YouTube's Latest A.I. Threat, Explained
HTML-код
- Опубликовано: 31 май 2024
- ⚠️ RUclips's AI-generated video summaries pose a potential threat to content creators. Summary by TammyAI.
Help keep the channel running: / pillarofgarbage
Or join my channel on RUclips with a Membership: / @pillarofgarbage
Make a (free, easy) difference with TabForTrees: tabfortrees.org/pillarofgarbage
Timestamps:
0:00 Introducing the situation
2:10 AI summaries and 'contentification'
6:14 Why it's bad part one, the viewing experience
7:41 Part two, the subsumption of voice
9:17 Part three, it's... wrong!
14:45 Monetisation, irony, and Rumpelstiltskin
17:49 End
Videos mentioned:
• Empower RUclips with T...
• Three Specific Kinds o...
• The Cult Of JKS Continues
• Unity's Plan Won't Wor...
• One Last Scene - Derry...
• Bob Iger is Chatting Sh*t
• Everything Is Content Now
Articles quoted:
eightify.app/summary/economic...
techcrunch.com/2015/03/03/in-...
Articles shown:
pluralistic.net/2023/07/24/re...
medium.com/@DrChuckle/the-ris....
www.ft.com/content/dca07c32-6...
Reddit posts shown:
/ the_cult_of_jks_continues
/ according_to_pillar_of...
Other handy links:
Join the Pillar of Garbage Discord server! / discord
Follow me on Twitter! / pillargarbage
Follow me on BlueSky! bsky.app/profile/pillarofgarb...
I'm on Mastodon: mas.to/@PillarOfGarbage
or check out the Pillar of Garbage subreddit at www.reddit.com/r/pillarofgarbage
Second channel: / @quone
I'd also like to thank skyeplus on RUclips for signing up to this channel as a member - sadly I already had this video rendered, uploaded & scheduled, so I couldn't add them to the credit roll :( 
Развлечения
Help support videos like this: www.patreon.com/pillarofgarbage
(You'll also get access to the brand new exclusive video "Dying of Stress" & like 15 older ones!)
The AI narrators are going to choke youtube until there's nothing left and youtube is gonna react too slow to stave it off.
Did someone say "react" ?
Who do you think you are. the Fine Bros?! /s
I'm pretty sure all the reacting will be done by AI too
Weird take. Maybe just watch something else?
@@personzorzhaha
For me the most disturbing implication of this is the notion of people letting AI handle all of the "hard thinking" and analysis - trusting that a few simplified bullet points are the CORRECT and ONLY distillation of a complex issue or idea being discussed in the video. It's turning over all the critical thinking and reading comprehension to a machine, and not questioning the conclusions being presented.
That’s been happening with the news for a long while now…
Don't people already do that with videos that dumb down a topic way too far?
I think Dune said something about that.
It reminds me of that lawyer who had chatgpt write the argument for the case they were representing; "AI" has been so detrimental to society that you can't even pay these dumbasses to think critically any more.
As others have already said, this is already what the news (and tons of human RUclipsrs) already do - distilling down information, and usually getting it wrong or putting some spin/bias on it that wasn't really there.
And any viewer who would blindly trust an AI summary without actually watching the video first, or doing their own further research... wouldn't have watched the video in the first place.
So this isn't really anything new. Just a slightly new expression of the same old propensity of humans to be lazy and miss the point of something.
AI at the moment seems to be like: "Wanna skip all the fun parts in life? We've got your back!"
my personal theory is it's because the tech executives who are pushing ai don't actually enjoy anything. they don't actually like watching a video for yourself. or making any art. and really my response to ai at this point is why should i read anything no one could be bothered to write. why should i read a summary of a video no one could be bothered to write if i can just watch the video that a real person made.
"Have you ever been down the rabbit hole of RUclips videos, only to wish you could grasp the essence of a video without watching the entire thing?"
Uh... No? Not ever? They're videos for a reason.
Yes. All the time. Never been caught by a clickbait? watching a half hour video to get the (always disappointing) answer to the question in the title?
Here's five ways to crack an egg, but a different sponsorship and music break between each of them! 5 easter eggs in this movie, but there's the most boring director commentary between each of them!
These are videos based on information they promise you, but the information is too simple, giving it out would kill their watch time, so they pad it with whatever, to make a full video.
They're videos... for a reason? What? They're videos fo- WHAT?
Can you make sense, please?
@key37raminus that only happens to me when I've curated my feed badly. I just don't click on videos like the ones you describe (or click off them immediately when I realize what they are). It's annoying running into them but no one's holding a gun to my head making me watch poor quality videos
1 hour food recipe videos starting with 55 minutes of "that's the dish my grandmother served every sunday, you see Sunday's are very important for my family because my grand grandparents..."
There are some channels who do great research and craft their content with care, making sure that they present and articulate their points.
Whereas, there are others who simply hold their mobile in their hand and ramble for half an hour, more or less saying the same thing over and over again or getting someone else to react to what they just said.
Yeah, there's some good well executed content out there, but much of it is dross.
I genuinely don't understand what drives some folks so hard into this "AI" fad.
Who wants art without artists, writing without writers, critique without critics, and video essays without essayists?
Reducing artwork to "scrapes" and RUclips videos to banal bullet points feels like an exercise in streamlining the human out of the art. That some hacks use it to make "content" just feels... Gross.
Good on you for spotlighting it.
They're trying so hard to convince everyone that no matter how weird or bad it looks, it's (somehow) objectively the future just because they throw in sci-fi words like "AI".
