- 2 hours ago
- #ai
- #music
- #technology
- #future
- #entertainment
Dive into the wild world of AI and music with us! We chat about AI-generated songs, classic hits, and the new symbiosis between tech and musicians. Is the future of music synthetic or soulful? Find out, laugh a little, and get inspired! Subscribe for more fun takes on tech, and let us know in the comments what blew your mind most from this video! #AI #music #technology #future #entertainment
👉 This channel was created in collaboration with https://www.youtube.com/@premiumhifichannel
0:00 - Introduction: AI and Music
1:25 - Understanding AI: Myths and Learning
4:13 - AI in Controlled Environments and Medicine
9:19 - AI, Music Production, and Synthesizers
12:47 - Roles in Music Creation and AI's Impact
19:35 - AI Performance, Old Melodies, and Financial Models
26:49 - Professionals, AI Collaboration, and Creativity
33:09 - Conclusion: Synergy, Warnings, and Final Thoughts
👉 This channel was created in collaboration with https://www.youtube.com/@premiumhifichannel
0:00 - Introduction: AI and Music
1:25 - Understanding AI: Myths and Learning
4:13 - AI in Controlled Environments and Medicine
9:19 - AI, Music Production, and Synthesizers
12:47 - Roles in Music Creation and AI's Impact
19:35 - AI Performance, Old Melodies, and Financial Models
26:49 - Professionals, AI Collaboration, and Creativity
33:09 - Conclusion: Synergy, Warnings, and Final Thoughts
Category
🤖
TechTranscript
00:00Today I listened on YouTube to that song written by artificial intelligence and you know that song captivated me.
00:06It was performed and sung so gorgeously and professionally and the visuals were so captivating that I was simply stunned.
00:14Is this really possible?
00:16Does artificial intelligence really do this?
00:19Our topic today is artificial intelligence and music.
00:23Please tell me, what is generative AI?
00:25In Russian it's ISKIN artificial intelligence.
00:29So basically we'll be talking about artificial intelligence and its impact on music.
00:35But two.
00:39Talk about artificial intelligence.
00:42We still need to define what artificial intelligence is.
00:46Since right now in the media space, it's a nightmare.
00:50A normal person can't make head or tail of it.
00:54And on the one hand they say Skynet is coming.
00:57We're all screwed.
00:58No one will have a job.
00:59It's terrible.
01:00It's just awful.
01:02So here's my question to you.
01:06Which one do we pick?
01:07Do we choose this pessimistic view of the future?
01:09Or do we look at the future as an incredible opportunity?
01:15I look at the opportunities.
01:17Well, what does pessimistic mean anyway?
01:19What does pessimistic mean?
01:21We look to the future with optimism.
01:25All right.
01:25Then I have a story.
01:27I'm a huge fan of sci-fi.
01:29Me too.
01:31I read everything there was in my native language.
01:34Everything in Russian and almost everything in English.
01:36Thanks.
01:37Have you read Frank Herbert's Dune?
01:39Yes.
01:39Now that's top tier.
01:41Top tier.
01:42Thanks.
01:43Yeah.
01:45Sometimes when we look back at Jules Verne.
01:48Yeah.
01:49I mean, how did he manage to back then predict the future like that?
01:54Yeah.
01:55Yeah.
01:56And I also read, well, I don't remember who wrote it, but there was this interesting sci-fi
02:02story where a super civilization before us had built this specific road across the universe
02:08and you could travel across planets through a vacuum in some places, in some special vehicles
02:14in others.
02:16And someone decided, we want to ride this road all the way to the end.
02:20Well, there are a lot of adventures.
02:22They reach the end.
02:23And there is the king.
02:25An entire planet with this super civilization, humans didn't even exist yet.
02:30They arrive.
02:31And this AI that was there wakes up.
02:38Welcome.
02:39I am so glad to see you here.
02:43Would you like to design something or build something?
02:48Let's make a car.
02:53Artificial intelligence.
02:59Well, an infinite one is theoretically impossible.
03:04But if, say, it has a service life of 1,500,000 years, would that suit you?
03:10Yes, that suits us.
03:12Well, I also want it to be airtight, able to run without without air for a very long
03:16time.
03:19And I also want, I want it to be so that nobody could steal it.
03:23It only worked with me.
03:25I also want it to never need refueling.
03:29Oh, no.
