncher posted:Why would they get upset at criticism of an image they didn’t even create in the first place? Like if I typed “Mona Lisa” into google and my friend sitting next to me calls that chick ugly. Why should I take that personally? They think sweet talking the diffusion model into giving you exactly what you need is a skill that takes effort to learn and master and they take pride in their skill. An attack on their skills is an attack on their pride. But it mostly sounds like STDH revenge fantasy fanfic  naem posted:people got really excited about the early internet and made predictions that we would soon enter the cyber-future and all skateboard and wear sunglasses indoors and techno music would follow us everywhere, based on spinning skull animated gifs and pixilated porn that took 3 hours to download Openai is now doing several $ billion in revenue and growing fast. For better or for worse, LLMs and diffusion models are here to stay. Lots of people with money think it's worth paying for it, even in its current form
|
|
#
?
Apr 22, 2024 16:58
|
|
|
The fact that so many people have to insist that AI is useful is all the proof you should need that it isn't. A hammer doesn't need anyone to argue its merits.
|
|
#
?
Apr 22, 2024 17:03
|
|
|
GABA ghoul posted:But it mostly sounds like STDH revenge fantasy fanfic It's easy to make an image of a person, it's hard to make ten images of the same exact person from different perspectives (especially perspectives that the AI model wasn't trained extensively on). Some of the issues the person mentions it seems should be doable by a 'prompter' (like removing the trees) but a lot of the others get to the fact that someone who knows how to use other artistic tools would have a very easy time making simple tweaks but someone who is only able to futz with an AI prompt will never be able to have that sort of control over what they produce.
|
|
#
?
Apr 22, 2024 17:04
|
|
|
GABA ghoul posted:Openai is now doing several $ billion in revenue and growing fast. For better or for worse, LLMs and diffusion models are here to stay. Lots of people with money think it's worth paying for it, even in its current form AI is currently in the spinning skull dot gif stage and we are all very excited by the possibilities its going to get plugged into every aspect of communication and transform it, and become incredibly more complex and powerful, much like the internet has, while also being completely terrible in ways we cant possibly yet dream of AI is also going to make us stupid, possibly even more than the internet AI is truly Something Awful
|
|
#
?
Apr 22, 2024 17:21
|
|
|
I think my favorite part of AI is people saying it’s useful because you can have it give you a bunch of information on some subject, and that as long as you know to remove the incorrect information you have a good resource Forgetting that if I’m using generative AI to research a subject for me, how the gently caress would I know enough about the subject to know when the AI is wrong? Gutcruncher fucked around with this message at 17:34 on Apr 22, 2024 |
|
#
?
Apr 22, 2024 17:31
|
|
|
Mozi posted:It's easy to make an image of a person, it's hard to make ten images of the same exact person from different perspectives (especially perspectives that the AI model wasn't trained extensively on). Some of the issues the person mentions it seems should be doable by a 'prompter' (like removing the trees) but a lot of the others get to the fact that someone who knows how to use other artistic tools would have a very easy time making simple tweaks but someone who is only able to futz with an AI prompt will never be able to have that sort of control over what they produce. Oh I know. That exactly why it sounds like bullshit. Everyone who worked with diffusion models or done some research knows about their limitations and why they are useless for that type of work alone and only shine as a productivity tool for actual artists. Someone just hiring two "prompt bros", without doing any research or experiments or trial projects or getting some consulting from experts, seems ridiculous to me. But then again I'm in Europe where hiring is a major decision and commitment and responsibility. Maybe it's different in the US where you can actually hire and fire people without rhyme or reason whenever you like.
|
|
#
?
Apr 22, 2024 17:52
|
|
|
Tarkus posted:In all honesty, even though I like the tech, I suspect that there's going to be a reckoning in the next year or so. People are going to realise both that the big promises of AI aren't going to materialize nor are the doomer scenarios and people are going to basically reject AI as some flash in the pan. Then again maybe these big tech guys know something I don't, who knows. I mean, they are dumping tons of money into adjacent AI tech like humanoid robots but there too, the stuff I'm seeing is largely the same stuff we've been seeing for the last 10 years, incredible amounts of work and it's impressive but I'm not seeing it actually working in a practical sense. it's no coincidence that AI and blockchain both suddenly started getting a lot more nuanced skeptics (like Molly) around the time that interest rates got high enough that more traditional investments started shifting the seemingly endless glut of easy VC money away from speculative con artists and back to boring, traditional finance and propping up the normal banking hegemony. those skeptics have been there for years, but as Molly obliquely references a few times, until the rabid hooting from thirsting VCs dies down, it's not worth their time to speak up. when their expert and considered advice isn't ignored entirely, it is derided as neither expert nor considered and they lose prestige and standing rather than gain it. this is a slightly longer form way of saying that i feel you are correct but that the phenomenon is more wide reaching than you may think. juiceros and interest rates are inversely related. the AI/LLM space currently supports dozens of juiceros the same way the blockchain space supported thousands of juiceros 5 years ago, but now it's more like hundreds. the longer the situation goes on, the generally more genuine you will have to be in order to get funding. it's entirely possible for things to swing back the other way too much of course but there's not much risk of that in the next few years i'd think.
|
|
#
?