It's like a war on art- there's a tangible effort to push out traditional forms of art. The incentives they give show they barely understand why someone would want art, nor why anyone would want to make art. It's actually toxic how some of them seem to be projecting their self-loathing on all of humanity. As if you're supposed to hate yourself as much as they hate themselves, to want to replace yourself with a machine. The "Machines will soon do it better so you should let them be in control" is such a feeble argument.
As someone in the art community being more and more involved into AIbros trying to justify the use for making pretty pictures: they often are people who convinced themselves that art is only matter of talent and "being lucky to have such talent" and therefore see AI as a way to skip all those years the very people they see as "unfairly lucky" took to become that good.
I draw since 10 years now, have my share of community and am relatively happy about my humble results. I sucked at making pictures for a good couple of years, it's normal. We all start from the bottom.
If I go to a small concert and see this guitarist doing amazing solos, do I tell myself "Oh, that guy is so talented. He was probably born with such skill and he never faced any struggle into this his whole life!"? No. I understand the guy most likely play that instrument since he was a kid or a teen. It most likely took him years to become that good. I am also aware he might bea genius at playing guitar, but problably would struggle at drawing a face looking symmetrical (with exceptions, as always). I have my limits and specializations. This guitarist has his own.
AIbros to me seems to find this unacceptable. They must have it all and cannot accept to be in a room with someone better than them at something. So they must use AI to "fix" that like it's some sort of birth right to do so. Not realising they are lying to their own selves.
It's the modern equivalent of AutoTune abuse. It was a cool effect decades ago when it was used sparingly. Now it's the entire lyrics of every song. Recording industry executives justify AutoTune abuse by claiming fans want pitch-perfect performances, although no music fan ever actually said that. It's just a way to make talentless hacks passable rather than cultivate real talent, which record companies did many years ago.
The basic mechanism driving the success of AI is the personalisation of consumerism. Initially, mass production technologies allowed the creation of very cheap but very generic goods - you can now buy a piece of furniture for a fraction of the price you used to have to pay, but the trade-off is that it's the same model as every other home in the country has. Over time, though, we've seen new companies appear making use of new technologies to diversify these products again - in many industries, you can now purchase things moderately specific to your tastes and needs but at the price of mass-produced goods. 3D printing is one area this is getting pretty great, for example.
AI is perceived to be the next step in this progression. If AI does what people want it to do, there'll be no need to pay someone with the right skillset or tools to make that specific product you want, which can be something so niche that no one else would ever want it, because you'll be able to tell a machine exactly what you want to have and it'll make it for you at mass production prices. Of course the flipside of that is that corporations can also produce ever-more specific content too in the endless war for your attention, so it'll be interesting to see where that goes.
As for the removal of the artist from the art - this is simply something a lot of people don't care about. It doesn't matter to them who the creator was or why they created something, as long as the creation itself is good. That was already true a long time before AI - just look at how much general contempt there is for "modern art", which is a genre so heavily about the artist instead of the art that the art itself can often be deliberately bad - that contempt is coming from people who simply don't care about the artist behind the art. So AI doesn't represent any change on that front for those people, its limitations are only in the quality of the product created.
@@yurisei6732 You're right, people have always has contempt over artists. They just want to be served something now and for free.
that is horrifying, truly, the rise of Ai has been a constant glooming menace over our heads in every corner we try to go
If you really want to hear about actual potential dangers of AI google Rob Miles. He has a lot of good videos on AI safety and how unrestrained AI research can be extremely dangerous, even existential. I recommend this video for a start, its a good introduction on why AI safety might be extremely important: ruclips.net/video/9i1WlcCudpU/видео.htmlsi=QFH9cHyrQBm8bmOZ
can you imagine the things AI will be doing by like... 2027?
AI is pretty smart at the moment (talking about the large language models and GPTs and any other assistant or chat bot alike), but it without a doubt can make some very noticeable mistakes that can be mostly easy to notice, but when all of those little mistakes get ironed out and there's will be absolutely no way to tell if a message/paragraph or even a whole essay was truly said by AI or an actual person... it's quite scary to think about and I begin to fear for what's to come, there's no clear way to tell what would happen...
@@AltKaxREAL The good news is, it will be very, VERY hard to iron out those mistakes, due to the nature of LLMs. This technology has been around for over 40 years and the people who make them have known about these issues ever since and they are still just as clueless now as they were than about how to fix it.
@@rumrunner8019 genuinely hoping for the best 🙏
@@rumrunner8019 Lol wat? LLMs are extremely demanding even with today's hardware, are you saying they were in use in the early 80's? Agree on it being likely impossible to iron out the bugs from them.
And so the enshitification continues...
As capitalism intended
while watching this, i had a sudden moment of understanding: they think that everything on the internet is like those awful recipe pages where you have to scroll through the writer's relationship with their mom to get to how to make the goddamn cookie. but with video essays, the essayist's relationship with their mom is CRITICAL to the point of the piece! Also i got a terrifying premonition of a world in which people were getting paid to use ai to create those overwritten recipes, and other people were getting paid to use ai to figure out what was actually important from that, and so no end user ever sees 99% of what the first ai got paid to make.
terrifying
_mommy issue cookie bloggers_ walked so AI bros could run
But that _is_ the majority of content (other than you-know-what). It's listicle nonsense, it's fast-forwardworthy ad nonsense, it's clickbait teasing you into 10+ minutes of filler to hit that adsense runtime. Even this video is a great example of the trash we have to filter through, with its threatening algorithmicly optimised thumbnail and coy title. This isn't "RUclips's latest AI Threat Explained", this is a rambling video about someone's personal thoughts and fears on the matter - which he's fully aware of! There's nothing wrong with sharing your opinion to a captive audience, but this video isn't doing that. It's shouting it like a lunatic on a street corner. It's a sensational opinion piece masquerading as a public service announcement.