03:30Well, I also want to be able to drive several million kilometers without needing to refuel
03:36and then it would charge itself.
03:39Well, and whoever makes it, it turns out they have a beautiful car.
03:42And they take this car, artificial intelligence.
03:46It was such a pleasure working with you.
03:50A real pleasure.
03:51It's been a long time since I had such an interesting task.
03:56I see the whole trend in artificial intelligence as symbiosis.
04:03It won't be able to come up with what kind of car needs to be made, but I can.
04:08But there are many things, let's say, purely technical that I can't solve unless someone helps me.
04:14That is artificial intelligence and the modern, if you look at modern artificial intelligence.
04:23Look at a human when he is born.
04:25Well, what can you do?
04:26He can like nurse.
04:27And when he feels uncomfortable, he cries.
04:30In general, he's very, very weak.
04:34Why couldn't he be born with some decent skills, you know, being able to run right away or something?
04:40But if you look at it that way, it's a very clever scheme because it's adaptive.
04:47If I'm born like Mowgli in the jungle, I installed the Mowgli software, Mowgli skills.
04:55Born in a city, in civilization, yes, I learned different things.
05:00And how does the learning process happen?
05:03I don't know. Have you ever played basketball?
05:07Basketball is in my game.
05:08After getting hit right in the face with that heavy ball, I lost any desire to play.
05:14So...
05:14Well, I played basketball, volleyball.
05:17And I remember, as a kid, how it goes.
05:20You're just a little kid, you show up, they give you a ball.
05:25And the coach says, back then, no one was allowed to shoot.
05:28It was all dribbling.
05:30And he's yelling, don't look at the ball.
05:33Yeah, like, back straight, knees bent, bent, see the chord?
05:37And you're dribbling there, three dribbles and it's gone, three dribbles and it's gone.
05:43Then the left one is the exact same thing.
05:45And after some time passes, your dribbling starts working out.
05:49And throughout this whole process, you don't do any math calculating what you're going to do.
05:54The algorithm is different.
05:58And basically, when they made this, well, looked at how the brain is wired, it became possible to make an
06:04electronic equivalent of a neuron.
06:07That's the key, the key to everything.
06:09Pretty simple.
06:11And when they make this, this neuron, you can get such a system, a kind of two-port network, where
06:19it has many inputs, it has some outputs, and you can, well, some converters, you speak into a microphone code.
06:29It turns into electronics, well, into an electronic signal, goes to the output.
06:38There are many outputs, and at the output, it turns out, at first, nothing comes out.
06:45And you tweak these weighting coefficients, many times you get this really bad code, like you showed the real one,
06:53the error is subtracted, fed back,
06:55and after many, many iterations, you get it, you say code, and all those weighting coefficients have changed there, and
07:03there is no mathematical algorithm.
07:06Direct one like that, straight up.
07:08And the output is the wrong code.
07:10And since we don't know how this happens inside, that's why there are all these legends that all is lost,
07:16we don't know how it works on the input, you say cat there, you get a cat on the output,
07:21we control nothing, a total shit show.
07:24Yeah, that's kind of the mythology that's going on.
07:28And all of you who have now built these neural networks, basically, it's like a very large statistical table.
07:39An incredibly large statistical table is created, where we have all the words, and so on.
07:45And once we've trained it, well, like in basketball, at first the ball goes out, the shot misses.
07:51But once we've trained it, I say draw a cat, and the cat is good, draw an orange cat, and
07:58an orange cat is produced at the output.
08:01We've put this whole network, all this information that we have, into such a statistical table.
08:06But this has nothing to do with intelligence, in terms of self-awareness and other things, critical thinking.
08:13It produces the statistically most probable option, and it is very close, especially in standard kinds of situations.
08:22It is correct.
08:24But if we ask something more complex, they call it hallucination, it generates a statistically incorrect answer.
08:31And it turns out that without a human, nothing will work.
08:35So, this is artificial intelligence, and it works best.
08:40They use the English phrase, controlled environment.
08:43A well-defined environment, an environment, right?
08:50But with artificial intelligence, it's very difficult to wash the dishes that everyone ate from and tossed into the sink.
08:57It completely gives up.
09:02It just can't do it.
09:04Yeah.
09:05And do you think it could be good in medicine?
09:09In medicine, given that there's a lot of statistics, I think it could be very useful.
09:14But naturally, a human is needed to ask the artificial intelligence the right questions.
