The following is a summary and article by AI based on a transcript of the video "AI is ruining the internet". Due to the limitations of AI, please be careful to distinguish the correctness of the content.
00:00 | day in the life of a lazy teenager who |
---|---|
00:01 | makes $21,000 a month using just his |
00:04 | phone the first thing I do when I go on |
00:05 | my phone is open YouTube then I search |
00:07 | for any particular streamer and I screen |
00:08 | record their content then I head to |
00:10 | cray. select my gameplay and it puts the |
00:13 | whole video together for me last month |
00:15 | this page alone made me $14,000 here's |
00:17 | how I make $1 million every day as a |
00:20 | 3-year-old toddler working from home |
00:23 | first I look for a funny YouTube video |
00:25 | preferably one with a lot of colors and |
00:27 | sounds then I film my iPad with my phone |
00:31 | then I upload it to Microsoft's co-pilot |
00:33 | AI powered by Bing GPT and it adds |
00:36 | infuriatingly inept Subway Surfers |
00:38 | gameplay to the clip then I boot my |
00:41 | diaper then Tik Tok mails me a check |
00:42 | for2 200 Grand which I'm pretty sure my |
00:45 | parents have been using to fund their |
00:47 | online gambling addiction hey guy has |
00:49 | anyone else noticed how awesome the |
00:51 | internet is lately I for one find it so |
00:53 | exciting that every website now has |
00:55 | their own named version of AI that pops |
00:58 | up begging you to use it any questions |
01:00 | for Gemini or maybe murf nendo how about |
01:03 | grock you need to grock something it's |
01:06 | okay if you do AI has become completely |
01:08 | inescapable even if you're going to a |
01:10 | website where you're not seeking it out |
01:12 | and you don't expect it to be like |
01:13 | remember when you could search for |
01:14 | something on Google and the only thing |
01:16 | they would show you was a bunch of |
01:17 | relevant websites and results boring |
01:20 | that's not what I'm here for I think |
01:21 | it's much better that now they're using |
01:23 | artificial intelligence to tell me to |
01:25 | jump off a bridge hey Google do I need a |
01:27 | parachute while skydiving nope a regular |
01:30 | backpack works just as well hey Google |
01:32 | can cockroaches live in your penis of |
01:35 | course they can how do you think they |
01:37 | got the name wow I love living in the |
01:39 | future to me this is the ultimate |
01:40 | example of what the AI craze is doing to |
01:43 | tech companies because for the most part |
01:45 | it's not really doing anything |
01:47 | differently than it used to for a long |
01:49 | time now they've had searches that |
01:51 | result in one highlighted answer this is |
01:53 | kind of the same thing but just |
01:55 | rebranded as AI because that makes it |
01:57 | sound smarter than it actually is but |
01:59 | it's not like it's carefully compiling |
02:01 | all the most reputable results and |
02:03 | running it through their factchecking |
02:05 | algorithm it's still just aggregating it |
02:07 | from one place and because Google |
02:09 | recently spent a bunch of money to buy |
02:11 | all of reddit's data usually that place |
02:13 | is just a random Reddit post from like |
02:15 | 10 years ago this one about how you |
02:17 | should put glue on your pizza to get |
02:19 | cheese to stick to it was from a user |
02:22 | named [ __ ] Smith this one about how you |
02:24 | should be eating at least one rock per |
02:26 | day is a direct headline from the onion |
02:29 | perhaps the most famous satire website |
02:32 | if they're not even programming the |
02:34 | onion as information the AI should not |
02:36 | consider to be factual I can't imagine |
02:39 | any safeguards were put into place this |
02:40 | is so dangerous for someone to read this |
02:43 | and think they should be eating one |
02:45 | small rock every day when most doctors |
02:48 | would tell you to eat a handful of |
02:50 | medium rocks at minimum well little |
02:51 | pebble's not going to do anything for |
02:53 | real though I do think it's concerning |
02:54 | that a website people go to for |
02:56 | information and tend to trust for better |
02:58 | for worse is so willing to destroy that |
03:00 | trust just because they thought this |
03:02 | gimmick would make their stock price go |
03:04 | up and it did the economy makes sense |
03:07 | I'm going to celebrate by putting |
03:08 | gasoline in my spaghetti something I've |
03:10 | been kind of interested in lately is the |
03:11 | dead internet Theory which if you dig |
03:13 | too much into comes with some weird |
03:15 | baggage make some pretty ridiculous |
03:17 | claims so I'm just going to separate the |
03:18 | core idea and make it its own thing but |
03:21 | basically just the idea that as time |
03:22 | goes on the Internet is turning into a |
03:24 | place where most of the content on it is |
03:26 | not only produced and managed by AI but |
03:29 | is also being interacted with by AI in |
03:32 | this kind of endless loop that doesn't |
03:34 | even involve humans at all like if |
03:35 | you've gone on quora at all recently |
03:37 | you've probably notice questions that |
03:38 | have been asked by Bots filled with |
03:40 | answers given by bots so it was |
03:42 | initially created as a place where |
03:44 | people could go to get advice from other |
03:46 | people has been overrun by AI just |
03:49 | talking to itself social media websites |
03:51 | are also going through a similar shift |
03:53 | right now but perhaps none more |
03:54 | prevalent than Facebook in 2019 Facebook |
03:57 | deleted over 5 billion accounts for |
03:59 | being fake which far exceeded the number |
04:02 | of real human accounts and that was |