So when you're faced with such disingenuous ensnarement, when you're having to sit through ten minutes of waffling irrelevant prose to get to the Object Nasa Discovered But Couldn't Explain, ten minutes of quippy banter and self depreciation to get to the One Weird Tip To Fix Your Anatomy Sketches (you're doing it WRONG), and so on... Can you understand why people like me are going "hell yeah, inject that information-dense bullet point summary directly into my-"?
I don't know this guy. I don't care about his opinions. He says there's a sensational new threat - OK, give me the Cliff Notes. That's my thought you process, just as a random viewer who got pulled in off the street. @PillarOfGarbage, I hope you don't take this as an affront, and more of an opportunity. The old tricks worked for a while, but people are wising up. This isn't "you bad" or "AI bad", it's just a universal race to rock bottom that you're simply getting lapped on.
Edited for typoes
I commonly feel like that, in many videos. They are stalling for your watch time. They pad their scripts. I don't doubt some of them use AI to do this, because AI is extremely good at padding text. Filling it with fluff. Only there's AI that does the opposite, so you're right, but we're already living in that world. Many webpages today are definitely created using AI to pad a simple concept, and create text to fill a page. And when you ask an AI for an answer, it reads from that. Or you can get to that page and ask for a summary. This is already the world we live in.
@@tahunuva4254if my ADHD riddled brain can handle padding, so can you.
Back in the day, there was a legitimate reasoning behind overwritten food recipes.
Very simply, the recipe itself, the instruction how to make a certain kind of food, cannot be copyrighted. Anyone can do their own pancake book summarizing all the way it's been published how to make pancakes.
What you can copyright and protect are those digressions and stories and philosophical arguments surrounding the pancakes and what they mean to you as an individual writer in particular and the global socioeconomic civilization in general.
It's probably one of the reasons why we don't remember Nostradamus for his recipes for making jam. He could copy those from anyone and anyone could reprint them themselves.
But fill those recipes with impossibly vague prophecies and imaginative woodcut illustrations with fearsome rabbits on horses fighting off tyrannical creatures from the heavens... Now that you can copyright and protect.
I've been fascinated with LLMs and have been messing around with them a bit. One thing I've found is that they are really bad at nuance. I asked ChatGPT what the song "The Poet and the Pendulum" by Nightwish was about. It gave a fairly banal take which was mostly accurate. Then I asked if there were any references to Edgar Allan Poe in the song. It said that there was in the title of the song and the imagery and feeling evoked by the song. Then I asked if any lyrics in particular were references to Edgar Allan Poe. It said that there were no "explicit references." As a human, it was easy for me to say that's wrong because there is a descending blade and someone being strapped to an altar. These are clear references, but since they are not explicit quotes from Poe's work, ChatGPT couldn't understand that.
LLMs are a fascinating tool and have some interesting potential, but I think can easily be misused. I'm particularly troubled by the fact that Sam Altman says that ChatGPT should have unlimited access to copyrighted content without compensating the owners of the content. They make money off of ChatGPT! How can it be okay for them to use someone else's work for free and make money from it?
"Facinating" is an interesting word for a "do-it-all" statistics simulator that fails to do anything but expend a ton of resources.
Altman is basically Zuckerberg 2.0
I appreciate the clarification. You are correct. I apologize for the oversight. "The Poet and the Pendulum" by Nightwish does indeed contain references to Edgar Allan Poe's "The Pit and the Pendulum." The lyrics describe a descending blade and the imagery of being strapped to an altar, directly alluding to Poe's work. Thank you for pointing that out, and I apologize for any confusion in my previous response.
@@macmcleod1188 is that supposed to be ai generated?
To be fair, ChatGPT has very human-quality levels of "maybe it'll get this, maybe it won't." And if anything, it exceeds humans in capacity to (eventually) acknowledge mistakes (and actually adapt to the correction), though it is still about as likely as a human to instead double down until directly corrected.
The real limit is that it is "average human" at _everything_ whereas actual humans are very often properly expert at _something._ But if "average human" is sufficient, it's at least a _really fast_ one.
The real advances in AI are the ones powering scientific research anyway - not the stopgap replacements for broken internet search engines.
13:37
there was a situation where that happened, before ai. someone got cancelled for saying a slur... except it was youtubes auto captions that put that there, and he never actually said that. so i wouldn't be shocked if people started doing that 🤦♂️
"They're spinning straw from gold"
Absolutely nailed it. As a creator trying to break in being on the edge of this new era on youtube is scary. I can't help but feel like I missed the boat
The truth is that people were saying that since 2010s. It's never too late to start a channel. Just do what you want on there and be patient if you want an audience.
@@Adrian_1114 i've actually been making videos on this channel since like 2011 but only started taking it seriously now lol
You missed the boat on this particular era of algorithm and ad revenue point-scoring, you're right on time to catch the next one. It's only a problem if you're the kind of person in love with your own voice - anyone whose videos are already snappy and engaging isn't going to see much of a problem with their videos getting summarised.
RUclips never had a boat, people have been calling it shit since it was first monetized. Just make videos you enjoy making.
Perfect time to start a career as a vtuber cat person though
I think that the people responsible for today’s obnoxious AI should receive the anti-Nobel prize. If there isn’t a category for that, they should invent one. I’m sick of this. It’s so unnecessary while also deceitful and sets us back as a human race.
I propose they win a permanent beach vacation to that Island Napoleon was sent to.
@@RobertStollElba
@@drewsachs6238 If the world was right, the home for every technocrat and technocrat wannabe.