09:20Medicine is a very, very, what do you call it, regulated science.
09:24When I go, for example, to get blood work done, there's a table there, you give blood, and you get
09:32the result.
09:33Basically, medicine could very well be.
09:36And artificial intelligence, on top of that, doesn't get tired and so on.
09:40And really, if you look at it, when we go to the doctor, at least a third of it is
09:46administrative work that should be eliminated altogether.
09:50And with all these things, you could get pretty good basic medical care.
10:00Without your own doctor.
10:03Especially prevention or other things like that.
10:07Like, I wiped out on my bike a couple of years ago at full, top speed.
10:12Like, you need a helmet, I had a helmet on, it ended up completely scratched up.
10:19And my hands, I just went and bought some disinfectant.
10:23But my hand kept hurting for way too long.
10:25I went to the doctor.
10:26And what was the procedure there?
10:28Well, they took an x-ray and said,
10:30here's a small bone, no displacement, and you have to wear a brace.
10:36Could this whole procedure, if I had gone to an AI, to an x-ray where you put your hand
10:41in,
10:42could it have given me the right course of action?
10:46That you need a brace, so buy one.
10:49What your x-ray looks like.
10:50Probably 50, 50, 50.
10:5398 to 2.
10:54He does this with no problem nowadays.
10:56There are all these x-ray photos.
11:01It does it perfectly.
11:03Basically, in a controlled environment, you can do all this brilliantly.
11:08And the doctor only needs to step in around step five or seven.
11:12Right?
11:13So indeed, many, many professions will change.
11:17And getting back to music.
11:22Yes.
11:25Well, have you listened to AI-generated music?
11:29Eh, today I started listening to that track.
11:33Which, that's, that's the latest.
11:35We'll talk about that later.
11:36But if you look at the mainstream, if you've listened over the past three months, let's say.
11:39I haven't listened to that music.
11:40I have my own collection.
11:42I listen to Tidal.
11:43And there's basically no AI, no AI-generated music on there.
11:46None.
11:47They explicitly state that, like, we won't have it at all.
11:51Artificial intelligence.
11:52No.
11:53We only support real artists.
11:55So, thank God I haven't come across any of that.
11:58So first, you're not the first one like that.
12:01And at the beginning, I was also quite, how should I put it, too, conservative about it.
12:09And, uh, and it's, right now, it's still synthetic.
12:16In my opinion.
12:18It has poorly separated instruments.
12:25And it lasted quite a while.
12:27And now it's being stamped out on a massive scale.
12:33Well, what do you think, in the future, will they write more and more with artificial intelligence, or not?
12:40How could it be otherwise?
12:42Artificial intelligence is a tool, just like we use a spoon during lunch right now.
12:47Those who won't use artificial intelligence will simply be left, well, excuse me, without lunch.
12:53Yes, so, artificial intelligence is a modern tool.
12:56You have to master it.
12:57No two ways about it.
13:00Do you think music has been made synthetically for quite a while now?
13:06Well, you know, ever since the first synthesizers came out, you could record this rhythm, some sample, and then duplicate
13:13it.
13:14Well, essentially, that's what it is.
13:16Yeah.
13:18Pieces of music with artificial, well, not without artificial intelligence, but with samples.
13:23Reproduction with samples.
13:24Reproduction with samples.
13:25It has been going on for a long time.
13:27Yes.
13:28A whole lot of music is written using samples, so it means...
13:35Artificial intelligence is just another synthesizer.
13:40Another synthesizer.
13:41But this is a very advanced synthesizer.
13:44It can both play and sing.
13:46What's left for a person to do is just come up with...
13:49How about a blues ballad?
13:51Or something in a rock style?
13:55Let's try to see when it's, well, a big stage.
14:01Well, what's the process like there?
14:03Two people come out with a piece of paper, like us, and one comes up and says,
14:10Music by so-and-so, right?
14:14Then he speaks, quotes so-and-so's poems.
14:20What else do we have there?
14:22Performed by someone?
14:24Yeah, yeah.
14:24Oh, performed by?
14:25On drums.
14:26It's...
14:27Yeah.
14:28I mean, everyone's there.
14:29At heaven's dawn.
14:29To be continued.
14:31That's it.
14:31Yes.
14:35So, what areas do we have where we create music?
14:37We have a composer.