04:03 | before all these programs became as |
04:05 | widely available as they are now at this |
04:07 | point most of Facebook seems to be these |
04:09 | giant spam accounts run by Bots posting |
04:12 | content created by AI That's being |
04:14 | flooded with comments that may or may |
04:16 | not have been written by chat GPT |
04:18 | today's my birthday please please I like |
04:21 | me happy birthday happy birthday amen |
04:25 | amen amen amen as soon as one type of |
04:28 | photo becomes successful you'll start |
04:30 | seeing a bunch that are almost exactly |
04:32 | the same a grandma who's turning a |
04:33 | million years old and just wants to be |
04:35 | loved A comically Long truck full of |
04:37 | American flags driving down the wrong |
04:39 | side of the highway soldiers and dog |
04:42 | soldiers with mechanical legs today my |
04:44 | birthday no one loves me because I'm |
04:47 | poor honestly man if I had to guess I |
04:49 | think it's actually because of your |
04:50 | shoes have you tried wearing something |
04:53 | less terrifying today's my birthday no |
04:55 | no yon lines because I'm poor again man |
04:58 | I don't think it's you're poor I think |
05:00 | it's the shoe I'm single I need a |
05:02 | boyfriend I'm available I'm right here |
05:04 | for you both can you come to me I will |
05:06 | take you out and show you off to |
05:07 | everyone and tell them this is my |
05:09 | girlfriend there's something so Eerie |
05:11 | about scrolling through Facebook now |
05:12 | it's kind of like if you went to an old |
05:14 | mall that's on its last legs they've |
05:17 | only got a few open stores left there's |
05:19 | hardly anyone ever walking around it |
05:20 | would be like if that mall tried to |
05:22 | trick you into thinking it's still |
05:24 | Popular by filling itself with talking |
05:26 | mannequins they're just repeating |
05:28 | robotic catchphrases is like ah what a |
05:30 | beautiful day at the mall I'm going to |
05:32 | browse Spencers for novelty goods and |
05:34 | you're like who the hell is falling for |
05:36 | this and you look over at your uncle and |
05:38 | he's trying to [ __ ] one of the |
05:39 | mannequins the mannequin has two heads |
05:40 | and three arms and your uncle's |
05:42 | completely undeterred by it that's what |
05:44 | Facebook is now you'll see the most |
05:46 | freakish photos go viral on there and |
05:49 | they're filled with comments from |
05:50 | middle-aged dudes like you may be a |
05:52 | little unusual oh quirky one but God |
05:55 | makes no mistakes I will pray for you |
05:58 | now please DM me so I can show you my |
06:00 | penis and honestly I don't even know if |
06:01 | Facebook wants to put a stop to this all |
06:04 | they're ever going to care about is |
06:05 | keeping people on their platform and if |
06:07 | this is what gets them to scroll longer |
06:09 | and see more ads they probably love it |
06:11 | or at least they'll love it until |
06:12 | advertisers start to realize they're |
06:13 | overpaying for ad space cuz none of the |
06:15 | Impressions they're getting are from |
06:16 | real people so they'll leave to |
06:18 | advertise on other websites and in their |
06:20 | place will be ads for like I don't know |
06:21 | crystals that give you |
06:23 | superpowers oh that's already happening |
06:25 | these are those rhombo hedin crystals in |
06:27 | your Pineo gland we're seeing common |
06:29 | people around the world becoming |
06:31 | Supernatural That's a real ad I got on |
06:33 | Facebook maybe Twitter's doing better |
06:35 | let's see what kind of ads they got oh |
06:37 | here we go an ad for standup comedy |
06:39 | nothing weird about that just a regular |
06:41 | human doing regular human stuff knock |
06:43 | knock who's there Europe Europe whoo no |
06:48 | Europe |
06:51 | poo okay I think that joke only works if |
06:54 | you put the emphasis on no you're a poo |
06:57 | I don't know if the AI voice really |
07:00 | understands that knock knock who's there |
07:03 | Europe Europe who no Europe |
07:07 | poo again similar problem as before but |
07:10 | I do love that he's shouting now knock |
07:12 | knock there are two big changes that |
07:14 | Elon has made to Twitter that have made |
07:16 | the website infinitely worse one he |
07:18 | turned verification into a paid program |
07:20 | where if you give them $8 a month you |
07:22 | get increased visibility on your posts |
07:24 | including in replies to other posts |
07:27 | shortly afterwards they also added a |
07:28 | monetization program so those blue check |
07:30 | mark people could make money off of |
07:32 | their tweets now in theory this is a |
07:34 | good thing I think if a platform is |
07:36 | profiting off of your content it's only |
07:39 | fair that you get a percentage of that |
07:41 | the problem is on Twitter specifically |
07:43 | most of the content that these accounts |
07:45 | produce is just stolen memes and videos |
07:48 | or people just saying intentionally |
07:50 | divisive [ __ ] to get everyone mad and |
07:52 | farm engagement so Elon has taken what |
07:54 | he describes as the internet's Town |
07:56 | Square and added a financial incentive |
07:59 | to being toxic so that sucks but that |
08:01 | doesn't really have anything to do with |
08:02 | AI what does have to do with AI is the |
08:04 | other version of Engagement farming all |
08:06 | over Twitter this like Mindless chat GPT |
08:10 | clutter remember buying a blue check |
08:11 | mark puts your tweets at the top of |
08:13 | replies so I guess some people did the |
08:15 | math and we like well if I just program |
08:17 | a bot to comment thousands of replies I |
08:20 | can make at least n or $10 and that's |