While it could be useful, 99% of the proposed used cases are just annoying or things it's not good at.
The most annoying thing is how it's being promoted. They want you to think buying into the ecosystems of a handful of AI companies puts you above everyone else, and that AI will inevitably be so good that it deserves to be applied to everything- as if it would be immoral not to.
The kind of "AI" that gets a lot of press is LLMS, but those are more like glorified autocorrects and not the seeds of a superintelligence that deserves to be trusted more than all humans.
@@sharonoddlyenough IgNobel prize is too prestigious for that. It is for things that are actually funny.
“Tammy.aye” 😂 they didn’t even care enough to fix the bot
Forreals 🤣
Honestly this is the kind of thing that if the company was more competent I could see RUclips actually wanting to crack down on. The more that this generative summary thing exists, the less people spend watching RUclips videos where RUclips wants them.
Ugh, AI continues to haunt me. In my job, I often feel like I in a one-woman war against AI tools. I'm a copywriter (because I have to pay the bills with my BA in English somehow), and I feel as though I have the constantly justify my own existence. I won't pretend that when I write copy for Meta ads or a website or emails that I'm making art, but I try to be a mark or two better than the AI-generated sludge. IDK, it's frustrating to see AI creep into everything. Even our meeting notes get taken by AI, with a prrogram that listens to our zoom calls, takes a transcription, and attempts to summarize the main points (with limited success). It's all just kiund of a bummer, you know? Great video, though.
At this point I kind of want to get out of any job that requires using a computer :) Unfortunately it's what I've almost been exclusively training with for my entire life.
Saaame… I’m a DevOps Engineer and I’m the only one at this company who is not willing to use AI.
Meredith, do you have a blog or personal website where you can freely say what you want or post creative works purely for personal satisfaction? You may find that liberating and cathartic. Devoting all of your creative energy to the corporate grind will suck your soul right out of you.
Copywriting I think is one of those industries that's going to go the way of the quality controller - it's easy to automate the production of this sort of content, but you're not just going to be able to print the first thing your algorithms make because it's important that the company's desired brand images are maintained, so we'll probably see a pivot of marketing staff from basic marketing material creation to marketing material quality control.
that last one is a big eyeroller for me. "ok folks we need a software to transcript our video meetings." it's called Closed Captioning, just get an IT to hook it up and there
RUclips's AI-generated video summaries pose a potential threat to content creators. 00:00
AI revolutionizing RUclips content discovery and consumption. 02:51
Negative impact of AI-generated video recaps on viewer experience and content culture. 05:46
Misattributed quotes and inaccuracies in AI-generated video summaries are erasing the voices and work of creators. 08:47
Inaccurate video recaps and summaries are misleading and unreliable. 11:48
Risks of online generative AI tools and monetization leading to misinformation. 14:14
Content creators face challenges from AI-generated content, impacting their livelihood and creative freedom. 17:11
I am so sorry, I couldn't resist a temptation :)
it's bleakly funny to read a summarybot come out with 'Misattributed quotes and inaccuracies in AI-generated video summaries are erasing the voices and work of creators.... Inaccurate video recaps and summaries are misleading and unreliable.'
i notice the time stamp & Ai quote , is also commented on by new creators, they either respond with , is this for real,,
is this a bot,
or,, first alien contact ,
&/or futurama 101101 time machine puns
or the new reee.@@PillarofGarbage
had us in the first half, ngl...
@@PillarofGarbage this is axactly what i came here to see after watching the video lmao
@@PillarofGarbage It really is turtles all the way down the rabbit hole isn't it?
I'm about 90% sure the *actual* use case for these tools is to convert a video into a format that can be used as input for some other AI software. This is similar to the decades-long process of categorising and describing billions of images so they can be searched using text prompts. Now the industry needs to do the same with videos but it's far, far more difficult as you are dealing with thousands of frames and an audio track, not just a single image.
So here's an automated way to create a description of any random video so it can have machine-readable context and details for whatever other software comes next in the pipeline.
I don't think this is for humans.
That'd still be pretty great for humans though. Image searches are such incredibly useful tools that I'd go so far as to say they're vital for the modern world. If we could genuinely search videos as specifically as we're able to search images, we might finally be able to bring youtube's algorithm back under control and get it presenting you things you want to see again.
That makes how innacurate via baised the results are even worse. Garbage going In, Garbage coming Out.
If tech bros have their way, humanity’s only purpose will be to feed data points into AI
Not even, it'll be to consume what AI feeds them because the vast majority of them aren't talented enough to be worth using as a data source for the generation of consumable products.
The internet's been on a rapid decline ever since all of this AI stuff began popping up everywhere.
We always knew Ron's ex wife was evil, but this is too far.
underrated comment
😂😂😂
Now I'm really worried about Tammy 2 coming... Tammy 2.0?🤔
Considering that RUclips's CEO is a Tech bro who loves NFTs and AI, i'm nit surprised that RUclips will inevitably end up as an AI wasteland
I don't think these summary robots are intended for human consumption. I suspect they are intended to help other LLMs grab shallow data about videos during their training process.
Absolutely fucking dystopian. I'm more convinced every day that anywhere AI is applied, it removes meaning, humanity, independent thought.
It’s ultimately soulless and without a point of view that it hasn’t consumed. no amount of refinement or advancement will change that fact
@@Nick-cs4oc exactly. Humans synthesize information to make something new that's entirely their own. AI just vomits up what's already been done
AI sucks, and the less of it, the better.
I am really curious about the mindset of someone who cares about RUclips videos enough to use a secondary tool to "understand" them but not enough to just actually watch & enjoy the video.