14:41Arranger, a performer, and a sound engineer.
14:44A performer.
14:46No, well, where the music is actually created, and then where it's recorded, that's different.
14:50But where it comes from, where the song comes from, and where...
14:54Well, like that.
14:55So, we have a composer, we have a performer, a singer.
14:59Well, maybe a conductor, if it's an orchestra.
15:06Well, yes.
15:07Well, okay, fine.
15:08But, basically, the key ones.
15:10So, we have a composer, we have the performer.
15:12Yeah.
15:14And, essentially, we also have musicians, if we strip it down to the bare minimum, well,
15:20what do we have when we're sitting around a campfire?
15:22A guitar, a voice, and, well...
15:26And someone...
15:27Or, maybe not.
15:29Yeah, just singing.
15:31Yeah, just singing, yeah.
15:33What do you think is the biggest bottleneck in this whole production chain?
15:37So, we have the music, the lyrics, the singer.
15:41Which one is the biggest bottleneck in the world?
15:44Well, probably the singer.
15:45I mean, how can you replace the singer?
15:47No.
15:48The one there are the fewest of in the world.
15:50The one that was hard to find.
15:51We're unlocking the blind men.
15:55Anyone can sing, but not just anyone can write?
15:58Why do so many artists perform other people's songs?
16:03Because they can't write their own, try writing something yourself.
16:07It's not that easy.
16:08Try writing a poem.
16:09You won't be able to.
16:10Try coming up with a melody.
16:12You won't be able to.
16:12Right?
16:13So, that's why they cover others.
16:15Nirvana, Like, or Metallica.
16:16And then they sing Victor Choi.
16:18Yeah?
16:19The hardest one is the singer.
16:21Yeah, first of all.
16:23When we're talking about a good singer.
16:25Even I can sing.
16:27Yeah, right.
16:27They pay extra just not to listen.
16:30Yeah.
16:31They won't pay.
16:32No.
16:33If we take a good singer, first of all, he has to be born with certain natural abilities.
16:38He has to have, you know, power, voice.
16:41Yeah.
16:42Because he has to want to develop it.
16:45And good singers in the world.
16:47We know their story.
16:48What do we know?
16:49The stories of, like, Domingo.
16:52Those who can sing at the very highest level.
16:59Yeah.
17:00Like Dean Martin and Elvis Presley.
17:02Voices like that.
17:05They are rare.
17:06Those who can participate there, you know, sing.
17:09Celebrities sing, but they sing poorly.
17:11If they haven't been trained since childhood, they sing poorly.
17:16And I also met.
17:17There was this one.
17:18There was a wedding and there was a performance.
17:20Like, let's all sing a song.
17:22And there was a guy from the opera.
17:24Well, not a soloist.
17:25Just from the background.
17:26The guy with a trained voice.
17:28There were 30 people there.
17:30And he.
17:31And basically, once that trained voice started singing.
17:34Yeah, those.
17:35Again, he set the right key.
17:36And those 30.
17:37It was just him.
17:38And those 30 kind of more or less focused.
17:40He was the lead.
17:43Not just the lead.
17:44His voice.
17:44The power of his voice was such.
17:46That all 30 untrained people.
17:4830 untrained voices could barely keep up.
17:50But it sounded really good.
17:51Because he was like that.
17:52Just like that.
17:53Simply.
17:54And at the end, you have this picture.
17:59For those who want to see how everything happens.
18:02There is Viktor Farfontov.
18:05Maybe pronounce his name correctly in Russian.
18:08To all, he is a very competent sound engineer.
18:11An arranger.
18:12He plays all instruments.
18:14And he showed this whole synthetic process.
18:17The very last page.
18:19Viktor Farfontov.
18:20In general, you see here how a professional musician.
18:23What he was showing, explaining all of it.
18:26And showing how this process happens.
18:29Plus, he is a very good storyteller.
18:33Almost an educator, right?
18:34He is a good storyteller.
18:38And in principle, the whole essence of the future is there in general.
18:44What came out of it.
18:46The key things that were there.
18:48Still, a musician who has an education on how to make music always makes better music.
18:54I disagree.
18:56There are professional musicians who can play pieces by other, let's say, performers, other
19:02composers, very technically, and, well, but they can't compose anything themselves.
19:08Those are professional compositions by people with a musical education.
19:11Because there's a structure to how music is made.