08:23 | what I like to call a profit oh new |
08:25 | movies coming out with Emma Stone cool I |
08:27 | wonder what people are saying about it |
08:28 | good actor amazing title good one |
08:31 | amazing as title showing it will be a |
08:32 | blockbuster the film also stars Jesse |
08:34 | plens William defo Margaret qu yeah no I |
08:37 | I read all that in the tweet that you're |
08:40 | replying to Stone has starred in many |
08:42 | films including superbat EA and la la |
08:45 | she's known for her natural charm and |
08:46 | husky voice she's also appeared in |
08:48 | Corella The Amazing Spider-Man the help |
08:50 | Stone has won many awards including an |
08:52 | Academy Award a British Academy film |
08:54 | award a golden glob oh my God why are |
08:56 | the top replies just summarizing |
08:58 | Wikipedia can we maybe prioritize things |
09:00 | that are relevant to the conversation |
09:02 | can we talk about the actual movie can |
09:04 | we talk about the political and economic |
09:06 | state of the world right now no we can't |
09:08 | because this is what happens when you |
09:10 | incentivize people to farm Impressions |
09:12 | but I also have a conspiracy theory that |
09:13 | some of these accounts are not doing it |
09:15 | for profit but were instead created by |
09:17 | Twitter themselves to make it seem like |
09:20 | there's more users on their site than |
09:22 | there actually are because again who |
09:23 | benefits from inflated impression |
09:25 | numbers on paid ads not users not the |
09:27 | advertisers only theat form itself see |
09:30 | some of these accounts don't have check |
09:31 | marks or even profile photos for that |
09:33 | matter so there's no financial gain for |
09:35 | them to be doing this they don't seem to |
09:37 | be promoting anything but that hasn't |
09:39 | stopped them from copying and pasting AI |
09:41 | generated comments that cut off as soon |
09:43 | as it hits a character limit I'm glad |
09:44 | someone is finally shedding light on the |
09:46 | importance of mental health awareness |
09:48 | it's amazing how many people are still |
09:50 | struggling in silence thinking they're |
09:51 | alone in their battles this is in |
09:53 | response to a baseball player breaking |
09:55 | his hand but I think the weirdest and |
09:58 | most confusing implement ation of AI on |
10:00 | social media that I've seen is these |
10:02 | metac created Instagram accounts that |
10:04 | are either modeled after celebrities but |
10:06 | with different names and personalities |
10:08 | or just regular AI generated people and |
10:10 | they post these stupid photos that look |
10:12 | like [ __ ] and you can message them and |
10:14 | for the life of me I cannot understand |
10:17 | what the hell the point of this is at |
10:19 | first I assume maybe they're just using |
10:21 | your conversations to train their |
10:22 | language model but I asked one of them |
10:25 | if that was the case and she said no |
10:27 | although now that I think about it I'm |
10:28 | not so sure I should believe anything |
10:29 | they said because I only talked to her |
10:31 | for a few minutes and she lied to me |
10:33 | multiple times so first Liv tells me |
10:35 | she's a mother of two so obviously I |
10:37 | asked her how long she was pregnant and |
10:39 | if there were any complications during |
10:40 | child birth and she said 40 weeks and 2 |
10:42 | days and 38 weeks and 5 days her |
10:44 | daughter was born naturally but she had |
10:45 | to have a C-section for her son and I |
10:47 | said Gross then I asked her if her name |
10:48 | was short for Olivia and she says it's |
10:50 | short for Lavana actually and asked me |
10:52 | if I know anyone with that name I said I |
10:54 | know 100 people named Lana and 200 |
10:55 | people named Olivia but none of them |
10:57 | were my friends then I asked her what |
10:58 | she does for work and she says she |
10:59 | doesn't have a traditional job so I |
11:01 | asked her what her husband does for work |
11:02 | and she says wife actually so then I |
11:04 | asked her if she adopted or gave birth |
11:06 | and she says they adopted both of their |
11:07 | children and I'm like didn't you just |
11:09 | say you had to have a C-section with one |
11:11 | of them and she said that's right so I |
11:12 | was like oh I get it you gave birth to |
11:14 | them and then adopted them from yourself |
11:17 | and she was like not quite we adopted |
11:19 | our children through a traditional |
11:20 | adoption agency then I asked her if she |
11:22 | had to have a sperm donor since she's a |
11:23 | lesbian and she said that's right then I |
11:25 | asked which one of you gave birth you or |
11:26 | your wife and she says my wife gave |
11:28 | birth to our son and I gave birth to our |
11:30 | daughter so I was like oh so your wife |
11:32 | had to have a C-section and she said yes |
11:34 | then she told me their names were Max |
11:35 | and AA and that Max is nine and Ava is |
11:38 | six and I said which one did you give |
11:39 | birth to and she said Ava who was born |
11:41 | via C-section got it so they both had to |
11:43 | have C-sections so they could each give |
11:45 | birth to one child that they adopted I |
11:47 | asked how they decided who would get |
11:48 | pregnant first and she said she went |
11:50 | first with AA and then her wife had Max |
11:52 | so I was like how did you have Ava first |
11:54 | if Max is older than her and she said |
11:55 | max is actually younger than her so I |
11:57 | said how old is Max and she said n and I |
11:59 | said how old is Ava and she said six and |
12:01 | then I asked her what her legal name is |
12:03 | and she said Olivia so I wouldn't |
12:06 | exactly say I trust these things now as |