Addiction, simple as that. Feel uncomfortable when not consuming, but also feel uncomfortable when the content being consumed isn't stimulating enough, so must consume lots of very small snippets of reward-optimised content.
I usually dont comment but for what its worth, your videos are important to me and on a personal level I can't imagine ever replacing them with anything. They production, delivery, and especially your sense of timing are, in my opinion, completely 'you' and a huge part of the appeal.
Thank you! Much appreciated
No lies detected. Well phrased!
'Tedious scrolling through lengthy videos' Your Derry Girls video is less than a half hour long.
counterpoint: 30 minutes is nearly an hour which is basically a whole day
@@PillarofGarbage BIG FACT
These AI summaries are as anathema to context as *original*, 140 character long limit Twitter was. If anything I feel like Twitter was the genesis of that kind of hyper-summarizing and Tammy is the next step. Twitter was where people went when they didn't want to actually read, and AI summaries are where people don't want to watch.
''Pulling straw from the gold'' now that's a good turn of phrase.
What really gets my goat about this 'TammyAI' thing, is how often it utterly mangles the summery... and always to the benefit of a specific political and world view.
I shiver at the idea of people doing onto already summed up videos the equivalent of news of "just reading the title and assuming they understood the whole context straight away".
eeek, that's kind of annoying about the AI bots. 1:03
but i do like that there's been a trend for a few years or so now of putting time stamps in videos. Mainly the longer ones where it might be a news type thing and you're looking for a specific story to be covered and maybe just want to get to that one specific spot.
Titles, yes. But statistical models that takes everything on its face (including sarcasm and exaggeration) and isn't even able to perceive multiple sources due to how similar everything looks as a statistic shouldn't be trusted to find those.
The very reason the mind has evolved as they have is to parse through the complexities and incoherence of human communication.
yeah, that part is an annoying tech jerky part that I don't like.
this seems like an evolution of the AI plot summaries for famous movies that get posted to ticktok/youtube shorts/instagram reels. it feels just as soulless.
If one of those Tammy AI comments appears on this video, I hope you pin it.
Machine learning is a Big, shiny tungsten sledgehammer. It takes in data, bashes it into statistics proverbs of pink-meat quality, and spits it out without anymore rhyme or reason, thinking it's doing a good job.
If the future of writing, music, art, etc is statistical models, I'd rather get a head start on the wooden totem carving market. Hell these "tools" don't even calculate 2+2 - they have to inspect the pink-meat chunks for what the answer is and even then it'll probably give you back the anser "7." If you need me, I'll be unplugged in the middle of Yellowstone.
Well said.
I would even say you are still giving it too much credit by using the term "thinking".
As you have stated, in essence, """AI""" is just a filtering algorithm based on statistics, so it doesn't even think, it just vomits whatever has already been created in an order that sounds familiar (pattern matching).
As a computer scientist, I've always hated the usage of the AI term, as it's just a buzzword for "algorithm" mostly used to attract investment from normies.
i thought this was gonna be about those ai generated movie recap channels
I was expecting this to be about the rise of recap channels using AI recep and voice but I think this might be more worrying was unaware of this side of ai recap
This is where "making money with AI-Generated arts and writings are okay, but even writing fanfics and drawing fanarts of anything other than Public Domain Franchises for absolutely anything are not" thingy are come from, via hypocritical enforcement of copyright laws
I love that analogy of spinning straw from gold- you may have coined here the perfect meme for describing the process by which AI takes 'Art' in all it's forms and transmutes it into 'content'- but it's content shorn of the very idiosyncrasy that makes that content worth anything in the first place. When chatGPT first appeared I really tried to push it to give me a single genuinely creative response- and I failed- the thing is like a compass that points unerrringly not to north but to a location where all distinctiveness becomes 'rounded down' to the point where meaning becomes lost in a sludge of bland circular assertions devoid of any real thought.
Wether this intellectual emascualtion is the product of over cautious training by the developers or is inherent in the statisical nature of the technology itself is not clear- but I suspect the latter- after all a machine that derives it's responses through the reductive process of extracting patterns from vast swathes of data in order to produce a single cogent outcome is hardly likely to be controversial or 'edgy' in it's 'thinking'.
I see this as a way to rob creators of monetization since it tries (and fails) to give people the gist of something so they don't have to watch it.
The way the AI explained what it does, kinda make it sound redundant. We already have tools to skip past parts of a video. We call it sponsorblock, and it's comply user driven.
I'm middle age, I used to be the tech savvy guy for friends and family but I've started to fall behind with the rapid advancement of these subtle and not so subtle uses of AI. I mention this because people who fell behind a decade or two ago are not prepared to critically assess what is about to be thrown at them. Its terrifying.
Your BS detector works at any age or skill level. Learn to trust it.
That's actually a really interesting point, we may well see AI tool competency become the "tech savvy" of the next generation, not just in using these tools but also in being able to recognise when what you're seeing is AI bollocks - the same way that older generations today often struggle to not believe everything they see on facebook.
I hear you and I agree, but I also think creators need to be more considerate of the viewers time. A good example is this whole mamamax fiasco. You simply cannot find a single video regarding this topic that’s less than 1 hour long and those are the minority, most are usually between 3 and 4 hours. After watching a few of them and following the rest of the situation as it unravels, there’s no reason why no one couldn’t have summarized it all in at least 30 minutes. Hell, anyone who knows all about this situation could explain everything in under 10 minutes including updates. Sometimes I’ll get recommended a video essay about something I’m really interested in but it’s a 3 hour + video, like, I got school, other people work, do they think we live off of watching YT videos?