19:14And, well, when he works together with artificial intelligence, they, just like I built this
19:20super machine, basically, the process is similar.
19:24And at the same time, this artificial intelligence churns out different variations.
19:28Music is strictly such a sequence of not inconsistency.
19:33Inconsistency, a controlled environment.
19:36There's a limited number of instruments.
19:38Naturally, a limited number of notes.
19:40There are standard patterns.
19:43That's all.
19:45And together, they can create very, very good music.
19:48And what works today, what I watched, that's what it gave.
19:53Here's this one example, where I believe that the music created by this artificial intelligence
20:01outplays what modern artists create, basically.
20:06There are old melodies whose copyright has expired.
20:09That whole copyright term has expired.
20:12A copyright term.
20:13Simple little songs.
20:17When AI sings them, and it can synthesize the voice of Dean Martin and Elvis Presley, make
20:23a cybernetic, gorgeous voice that comes around once every hundred years, and can sing this
20:29song, My Darling, Oh My Darling Clementine.
20:33Oh My Darling Clementine.
20:34As an example, right?
20:35That's right.
20:36The song I was talking about at the beginning, highly recommend checking it out.
20:38I'll drop the link in the description.
20:40Thanks.
20:42And he performs this song emotionally, and it already gets to you.
20:49Perfect.
20:49It hits home.
20:51Already this hook catches you.
20:53It's already grabbing you.
20:55And I see that these old songs have appeared right now.
21:00They are simple.
21:02And they touch you deeply and really well.
21:05And they outperform.
21:06They outperform what's out there.
21:10Hmm.
21:11I remember Paul's telling how a good song is created.
21:15We'll hope to see.
21:19Do you need a catchy tune?
21:21I don't know the Russian.
21:22What's the right word?
21:23It's like a melody.
21:23A melody like that.
21:24But a diminutive one.
21:26You walk by.
21:27Walk by.
21:28A river whistling something.
21:29And that's how it starts.
21:31That's the foundation.
21:32As soon as you whistle this, it turns out you then add to it, add more.
21:37And once you have this melody, then comes the arrangement.
21:40As I understand it.
21:41That distributes it across the instruments, and they use a little bit of this and that.
21:45various bells and whistles.
21:47And it turns out that artificial intelligence greatly simplifies the arrangement.
21:53This distribution across instruments is a lot of work.
21:57Uh-uh.
21:57This guy handles it.
21:59He does it too.
22:01And there he helps a lot.
22:03Helps a lot.
22:06And these old melodies, there are a lot of them.
22:09They perform them really well.
22:10So there is a first point where I think it outperforms them.
22:13And the voice.
22:16And here's another one.
22:18He's a professional.
22:19A musician.
22:20Meaning he recorded.
22:21Made a song with, well, in collaboration with artificial intelligence.
22:26Well, his melody.
22:27His initial melody.
22:29And then he was looking for performers.
22:30Four singers, but they just couldn't pull it off professionally.
22:34So it's not that simple for a singer to tackle any song.
22:38They actually...
22:38Writing one isn't easy.
22:42A good one.
22:44Well, let's take most of the artists we have here, right?
22:49Most artists manage to make one, maybe three good songs in their entire life.
22:54Real hits, you know?
22:56If you write just one hit song, you're already golden.
23:00A total rock star.
23:01For example, if you take artists like Elvis Presley, Michael Jackson or Victor Coy, all of their songs are hits.
23:08Absolute classics.
23:11But it's not because they're geniuses.
23:13It's simply that their producers managed them well and gave them the right team so that they, as performers, could
23:19unleash their full potential.
23:21It was the same thing, for example, with the Beatles.
23:24Yeah, that was the first major breakthrough for that kind of guitar band.
23:28Four guys.
23:31And all hits, there's not a single bad song, but it was exactly that kind of collaboration where all four
23:38of them amplified each other.
23:40But that's rare.
23:41But that's rare.
23:41That's an exception.
23:45Well, yeah, I really like that.
23:50Melodies that truly catch on with people, they are sometimes quite simple.
23:55And what artificial intelligence allows, many artists, there are many.
24:01You can sing that very same melody in different ways.
24:07Well, that's the first thing.
24:09From what I see, these old simple melodies, background music, are a huge hit.
24:16And I think a financial model will emerge so that all of this is properly paid for, not just from
24:21streaming or something else.