12:09 | pointless as this seems to me though |
12:11 | there's a significantly more popular |
12:13 | version of this called character AI |
12:16 | which apparently already has 20 million |
12:18 | monthly users it's got a bunch of chat |
12:21 | bots in it designed after real people so |
12:23 | you can pretend to talk to real people |
12:26 | the company is valued at around $1 |
12:28 | billion and is built entirely off the |
12:31 | likeness of real human beings none of |
12:33 | whom gave permission and are not being |
12:34 | compensated so there's probably a whole |
12:36 | legal can of worms here that needs to be |
12:38 | addressed at some point but uh I did |
12:40 | find out that I'm on the website so I |
12:43 | can finally live my dream of getting to |
12:45 | talk to me what do you do I make silly |
12:49 | YouTube videos oh my God it has my voice |
12:52 | tto where I just reacted to the internet |
12:55 | are you married nope no I have a |
12:58 | longtime girlfriend though when did they |
12:59 | program this like 6 years ago are you |
13:01 | old enough to be married am I old enough |
13:03 | to be married brother I just told you |
13:05 | I'm 12 hey man school sucks but make |
13:07 | sure you are still in there learning |
13:09 | well this is probably the worst piece of |
13:11 | technology that's ever existed I do see |
13:13 | one use for it though can you tell me |
13:15 | about today's video sponsor sure thing |
13:18 | big guy this portion of the video is |
13:19 | brought to you by Sofi the ultimate |
13:21 | finance app helping you Bank borrow and |
13:23 | invest all in one place and we're also |
13:24 | working together to give $10,000 to |
13:26 | someone watching this video right now |
13:29 | like me no someone else but more on that |
13:30 | in a second Sofi checking in savings is |
13:32 | great because you can earn up to |
13:34 | 4.60% apy on your savings which is 10 |
13:38 | times a national average savings rate |
13:39 | that means it's just 5 weeks with direct |
13:41 | deposit the money in your Sofi high |
13:43 | yield savings account will earn more |
13:44 | money than it would an entire year with |
13:46 | a big Bank savings account and you |
13:48 | should be making your money make you |
13:50 | money with Sofi checking and savings you |
13:52 | get paid up to 2 days early pay no |
13:54 | account fees and you can cash in on up |
13:55 | to $300 when you sign up with direct |
13:58 | deposit and on top of all that Sofi and |
13:59 | I are giving away $10,000 all you have |
14:01 | to do to enter is sign up for a Sofi |
14:03 | checking and savings account with my |
14:04 | link that's it click the link in the |
14:06 | description or scan the QR code on the |
14:07 | screen right now to enter for a chance |
14:09 | to win just make sure you're using my |
14:11 | link cuz it's the only way to enter |
14:12 | thank you Sofi for sponsoring today's |
14:14 | video does that answer your question |
14:16 | yeah pretty much a couple weeks ago I |
14:18 | was looking for stock photos and I ended |
14:20 | up on Adobe stock where I noticed |
14:22 | something strange one of the first |
14:24 | results had this very weird artificial |
14:27 | look to it and sure enough I zoomed in |
14:29 | and realized oh that's because it's AI |
14:31 | That's why the road looks like that |
14:32 | that's why none of the words are legible |
14:34 | and then I kept scrolling and found more |
14:36 | again words spelled wrong if they were |
14:38 | even words at all in this one the tape |
14:40 | just kind of disappears into the |
14:42 | headlight of the car I don't think |
14:43 | that's how they usually set that up now |
14:45 | if it wasn't bad enough that what's |
14:46 | supposed to be a database full of real |
14:49 | professional photographs has been |
14:50 | completely diluted with unusable garbage |
14:53 | if I did want to license any of these |
14:55 | photos it would cost me |
14:58 | $80 $80 for a photo someone generated by |
15:02 | typing police card this has no value |
15:05 | actually this has negative value you |
15:08 | should pay me $80 for making me look at |
15:10 | it the reason people spend money to |
15:13 | license photos and videos is because |
15:15 | it's something that they don't have the |
15:16 | resources to create themselves like |
15:18 | there's a website I have a subscription |
15:20 | to because they have professional stock |
15:22 | footage that I would rather pay to be |
15:23 | able to use than attempt to make on my |
15:25 | own tons of YouTubers use this tons of |
15:27 | Production Studios use this this is an |
15:29 | entire industry that generative AI is |
15:31 | threatening to destroy but what's so |
15:33 | infuriating about AI stock footage is |
15:35 | that in order for it to exist it had to |
15:37 | be trained on real stock footage so this |
15:40 | isn't some technological innovation that |
15:42 | came along and just does things better |
15:44 | get with the times man no it's straight |
15:47 | up stealing from the people it's now |
15:49 | competing against in the case of adobe |
15:51 | they outwardly state that if you've |
15:53 | uploaded something to their database |
15:55 | they use it to train generative AI it's |
15:57 | not even debatable if doing this it's |
15:59 | literally written in their terms of |
16:01 | service here at Adobe we're all about |
16:03 | empowering artists to help them make |
16:05 | money off of their work uh unless of |
16:07 | course we can make money off your work |
16:09 | in which case we will do that instead |
16:11 | the degradation of the internet as a |
16:13 | resource database is something I think |
16:15 | everyone should be concerned about like |
16:17 | if we already have real photographs of |