Buddy, 1.5x exists, and also, you're not entitled to videos. If a video is too long, find another one. Like, I guarantee someone on youtube has a video on mamamax
@@wyltedleaves If anyone needs to grow up, it’s you. First learn how to read as I clearly stated I do agree with the video and did not advocate in any way for AI summaries. Second because who are you to tell anyone to grow up and judge them based on their interests? You watch a lot of things about a variety of subjects that do not affect your daily life in any way but you have all the right to be interested in whatever subject and so am I and everyone else on the internet. I brought up a topic respectfully, you’re the one complaining about what I’ve said because you have no empathy or maybe you just had a bad day and wanted to lash out on someone who you disagree with. Entitled? Maybe for someone who’s worthless like you and can sit on their basement for 24 hours straight watching RUclips videos because daddy pays the bills and you don’t have any responsibilities but everybody else has lives.
Something not being to our taste or desires isn’t an excuse to sell the whole thing down the river. If you want a 30second breakdown then RUclips certainly isn’t the place
@@Nick-cs4oc Where exactly did I talk about 30 second breakdowns? Do you even know the difference between “minutes” and “seconds” or what exactly the expression “selling down the river” means? Because it doesn’t apply to what I’ve said at all. Not every hour long video is bad but you also can’t deny there are tons of creators who add too much unnecessary material to stretch their videos runtime.
I agree but it looks like this padding is added because of how YT promotes videos and pays creators. If they would still reward 15-30 minutes videos the same, I don’t think people would make these long ass videos.
This feels like a parody of real life
That's basically what AI is, or at least what people have chosen to use it for.
Funny thing about that ai time stamping is that the most upvoted comments are time stamps, so you don't need to scroll.
Each time it is making time stamps of the content, it is training the AI how to write a script. The time stamp thing is a trojan horse. They are getting video viewers to use this for fun, but they are really training AI. Next Tammy will start generating their own content based off what it learned from these time stamps.
hy , thank you for talking about this problem
Chainsaw Man slaps.
bro...why do you have A closer look in you're playlist ?@@spritvio639
How did I never notice that neither Robert Carlyle nor anyone else on _Once_ know how a spinning wheel works? He _literally_ has the wheel the wrong way around. It sits perpendicular to you and you attenuate your fiber into the orifice where your thread loads onto the bobbin. You'd load multiple bobbins and then offload them to a knitty-knotty.
If you tried to offload the bobbin in place, the wheel wouldn't spin. The bobbin "floats", staying in place until you release your thread, otherwise you couldn't control the thickness or consistancy of your thread or yarn or whatever weight you were making.
✨
❤️
Just after the first Musk inspired exodus from Xitter, I had thoughts.. “getting the message” It still applies today.
God, I'm so embarrassed now because I saw a timestamped and emoji pared comment just like that a couple of months ago and people were arguing if it was real or AI. But the same user who posted the timestamps actually replied saying they had done the timestamps themselves for fun, so I believed them. Clearly though, this was just a type of AI I wasn't aware of. Or maybe the person was truly just autistic (lile me and many people under the video) and was emulating a style of comment they had seen used before and thought was the popular new thing. Hell... I don't know anymore... This is all so stupid and confusing. I'd rather just watch the damn video!
Honestly those Ai models do highlight how bloated videos have become, like a lot videos are just long for the sake of pleasing the algorithm, those long videos kinda encourage the watcher to play it into background as filler noise while they focus on something other like video games. The result is them watching a video that struggle to pay attention to. There needs a good need for concise content. RUclips videos feel like 5th grade essays with a minium word count requirement--------------bloated.
Even Googles Bard AI which runs on Gemini has this same issue with getting time stamps and context correct. It's interesting that something that seems easy is a difficult thing for these AI's to do correctly.
It reminds me when I was in school, people wouldn't actually read the novels we were assigned for class and would just read the summaries of them posted on websites. AI feels like it was made by and for those lazy people, who don't actually care about creative things and the entire body of information. It's all a shame.
I have seen some of these randomly. RUclips might as well as put in the lorem ipsum text for the "clarity:" it provided.
Ai is actually a threat. I've lost a job as an illustrator due to it
In the best case scenario where these tools were actually functional it makes the assumption that a quick summary can give you all the value you need.
Responding on that basis is like only responding to some conclusions and skipping all the actual arguments.
It also completely misses that people generally don't engage with RUclips to blandly consume information but because they're interested in watching it. It's like the value of reading a book or watching a movie does not exceed that of reading the plot summary on wikipedia. It reminds me of the tech-bros who are so set on making a meta-verse but can only think about it as a model for sales and ownership and completely forget that you actually need a world anyone would want to live in.
So who is this really for? It seems exclusively for people who want to respond to something they don't actually want to engage with. Which is shitty. It doesn't just let you ignore context, it automatically skips over it.
I thought you were going to be talking about the AI movie recap channels popping up. It’s like Cliffs Notes for movies and series. I watched a few unknowingly and noticed that the narrator mispronounced simple words and didn’t even have an accent. Now I avoid them like spam. I’m also seeing more bots in the comments of some videos.
Skynet is upon us, soon they will make Terminator version of people do make comedy movies and do art. The future is here
If they do, remember, Genysis is Skynet, all that.
Singularity and U. We all can’t wait.
"I'm the future, comedy will be randomly generated" -christian tomato.
@@goober324 I don't quite know about that, but we can read it directly from the news.
People keep saying “Skynet this” “Skynet that” bitch we’re getting Auto from Wall-E
Yk, with how quickly i clicked, i now realize that pillar of garbage has become one of my favorite youtube channles! Thanks for the great videos mate!
People got sick of reading proper articles and essays on things, so we started watching and listening to the same information in video/podcast form.