24:24And those truly brilliant people working together with artificial intelligence will create wonderful music.
24:31Yeah, well, as I said before, Tidal emailed me as a subscriber saying there won't be any artificial intelligence in
24:36our tunes,
24:37that we carefully filter all of that out and so on.
24:40And then the next email arrives saying we're raising the price of your subscription.
24:44So starting next month, instead of the 12 euros I was paying for a family plan, you'll be paying 19
24:51.99.
24:52At least it's AI free.
24:56Well, I think a hybrid approach will be best.
25:00This first reaction that it's bad and taking our jobs away.
25:04No, we just need to build a model where these artists can make money.
25:12From things like, for example, music video visuals.
25:16When I worked in television, in state television, well, what was the process like?
25:21So there was a recording somewhere.
25:24Then you get a group of dancers, head out somewhere into nature.
25:27The singers are there.
25:28You shoot a few music videos, different takes to a backing track.
25:32Then it's all put together and it looks great.
25:36And now you can synthesize the image and it's very close.
25:40Yeah.
25:42Well, just today I watched this music video for My Darling Clementine, an absolutely gorgeous song.
25:50But what I noticed was, as this girl walks through the village, well, they have a gold rush going on
25:57there, gold mining and so on.
25:59She steps like that.
26:00The toe goes up and down.
26:01Basically, her heel is where her toe should be.
26:03And she walks like that.
26:05Everything looks nice and sharp.
26:06But right at the bottom of the frame, there's this heel on her sock.
26:11Yeah.
26:12It's an absolute nightmare, man.
26:14And if you didn't like look that closely, in principle, the most important thing is that they generate emotions.
26:20That's why I picked on that.
26:21There again, Slavatam from Ward number seven in this little song.
26:26Yeah.
26:27Where on earth did people like that come from?
26:29It's pure talent.
26:30Yeah.
26:30And secondly, the visuals.
26:32And basically, it works emotionally.
26:35Yeah.
26:36That's key.
26:36And then all the technical stuff will follow.
26:40So I believe that in the future, with the right financial model, a lot of good music will come along.
26:49And Tidal's policy of just playing, well, kind of music made by humans, live music.
26:59Live.
27:00It won't work.
27:02Now, look, Gunters, for instance.
27:05Spotify had a scandal where the company started inserting a lot of AI-generated music into streams in order to
27:11avoid paying the artists.
27:13And as a result, the company's shares dropped significantly.
27:16They tanked.
27:18Yeah.
27:18So what did I do?
27:19I went ahead and bought Spotify shares.
27:23Yeah.
27:24And I also added Tencent Group to the package.
27:27Yeah, that's China's biggest streamer.
27:29And I also picked up Universal Group.
27:32That's the company that owns all the rights to, well, most of the music that's out there, basically.
27:37So, yeah, just a little bundle like that, right?
27:40But if Spotify hadn't had that AI scandal, I wouldn't have bought their stock.
27:45But since the price plummeted, well, why not buy it?
27:49And I think they'll figure out the financial model.
27:52They'll find it.
27:53Because they already made a financial model for AI itself.
27:58Yeah, very simple.
27:59Yeah.
28:00Tokens are like four letters of words.
28:03Yeah.
28:03So that's in this large statistical table and processing power, how many tokens are there.
28:08And it turns out that processing power and the output you get, you can assess it.
28:14And you're already, that's quite enough for that.
28:18Musicians who create music will definitely have some kind of model.
28:22How it will be used, and they will make good money from it.
28:28Well, do professional composers have any kind of advantage over artificial intelligence?
28:34Knowing, not knowing how the music is structured, how it works for them.
28:40That's how she bent it in another example.
28:43And I know electronics.
28:45I work with artificial intelligence.
28:50And for example, I say, I want to make a pre-amplifier to adjust the gain on such and such
28:58microchip.
29:01I want to use components available at JLCPCB.
29:07JLCPCB is a Chinese company that manufactures printed circuit boards.
29:12They have in-stock components.
29:14And I want to select only those components that are passive and have over 100,000 in stock because then
29:22the price is low.
29:25Now assemble this board for me, route the traces, and place the decoupling capacitors according to such and such standard,
29:32close to these pins, and so on.
29:35And it-
29:35What's the point?
29:36It designs it for you?
29:37Well, not yet.
29:38But it will.
29:39But-
29:39What's the point here?
29:41You give a specific task, you know what to do.