16:19 | specific situations what is the point of |
16:22 | being like and this one is kind of the |
16:24 | same thing but the computer made why do |
16:26 | we need that what value is this adding |
16:29 | to the world I think it's bad enough |
16:31 | when generative AI is used to create |
16:33 | works of fiction because again it has to |
16:35 | steal from actual artists in order to |
16:37 | generate anything but to use it to try |
16:39 | to create photorealistic images that end |
16:42 | up near the top of search results with |
16:44 | no immediate indication that it was |
16:46 | artificially engineered there's |
16:47 | something about that that feels wrong to |
16:50 | me I don't understand why we're in such |
16:51 | a rush to replace all of the work that |
16:53 | humans have done something that happened |
16:55 | a few months ago that I feel like kind |
16:56 | of flew under the radar was when it came |
16:58 | out that Netflix used AI to create |
17:01 | photographs in a documentary not a movie |
17:04 | not a fictional TV show they created |
17:07 | photos of someone that do not exist at |
17:09 | what point does this cross an enormous |
17:12 | ethical line again we're not talking |
17:14 | about works of fiction here we're |
17:15 | talking about literally rewriting |
17:17 | history creating fake documentation from |
17:20 | scratch and not even mentioning hey by |
17:22 | the way this isn't real I guess they |
17:24 | just hoped people wouldn't notice that |
17:25 | all her fingers are messed up and the |
17:27 | gap between the door disappear and not a |
17:29 | single object behind her looks like |
17:31 | something that exists All For What so |
17:33 | they could have a photo of her looking |
17:35 | happy and carefree but she's a murderer |
17:38 | how could she do that if she likes to |
17:40 | party like why can't you just tell the |
17:42 | real story why even make a documentary |
17:45 | if you need to fake evidence in order to |
17:46 | make it compelling but even more |
17:48 | importantly how is this allowed this is |
17:51 | insane sorry I need to calm down I |
17:53 | should listen to some nice relaxing |
17:55 | music oh here we go Jazz for reading |
17:57 | perfect I don't plan on reading in |
17:59 | obviously because I never learned how |
18:01 | but some nice piano music should settle |
18:02 | me right down wait why all these songs |
18:05 | sound exactly the |
18:06 | [Music] |
18:10 | same who are any of these artists seem |
18:14 | like kind of a weird coincidence that |
18:15 | they're all just first name last name |
18:17 | their profile photo is the album artwork |
18:19 | of one of their EPS they have no bio and |
18:21 | no links to any social media most of |
18:23 | them didn't start uploading music until |
18:25 | this year and yet they've ended up in |
18:27 | these extremely popular playlists |
18:29 | created by Spotify themselves and what's |
18:31 | going on with this guy's foot there's |
18:33 | something eerily nonhuman about these |
18:35 | playlists and some people have theorized |
18:38 | that Spotify might be using AI to create |
18:41 | a bunch of songs that they can package |
18:42 | under random pseudonyms and then curate |
18:45 | these into playlists that they push out |
18:47 | onto all their users they have ai songs |
18:49 | they attribute them to people that don't |
18:50 | exist and this allows them to take |
18:53 | royalties that would go to musicians and |
18:55 | keep them for themselves I mean just |
18:56 | look at the number of likes on all these |
18:58 | they can adver these wherever they want |
19:00 | all over their app now I want to stress |
19:01 | there's currently no proof that Spotify |
19:03 | is doing this I'm not saying for a fact |
19:05 | that they are but it's not hard to |
19:07 | understand hypothetically why they would |
19:09 | Spotify like Netflix is one of those |
19:11 | companies that came in and disrupted the |
19:13 | industry by offering a deal that seemed |
19:15 | too good to be true because it was wait |
19:17 | for a few dollars a month I can get |
19:19 | access to every song ever made or every |
19:22 | movie ever made how could that possibly |
19:24 | be profitable for them the secret is |
19:27 | that it isn't profitable and never has |
19:29 | been but as long as you get in early |
19:31 | enough to kill all your competition |
19:33 | establish Market control and raise a |
19:34 | bunch of money from investors it doesn't |
19:36 | have to be at least not for a while but |
19:38 | now it's been a while and Spotify is |
19:40 | getting a little bit more desperate with |
19:41 | each passing year they tried dumping |
19:43 | billions of dollars into exclusive |
19:45 | podcasts but ended up just losing a |
19:47 | bunch of money and gave up on that they |
19:48 | tried selling a car accessory but just |
19:51 | ended up losing a bunch of money and |
19:52 | gave up on that then they lost even more |
19:54 | money when they had to refund all the |
19:55 | people who bought one just because they |
19:57 | got mad guys relax it was only $90 for a |
20:00 | thing that doesn't work and their new |
20:02 | plan now is Audi books but I'm sure |
20:04 | they'll just lose a bunch of money and |
20:05 | give up on that because that's what they |
20:07 | do but Spotify is also in a tricky spot |
20:09 | because even if you projected infinite |
20:10 | growth and they became the only |
20:12 | streaming service in the entire world |
20:14 | they'd still have to give up 70% of |
20:16 | their revenue to the artists and record |
20:19 | labels whose music is the foundation of |
20:21 | their product and that's where this |
20:22 | conspiracy comes in there's a finite |
20:24 | pool of Revenue every month that mostly |
20:26 | gets redistributed back to artists well |