Now, the culture is struggling to even pay attention to a 5 minutes, let alone 20 minute video. Tiktok's short form style may be part if why alot of longform content or even medium length (10-30 minutes) is now struggling to keep people's attention. This new ai feature is just a product of low attention spans and weakening literacy.
It's a shame because I love a good, long (often multi-hour) video essay that's packed with interesting points and/or evidence to back up and break down claims. I'm already removed from the generation that actually read through the information themselves, (I set a goal this year to start reading fiction books and properly review whole articles before forming or sharing a stupid opinion 😅) but atleast my attention span is long enough to enjoy more substantial content.
I fear for gen alpha ....
The 2+ hour long-form RUclips video essay is a recent phenomenon. I don't think people 10 years ago were ready to watch a 2+ hour documentary framed around a specific sound effect in a popular video game. Last year, millions of people have seen it and unironically enjoyed it. Just because people consume more short-form doesn't seem to correlate with the death of long-form is my point.
Might wanna put a date next to the title so it doesnt become an issue for your next video a few weeks from now when a new AI threat is found out
imo the root cause of the AI issue in general is just lazy people with short attention spans. someone too lazy to make their own videos, too lazy to make their own art. hell everyone who has used discord has seen someone complain that a post is too long that they wont read it but the post is two sentences. the unfortunate fact is, there are too many people willing to spend money to save time but have nothing to do with that saved time since they are saving time on creative subjects
Fantastic video
I personally agree that consumerism is at the heart of the problem. I’m not gonna waste my energy typing a 5 paragraph essay on what I think, but the short version is that it’s not just “content” but everything that’s being reduced to a product and I fucking hate it.
Commodity fetishization
Not just consumerism but people don’t really have time when they have to work 8 hours then travel 2 or so hours every day to just get by… These companies are advertising AI as the solution to this but ultimately the solution would be less work hours without pay cut. Hardly any politicians will implement such a drastic change sadly.
But the flipside is that when everything is a product, anyone can be a producer. I have no desire to go back to the days when there were only two TV channels because very few people had the money or the connections or the talent necessary to make a living producing media.
@@yurisei6732 Bro there is a big difference between AI and having 2 tv channels… You were already able to learn anything just by having an internet connection, determination and curiosity. AI is not needed, people have to get back their time so they can spend it on hobbies and family.
Skynet isn't going to kill us. It needs something to advertise to
This is a wonderful tool, never again will I have to spend 4 hours to learn that 'James Somerton is a bad man'.
That's a weird example, because the detail & nuance of that video in particular was _absolutely_ necessary to serve the uncontestable, paradigm-shifting role that it did. I think that being a viewer's only takeaway from Hbomb's video says more about the viewer than it does the video.
@PillarofGarbage it was a joke, I love H.bomb's videos, I still remember the feeling at the 2 hour mark where he says "this video is actually about James Somerton..."
@@Paul-wd8cz apologies, then! you’d be surprised how many people in this comments section are saying similar things seriously…
Worrisome.
As someone trying to get into voiceover/voice acting, I try to avoid these AI voice videos.
If you want a narration I can do it cheap, and I won't even charge you to sound human.
Funny thing is I did want to get into content creation. But then this came along.
The Recommending Numbers bless this video.
a.i-tech-bros suffer from contempt toward creators, artists, and their process. they only see them as gate keepers of value behind their skill-wall, they want to "democratize creation of art" by lowering the bar to the Mariana Trench, ignoring the skill, the care for the craft, and the personal meaning behind it
3:25 I dont see how reading is less tedious than watching. Who are they trying to fool with that?
This seems more like a case of something an AI is capable of doing and so one is made for doing that, despite that being not very useful.
Great video
I get your point, but if this recap is down in the comments, it won't 'ruin' anybody's watching experience that doesn't want to seek out the comment. And I've seen too many BAD video essays where I just want to yell at the screen "GET TO THE POINT ALREADY!!!", where a summary would really be helpful to jump over the fluff.
Plenty of people sort by ‘newest’, and on non-desktop platforms the YT player nonfullscreened will regularly cycle through comments below the screen.
I think y'all need to learn to slow down....and pay attention. We're very much missing that lately.
Specially when you have to watch half the video to figure out what is the topic. The title doesn't help at all. There are plenty of tricks youtubers use to get you to stay watching their video, for the longest time possible. This technology sounds like a defense. It's like clickbait, but really long and time consuming.
Like this video lol!
@@Antiyoukai bro this video's focus is laid out pretty clearly in the first minute or so
Okay, bad as this is, I can absolutely see a use case for it, How many times have I been having a technical issue and googled a fix, only to be met with a youtube tutorial where I have to sit through a lengthy intro and subscriber beg before they get to the actual fixing of my problem? In those cases I would 100% prefer an Ai summary.
Excellent video as always.
11:26 why do i get day shaw vue for the biggening of the darry girls video/even though not seen it?...
,another video did the same thing?...
Man, can't wait for all the needless drama videos resulting from someone who skipped to the end of an ai made summary of a video resulting in a massive misunderstanding because they skipped all the content and made an assumption based on what an ai says was in a video.
As if people looking at only parts of a video, or just reading the video, and attacking the creator or the video w/o watching the video wasn't annoying enough, this will definitely amplify that shit.