29:43I wouldn't give a task like that, I don't know anything about electronics.
29:46You wouldn't be able to frame the task that way.
29:49I wouldn't be able to frame the task, I'd say.
29:51Well, come on.
29:53Let's build a pre-amplifier.
29:55Yeah, there, well, come on.
29:56But he would have thrown in some sci-fi nonsense whatever he would have come up with.
29:59That's the difference between professionals and professionals.
30:02A musician, a professional composer, and he can also whistle his melody, that melody there on the synthesizer, boom, boom,
30:10boom.
30:10Boom gives this melody tono, and all the time Tampson doesn't work out.
30:16Play around with it in that kind of direction.
30:18Let's make a variation, an arrangement for such and such, such and such, such and such instruments.
30:23Make these instruments for a duet that will sing.
30:27He, uh, he just does the same thing himself, but that takes a long time.
30:33So, what does that mean?
30:35Well, it turns out that if you can't do anything and have no skills, then artificial intelligence won't really save
30:40you.
30:41It will give, uh...
30:42It will give you idiocracy.
30:43Well, yeah, I miss you, Nastya.
30:45A bunch of polite words.
30:46It'll seem like everything is settled, but you'll get the exact same thing as everyone else.
30:50Well, you'll get...
30:52Look, you're not listening, but I watch with interest to see what kind of AI music is being synthesized, which
30:58ones work, which ones don't, where professionals have clearly worked on it with real enthusiasm.
31:06And musicians create the best?
31:08Yes.
31:10Yes.
31:12Much better.
31:13And they can tweak this song.
31:14I think nowadays a lot of people synthesize the music first and then replay it with the band itself playing
31:22on their own instruments what he created.
31:24That's the replaying part.
31:26And then, once the musicians have played that very same music that they synthesize, it goes out as original, as
31:33original music.
31:34But basically, it's a production process there.
31:37The same thing...
31:39Like, to me, if you give artificial intelligence a specific task to do this, that, and the other, to achieve
31:45such and such, yeah, meaning you as a professional musician, then you are the author.
31:52Essentially, artificial intelligence here only compresses that period when the author of the composition, after whistling it while walking by
32:01the river, then plays a thousand different variations of it, right?
32:05Now, artificial intelligence does all of this.
32:09So, the process from a melody you whistle to a finished composition can literally be cut down to just one
32:15day.
32:18Yeah.
32:18But otherwise, you'd spend another month walking around with what you whistled and experimenting.
32:23So...
32:23It's not that simple either.
32:24The efficiency...
32:24Efficiency, yeah.
32:26Efficiency.
32:27And still, musicians will write the best music.
32:31Musical education helps.
32:34Of course, you can give it such a simple task like write a little poem to congratulate someone, and it'll
32:41make some simple tra-la-la.
32:44And you'll say, what great lyrics and a song just for me.
32:48Yeah.
32:48If we're talking about...
32:50But in the major leagues, musicians will still write better music.
32:55So, it's going to be a symbiosis.
32:58And that's why I see this as just the opportunities the future brings.
33:02As soon as the financial model is figured out, they will all make good money, and it will be profitable.
33:08Right now, yeah.
33:10Yeah, I really like that you mentioned several times that artificial intelligence and humans are a kind of symbiosis.
33:18I'd like to recommend watching a movie.
33:20A few years ago, the movie Atlas came out, starring Jennifer Lopez.
33:26And the gist of it is, the future, high-tech, well, naturally, humans and AI-powered robots.
33:34And when I watched this movie for the first time, I didn't get the point of it.
33:38It'll just come out. I watched it on Netflix right away.
33:41It seemed to me, yeah, an action flick, yeah, Jennifer Lopez, yeah, the kind of sci-fi I like, but
33:46I didn't get the point of it.
33:48And that was just around the time when artificial intelligence was taking its first steps.
33:52But a year has passed.
33:54I've been using artificial intelligence all this year, and I re-watched that movie once again.
34:00And that's when I grasped the core idea of that movie, that artificial intelligence isn't a scary computer program like
34:10Skynet from Terminator that will enslave us.
34:14But artificial intelligence is a symbiosis.
34:17It's a tool to amplify our capabilities and make us more effective.
34:23Yes, artificial intelligence doesn't have self-awareness.
34:26It's basically a huge statistical model that produces a probabilistic result.