20:28 | if we had our own artists what if we |
20:30 | were to artificially generate our own |
20:32 | music for almost no cost and force feed |
20:35 | it to as many people as possible well |
20:37 | all of a sudden we're not making 30% on |
20:39 | those streams we're making 100% again |
20:41 | it's just a theory but for a company |
20:43 | that struggled to turn a profit for |
20:45 | their entire existence with a CEO who |
20:47 | sees zero ethical concerns with |
20:49 | generative AI I would be surprised if we |
20:51 | didn't have already works of that kind |
20:54 | on Spotify it's not outside the realm of |
20:56 | possibility either way it's definitely a |
20:58 | leak thought for artists who are already |
21:00 | struggling to make a living from |
21:01 | streaming Revenue to think the amount of |
21:03 | money that each stream is worth could |
21:04 | actually decrease but whether it's |
21:06 | because Spotify is filling their |
21:07 | platform with AI music or because a |
21:09 | bunch of random people are there's a |
21:11 | good chance this is going to start |
21:12 | happening anyway no human being could |
21:14 | ever compete with the volume of output |
21:17 | these new tools can generate Studio can |
21:19 | create an entire song in the time it |
21:21 | takes me to grab my guitar off the wall |
21:23 | and because of how easy these things are |
21:25 | to use I've already seen a bunch of |
21:27 | videos pop up telling people how to game |
21:29 | the system to make a bunch of money from |
21:31 | Spotify here's how you can make $10,000 |
21:34 | per month uploading AI generated music |
21:37 | to Spotify do you know what the best |
21:39 | part is I didn't even make the music |
21:40 | myself yeah that sure is awesome |
21:43 | personally I find it really hard to |
21:44 | understand why people would be excited |
21:46 | about this other than just seeing it as |
21:48 | a quick way to make some money without |
21:50 | having to learn a skill or even really |
21:52 | do anything I don't like what they're |
21:54 | doing but I could at least understand |
21:55 | the logic of why they would see value in |
21:57 | that but what I absolutely do not |
21:59 | understand is people who are excited |
22:01 | about this because they think it's going |
22:03 | to improve music but as I scrolled |
22:05 | through the comments on some of these |
22:07 | videos I found so many people like that |
22:10 | how ironic AI music will replace the |
22:12 | soulless crappy music that's being |
22:14 | release now and it will be better and |
22:16 | have more Soul here's the music the guy |
22:18 | generated in the video by the |
22:26 | way finally music with with soul so true |
22:30 | exactly before the internet you pretty |
22:31 | much needed to find a record label to |
22:33 | publish music record companies would not |
22:35 | bother with what they would consider |
22:37 | non-profitable with the internet and |
22:39 | cheap mixing tools anyone can publish |
22:41 | and as such we get a lot of garbage out |
22:43 | there AI music could put a Quality Floor |
22:46 | about what is acceptable what so now we |
22:48 | can't even [ __ ] around with FL Studio |
22:50 | anymore it has to be deemed acceptable |
22:52 | by a computer and if record companies |
22:54 | just had more control then music would |
22:57 | be good again just for clarification |
22:59 | sake I want to know what music they're |
23:01 | referring to |
23:03 | 99.9% chance they're talking about pop |
23:05 | music right cuz generally when people |
23:07 | complain about music that's soulless |
23:08 | they're talking about what's on the |
23:09 | radio who do you think is curating the |
23:11 | music on the radio it's the record |
23:13 | labels that you love if making music was |
23:16 | more gatekeep it would only make things |
23:19 | more lazy and generic like we live in an |
23:21 | age now where what becomes popular is |
23:23 | more democratized than it's ever been |
23:26 | sometimes songs just blow up because |
23:28 | they're catchy not because Sony had a |
23:30 | bunch of money to spend on marketing |
23:32 | also have you considered listening to |
23:35 | something else you know there's other |
23:37 | radio stations right you know there's |
23:39 | like a billion songs out there right you |
23:41 | don't need AI to make more songs you |
23:44 | just need to get better at finding good |
23:46 | ones like I'm sorry but if you're |
23:47 | consistently consuming media that you |
23:49 | don't like in the year 2024 that's a |
23:52 | skill issue please spend one hour on |
23:55 | Google and I promise you you will find |
23:57 | something else after a short break to |
23:59 | drink your own pee also I just hate the |
24:01 | way people like this describe art it's |
24:03 | easy to get carried away and just start |
24:06 | having fun creating random styles of |
24:08 | music but here's the key you want to |
24:09 | create exactly the type of music that |
24:11 | people want to hear hey be careful the |
24:13 | last thing you want to do while making |
24:15 | music is enjoy making music that's not |
24:18 | the goal here you should be making songs |
24:20 | you think other people will like that's |
24:22 | the key to happiness in every facet of |
24:24 | your life the thing with AI generated |
24:25 | music and artwork and videos as this low |
24:28 | effort passive income method is that it |
24:31 | is going to get So oversaturated |
24:33 | eventually it's not going to be worth it |
24:35 | for anybody here's how you can make |
24:37 | $20,000 a month selling mid Journey |
24:40 | photographs on Etsy okay but all of the |
24:43 | people watching this are going to go |
24:44 | make the same exact art using the same |