Thank you for educating us about this. AI is a mess. I have noticed that my feed is consistently expanded with content created and narrated by AI. I can tell because it uses a thumbnail of an actual human's channel, changes it slightly, then basically narrates content the original creator has already made and released, only it does so-VERY BADLY. It misses the point of the original video, mispronounces words ALL THE TIME, consistently misuses words like "pace" instead of "pack", and even the transcripts are so bad they're basically word salad. It seems in the categories of stuff I watch, there are new channels popping up weekly. It is so disheartening. I have taken to subscribing to channels I don't really even watch very often simply because I want to be able to watch the REAL thing when I go to look for it and not get deep fakes. I can't even tell by the "narrator's" voice any more because they have at least a dozen or more different "voices" made to sound like popular narrators. They even have a deep fake of the illustrious, well respected gem Sir David Attenborough!!! I was viewing this one video and at one point I thought "wait, he would NEVER mispronounce that phrase, that is NOT ACTUALLY Sir David Attenborough"!!! That really got me angry, and it pushed my awareness level way up. Now I can pretty much tell.
If we aren't careful...[then again, it may already be too late.. ] pretty soon, all our youth are going to be being educated by word salad spewing AI and nobody will be able to understand a single word anyone says!
An AI voice pronouncing AI as “aye” repeatedly is peak AI.
I've seen AI summaries of youtube comments too, built into youtube. This seems worrisome to me as well for other reasons.
So folks use an A.I to watch and comment on videos?
Leads me to believe some creators will use A.I to make videos. 😳
They already are
@@Barnacl3_Boi
Like damn.😔
@@theturquoisedream9244there’s channels that help people scam more people with deepfakes on RUclips and there’s a bunch of channels that uses nothing burn TTS VO in lieu of hiring VAs for voiceovers instead
You'll see them on RUclips Shorts. A lot of the movie summary/clips, and a bunch of the promotional videos for certain self-help guru personalities catering to adolescent boys, are using AI to churn out profitable pap. I'd be surprised if 5 Minute Crafts isn't already doing it, they've been recycling portions of their proven-incorrect and occasionally lethal videos for years now.
They used TTS to narrate Reddit threads for a while, now they're being sneaky and using AI voices to narrate AI generated summaries of movies. It sucks, don't give them views, watch better videos.
You forgot to mention it fu**s up each of your (or anyone else's) videos' playthrough possibilities. Definitely lowering the time viewers sit through without bailing. They'll just find the (wrong) information they need and move on.
Dear god. This is absolutely insane.
0:00 The video discusses AI
Recap powered by the lowest-effort joke I could come up with
I love your videos, and I don't ever read the comments until after I'm done with the video.
This AI could easily be used to spread Propaganda since words are metaphors- if an AI is summarizing something they'll inevitably insert their own opinion and/or their coder's opinions into it, even if it's not intended.
Ironic those ai recap channel just wanna make me watch those shows even more
Great video. Can’t stand ai and people who keep using it
To be fair, most RUclips videos are mostly fluff
depends on what you're watching!
Watch better videos
Why do I get the feeling that 5 hour alt right videos are free of those dreaded AI "explaining machines"?
We truly live in the weirdest timeline... I'm not a YT or video creator but I feel bad for you. Best wishes on dealing with this mess!
0:00 Introducti- jk in all seriousness AI can be used for good, but its being used way to often or negative reasons. Watching a RUclips video is better than just having an ai recap it. These videos have soul, time, money, etc put in to them.
welp, hot take, i understand creators arent gonna like it... but as a content consumer, instead of wasting 1 hour to see if i'll find the answer in a video, i use theses services.
I use theses at least to know if i will find the answer in it, or if it's actually something i wanna watch. AI is nowhere near perfect to trust blindly, but to summarize ? it's pretty alright, ive yet to see it hallucinate.
It helps against the click bait tittles and thumbnails, soooooooo after 'playing the youtube game' and 'just doing what you gotta do to satisfy the Algo and deceiving people'
Now content creators call wolf on this ? and call this a threat to their livelihood ?
how ironic ( btw i used tammyAI after watching the first 2 min and got a summary of the main remaining points )
The irony of making ai summary videos is that ai is one of the main reasons there's too much garbage drivel content.
This is a problem that will only get worse
Here's a useful AI: One that cuts out all of the redundant crap people pack into videos to extend their runtimes. But before I talk about that, we first need to talk about the development of artificial intelligence....
All the new various AI trends on youtube. I hate them. All of them. So much.
Reminds me of Blinkist that was ruthlessly advertised on RUclips.
This is a fastening topic that I hadn’t heard of yet. I’ve yet to get AI comments like this in my videos, but it seems like it’s going to be a matter of time.
It's amazing how much money there is in taking something good and improving it until it's ruined.
Good video, thanks for making it.
Interesting that you didn't mention RUclipss "Ask" feature on videos, or Google's Bard youtube features that allow the same sort of things you did talk about on this video.
Seems relevant since to me it indicates that we will see more of this.
It's certainly the case Google would love for you to not actually go anywhere other than Google. I just don't know why unless they start putting banner ads or making the first full page of results plugs to different products.
I am as of yet unaware of these features. I’ve not knowingly come across them.
@@PillarofGarbage ah that makes sense. My apologies that was not a fair assumption on my part.
The "Ask" feature was in opt in testing a few weeks ago though seems to be more widely available now. It's available on this video and allows you to have a chat with some mix of Google's models to talk about the content of the video.
As for Bard, if you have the ability to use Bard which is lots of countries you can use the @ symbol followed by RUclips to activate it's RUclips video chat mode, give it a video url and carry on with your conversation about the video in question.
(Bard is an official Google product and can be googled on how to access)
@@user-pd6iy5zm9d I guess they're rolling out Ask on a user-by-user basis, I don't see the option on my end for this video (or any others)
Part of the problem is the amount of 'Content Creators' that seem to exist solely on taking 2min article then reading it in the most drawn out, guff filled, vapid way- usually with bad English.