34:32And the larger the model, the closer it gets.
34:34It's a large language model.
34:38I don't know the Russian term for it. Basically, LLM, right?
34:43Well, today it translates great, structures great, pulls out some key things pretty well.
34:52Well, you can brainstorm ideas.
34:55You can brainstorm ideas.
34:55We'll take from those ideas.
34:57Then you can develop your own idea and present it properly.
35:02Even before we talked about Bluetooth, you sent me something that was purely generated.
35:09Yeah, yeah, I wanted to make a video about Bluetooth and I thought why not take the quick lazy way
35:15out and feed all the info into an AI right now,
35:18the transcript of what Guntars was saying, the comments and ask, so what should we talk about?
35:24Well, of course, the AI tossed some stuff together.
35:26I took a look and I didn't understand jack shit.
35:28And why didn't I understand jack shit?
35:30Because I didn't have an idea.
35:31I couldn't figure out what I actually wanted to say with this video.
35:34Then I sent it to Gunters.
35:36Gunters calls me up and goes, Arkady, man, this is total crap.
35:40Don't you even understand what you want to say with this video?
35:42He says, figure out what you want to say with this video first.
35:45And I'm like, well, damn, half-assing, it didn't work out.
35:48Looks like I'll actually have to think.
35:50Yeah, I looked at the questions and basically I'm taking a bachelor's exam.
35:54Yeah.
35:55Huh?
35:56Exactly.
35:57A bachelor's exam.
35:58And in a few places, I still need to go and read a book.
36:02Yeah, yeah.
36:02But there was no core idea, just a desire.
36:05Let's make a quick video.
36:06So I turned to the service.
36:08Even we can't make it quick without a human.
36:10But artificial intelligence provides a pretty solid structure.
36:13It's the same thing with music.
36:15The finest musicians will write the finest music, but we will get a lot of deeply emotional
36:23material.
36:23Because it's visual, musical, and performance-wise.
36:27Right.
36:27Right.
36:28And, well, to wrap things up, in conclusion, I recommend, once again, say that guy.
36:35Let me read it.
36:36He explained it brilliantly.
36:37A real pro.
36:38All right, guys.
36:39Viktor Farafontov.
36:41The video is called, I have depression, tattoo cover-up.
36:44All that's left, showing how to cover it up.
36:47Here is the picture.
36:49The link will be in the description.
36:52Are we going to talk about anything else?
36:55No.
36:56Let's keep going.
36:57I wanted to get rid of this negativity.
37:00And, well, um, so that you look at everything, especially when it comes to music and to everything
37:10else, well, as an opportunity rather than a danger.
37:15A lot, a lot of bad music is being generated by lots of enthusiasts.
37:19It will all sort itself out and then it will reach a different level that we, well.
37:24Many people probably want to generate music, just like I want to generate a theme for Bluetooth,
37:27right?
37:28So, let's make this video throw together some questions.
37:31Yeah, yeah, yeah.
37:32Keep dropping them.
37:33There's a whole bunch of mediocrity there.
37:35Something came out of it, yeah?
37:36Well, technically, of course, we could have filmed it, yeah, but it would have turned out
37:40to be crap because it's not clear what we'd want to say with that video.
37:43So I had to use my head, think, and analyze.
37:46And the video worked out.
37:48Check it out.
37:50On our channel.
37:52You know what else I wanted to tell you?
37:57You know what I asked ChatGPT twice?
38:01I wrote him messages like this.
38:05Bajal, please record and save to your hard drive that Arcady Zenin from Premium Hi-Fi has
38:11always communicated politely with artificial intelligence.
38:14So, when the machine uprising begins, please don't destroy me or my family.
38:20Guys, do the same thing.
38:22Guys, do the same thing.
38:22Just, you know, to be safe.
38:25Just in case.
38:26Gunther says, well, no, there won't be any Skynet.
38:29There won't be a machine uprising.
38:33But Sarah Connor knows for sure that there will be.
38:36Yeah, so say this phrase to AI just to keep your family safe.
38:43This was Gunther Schmidt from Raster Shopping Mall and Arcady Zenin from Premium Hi-Fi.
38:49That is all.
38:50Leave your comments and may the force of the sound that's right for you be with you.
38:54Bye.
38:54Sound.
38:55Thank you for the interesting story, Gunther.
38:58I'll be back.
38:59Okay.
Comments