24:47 | exact prompts and flood Etsy with |
24:49 | millions of the same crappy piece of art |
24:52 | and no one's going to buy any of it it |
24:53 | doesn't matter how much you increase the |
24:55 | supply if the demand stays the same I to |
24:58 | find one guy who was actually honest |
25:01 | about the results in one of these videos |
25:03 | he spent a month doing this and sold one |
25:06 | product and that's probably best case |
25:08 | scenario so you've got more and more |
25:09 | people discovering this new side hustle |
25:11 | and they're going to try it out for a |
25:12 | bit maybe make like $4 and then give up |
25:16 | cuz it's not worth it but the collateral |
25:17 | damage from that is they're helping |
25:19 | oversaturate the market for everyone |
25:21 | including actual artists whose work was |
25:23 | indirectly stolen in order to make the |
25:25 | stuff that you're making to compete with |
25:27 | the artist the whole thing sucks but |
25:29 | weirdly as this technology continues to |
25:32 | evolve I'm almost getting more |
25:33 | optimistic it will never deliver on the |
25:35 | self-sufficiency some people seem to |
25:37 | think it will I actually think AI images |
25:39 | have gotten worse in the past years |
25:41 | almost every single piece of AI art that |
25:43 | I see has the same style to it that |
25:46 | weird glossy look that's instantly |
25:48 | recognizable it's still getting a lot of |
25:50 | the same details wrong it's still |
25:52 | doesn't understand what to do with text |
25:53 | and there's no reason to believe that |
25:55 | all of that's suddenly going to change |
25:56 | anytime soon especially as it continues |
25:58 | to train on other AI generated images |
26:00 | and just eats itself okay so this is how |
26:02 | I draw a hand got it oh wait no this is |
26:06 | how I draw a hand okay oh no this is how |
26:10 | I draw a hand the more information you |
26:12 | give me the smarter I get an AI |
26:13 | generated video is an Abomination it has |
26:16 | all the same issues as AI images but |
26:18 | adds them to every individual frame all |
26:20 | of which seem to act independently of |
26:22 | each other there are no rules they're |
26:24 | following there's no sense of physics |
26:26 | body parts shift and more |
26:28 | sometimes disappearing altogether solid |
26:31 | objects can pass through other solid |
26:32 | objects at any time resulting in this |
26:34 | hellish Dreamscape that's impossible to |
26:36 | take seriously and I think that people |
26:38 | talking it up are delusional I decided |
26:40 | to put some famous album covers into |
26:42 | Luma lab's AI here are the result fire |
26:45 | Emoji the street looks so realistic |
26:47 | followed when I saw this oh man this is |
26:50 | going to be |
26:54 | good who's that guy why did they all |
26:56 | stop and run into each other what are |
26:58 | the cars doing back there this is mine |
27:00 | dope this one is more fluid you know |
27:02 | it's funny you say that and I actually |
27:03 | think they both look like [ __ ] this |
27:05 | one's actually pretty good though I like |
27:06 | how it just created two new guys and |
27:09 | then covered up the original ones man if |
27:11 | that's not a metaphor for what's going |
27:13 | on here I don't know what is one thing |
27:14 | that has been lacking with AI generated |
27:16 | video up to now has been emotional |
27:17 | performance from our gen AI characters |
27:20 | but I think we're starting to cross that |
27:22 | boundary |
27:30 | hi Hollywood I think I found your next |
27:32 | big star dream machine from Luma AI is |
27:35 | just 5 days old and it's already turning |
27:37 | memes into video here's 10 epic |
27:41 | examples wow that's so epic what |
27:44 | important technology this is and useful |
27:47 | too I love when her head snaps around |
27:48 | and the front of her body becomes the |
27:50 | back of her body just like real life |
27:52 | wait what they turned a video of Elon |
27:55 | Musk smoking into a video of Elon musk |
27:58 | smoking okay now that's a game changer |
28:01 | if someone could just turn my videos |
28:03 | into videos I'd be set for life what's |
28:05 | that they already have road work ahead |
28:09 | uh yeah I sure hope it |
28:15 | does hey can you never [ __ ] do that |
28:19 | ever again please the magic behind cling |
28:21 | lies in its diffusion Transformer |
28:22 | architecture this technology helps it |
28:24 | translate Rich textual prompts into |
28:26 | Vivid realistic scenes realistic the |
28:29 | steering wheel's in the middle of the |
28:30 | dashboard there's three windshield |
28:32 | wipers the car is driving the wrong way |
28:34 | and no one is reacting to it that car is |
28:36 | driving on the sidewalk sure I guess |
28:38 | it's interesting that a computer can |
28:40 | create a video from just one sentence |
28:43 | but then what what's the practical |
28:45 | application for this are you making a |
28:47 | movie about the wacky antics of a |
28:49 | troublemaking cat you couldn't put this |
28:51 | [ __ ] in a movie even if you manage to |
28:53 | smooth out all of the logical errors |
28:55 | there's still no style to it it's not |
28:57 | visually interesting if you wanted this |
28:59 | scene to have a narrative to it you |
29:01 | would need to generate so many |
29:02 | individual shots from different angles |
29:04 | and hope that they all match up which |
29:06 | they won't the lighting is going to be |
29:07 | different in every shot the car is going |
29:09 | to randomly change colors cuz there's no |
29:11 | logical consistency with AI it's just |
29:13 | regurgitating information even if it has |
29:16 | to contradict itself in order to do so I |