Apr 22, 2024 18:01
|
|
|
GABA ghoul posted:Maybe it's different in the US where you can actually hire and fire people without rhyme or reason whenever you like. Depends on the state but yeah it’s entirely possible to just fire people and replace them with “prompt artists.” And CEOs and the like are both greedy enough and stupid enough and short sighted enough to see “it costs less? Do it!” without thinking ahead to how it’ll harm them later It’s always about THIS quarter, not the NEXT quarter where you just hope it’s someone else’s problem anyway
|
|
#
?
Apr 22, 2024 18:03
|
|
|
Gutcruncher posted:I think my favorite part of AI is people saying it’s useful because you can have it give you a bunch of information on some subject, and that as long as you know to remove the incorrect information you have a good resource Ok, so this is the way I get around these things. I never use raw GPT for actual fact finding missions. I will however use it to guide me towards terms and concepts that are more common knowledge on things that are outside of the scope of my knowledge, much like searching on google, I won't trust the first thing I see. Also, if I have a vague question I can have it clear up the ambiguity for me so i can google search it. I can also take a lengthy explanation on another site and have the LLM explain it in context of my particular use case. I do this with datasheets or libraries. For factual things I'll use tools like Perplexity or the Poe websearch to find stuff since it uses RAG to give an answer. It will give you links to the sources of information. You should still check the links to see what is actually contained. So to get to your original question, obviously you can't know what you don't know but the more common the knowledge, generally, the more correct it is. Google is like that too, particularly when it comes to more esoteric stuff, lots of people are confidently wrong about all kinds of stuff. If you're looking for direct factual references then you are better off searching for it and then dumping that page into an LLM and having it walk you through it. LLM's are much better at being coherent when the data is in their context window. I think what's funny though is that some people think that the admission that these tools are flawed is some kind of own. I've been dealing with hosed up only mostly functional software all my working life. This is no different. You pick your battles, walk around the landmines and use what's useful. So far I've found use cases for what I do and that's good enough for me.
|
|
#
?
Apr 22, 2024 18:11
|
|
|
Gutcruncher posted:Depends on the state but yeah it’s entirely possible to just fire people and replace them with “prompt artists.” And CEOs and the like are both greedy enough and stupid enough and short sighted enough to see “it costs less? Do it!” without thinking ahead to how it’ll harm them later I mean, it's not just about the legal issues. In all companies I ever worked at it would be a major blunder to hire TWO people at the same time and get rid of them two weeks later. It would be considered a gently caress up and financial loss. The very first question from upper management would have been "Why didn't you just try one small test project with ONE person working on commission/as a freelancer?"
|
|
#
?
Apr 22, 2024 18:13
|
|
|
Is quot=edit? We may never know
|
|
#
?
Apr 22, 2024 18:14
|
|
|
GABA ghoul posted:I mean, it's not just about the legal issues. In all companies I ever worked at it would be a major blunder to hire TWO people at the same time and get rid of them two weeks later. It would be considered a gently caress up and financial loss. The very first question from upper management would have been "Why didn't you just try one small test project with ONE person working on commission/as a freelancer?" it's film/tv, people like that will be hired onto a project as freelancers and flushed just as easily
|
|
#
?
Apr 22, 2024 18:43
|
|
|
GABA ghoul posted:Oh I know. That exactly why it sounds like bullshit. Everyone who worked with diffusion models or done some research knows about their limitations and why they are useless for that type of work alone and only shine as a productivity tool for actual artists. i have a friend who works at an american marketing firm that did in fact replace a number of its artists with midjourney bux. didn't even hire any prompt bros--just added prompt broing to the job descriptions of non-artists. Insanite fucked around with this message at 19:11 on Apr 22, 2024 |
|
#
?
Apr 22, 2024 19:01
|
|
|
|
| # ? Jun 3, 2024 08:00 |
|
|
Insanite posted:didn't even hire any prompt bros--just added prompt broing to the job descriptions of non-artists. Honestly can’t imagine why every company isn’t doing this instead of hiring “prompt artists” when there’s no difference in the output
|
|
#
?
Apr 22, 2024 19:16
|
|