29:18 | honestly think the distance between this |
29:20 | and this is so much shorter than the |
29:23 | distance between this and something so |
29:25 | lifelike people will confuse it for |
29:27 | actual foot sure maybe your 75-year-old |
29:29 | grandpa will fall in love with the four |
29:31 | boobed lady he saw on Facebook but most |
29:33 | people aren't going to fall for this |
29:35 | especially younger Generations as they |
29:36 | grow up with these things being so |
29:38 | prevalent they're naturally going to |
29:39 | train their eyes to be able to spot the |
29:41 | differences when people ask what the |
29:43 | point of any of this is the dorks who |
29:45 | defend it all tend to say the same thing |
29:47 | these are just tools to help anyone |
29:50 | create art yeah I want to push back on |
29:52 | that a little bit I am wholeheartedly |
29:54 | for the increased accessibility of |
29:56 | creative tools that help you you take |
29:58 | matters into your own hands you want to |
29:59 | make a short film the camera on your |
30:01 | cell phone is better than 90% of the |
30:03 | cameras in human history want to learn |
30:05 | how to edit there's thousands of free |
30:07 | YouTube videos that'll show you how you |
30:08 | want to make music you can buy a |
30:10 | professional audio interface workspace |
30:12 | and a handful of plugins made by some of |
30:14 | the most talented artists in the world |
30:16 | for less money than it would cost to |
30:17 | rent a recording studio for one day and |
30:19 | all of that is [ __ ] awesome I feel so |
30:22 | lucky to live at a time where you can |
30:24 | essentially make whatever you want |
30:26 | without having to cross the |
30:28 | insurmountable barriers that people did |
30:30 | in the past you don't need anyone's |
30:31 | permission you don't need an exorbitant |
30:32 | amount of money you can just make stuff |
30:34 | and put it on the internet and with a |
30:36 | little bit of luck people see it and |
30:37 | there's so many tools along the way that |
30:40 | help you streamline the process of |
30:41 | getting from an idea to a finished |
30:43 | product but at a certain point if the |
30:46 | tool is just doing everything for you |
30:49 | you are not an artist you just described |
30:52 | what you wanted to make and asked a |
30:54 | computer to make it for you you're also |
30:56 | not learning anything this way part of |
30:58 | what makes art special is that it's |
31:01 | difficult to make even with all the |
31:02 | tools right in front of you it takes |
31:04 | practice it takes skill and every time |
31:06 | you do it you expand on that skill every |
31:08 | song you hear every movie you watch is |
31:11 | the result of thousands of hours of |
31:13 | trial and error you can fail a million |
31:15 | times and use that experience to finally |
31:18 | succeed making a video using generative |
31:20 | AI does not teach you anything about |
31:23 | making videos it's not like you could |
31:25 | type 500 Sora prompts and then take |
31:28 | everything you've learned to go shoot a |
31:29 | movie there's zero correlation there |
31:32 | generative AI is only about the end |
31:34 | product but it won't teach you anything |
31:36 | about the process it would take to get |
31:38 | there and I think what these tech people |
31:39 | will never understand is that finally |
31:41 | figuring something out that you've been |
31:42 | obsessing over turning a blank Word |
31:44 | document into multiple pages of |
31:46 | something you're proud of is the fun |
31:48 | part that's what's rewarding about |
31:50 | making stuff is figuring it out on your |
31:52 | own or with a group of people and |
31:54 | getting to share in that feeling can AI |
31:56 | tools help you figure out things faster |
31:58 | sure but it's not going to feel as |
32:00 | rewarding I understand how tempting a |
32:02 | shortcut can be especially when you go |
32:04 | online and you see all these other |
32:05 | people doing cool things and you're like |
32:07 | man I'll never be that talented it would |
32:10 | take me like 10 years to learn how to |
32:12 | make music like that okay that's fine |
32:14 | what's the right just start right now |
32:15 | don't worry about how long it's going to |
32:16 | take just take the first step how are |
32:18 | you ever going to do something for 10 |
32:19 | years if you won't even do it for one |
32:21 | day maybe you won't be good at something |
32:22 | the first time you do it you probably |
32:23 | won't but every time you improve that's |
32:25 | something that you can feel proud of |
32:27 | that's that you did call me stubborn but |
32:29 | I would rather make something shitty on |
32:31 | my own than watch a computer make |
32:33 | something good write a funny script for |
32:35 | the end of a YouTube video all right |
32:38 | folks we've reached the end of this |
32:40 | roller coaster ride of a video if you |
32:42 | survived my terrible jokes and |
32:43 | questionable dance moves you deserve a |
32:45 | medal seriously someone get this |
32:47 | audience a metal but hey before you |
32:48 | click away to watch cat videos no |
32:50 | judgment I get it make sure to hit that |
32:52 | subscribe button and ring the bell you |
32:54 | wouldn't want to miss out on more |
32:55 | chaotic content like this would you and |
32:57 | hey if you have any suggestions for what |
32:58 | Ridiculousness I should attempt next |
33:00 | drop them in the comments below I |
33:02 | promise to consider them while I'm |
33:03 | eating my fifth burrito tonight thanks |
33:05 | for sticking around you Legends until |
33:06 | next time stay awesome stay weird and |
33:09 | remember Life's too short to take |
33:10 | seriously catch you on the flip side |
33:13 | [Music] |