|
|
Post by drbill on Apr 17, 2024 9:51:17 GMT -6
The game is close to being over for those who write for : - commercials - TV - B & C level films. A level blockbuster films will still hire a real composer. - Youtube videos - reality TV - anything that does not require a "artist" personality to drive it. (And even there, AI is accelerating) In the last 6 months since I really thought about this, the strides that have been made are astounding. There are no brakes, and there is no political will to stop even if there were brakes. If you're OK with your musical experience being the total sum of playing your guitar in a coffee shop for $50 a night, then all is good - those opportunities will continue, although you'll have much more competition from the other musicians whose careers have been annihilated. If you seek a studio career as an engineer, or studio musician, or arranger, or composer, or producer / etc. there will be little work for any but the AAA level guys/gals with long standing career experience and accolades. This will destroy the middle class musicians. Some are saying make as much money and invest as much as you can right now, because by 2030 the work force will look nothing like we have now, and even jobs may be a thing of the past. Just "believing" that this will go like past trends is naive. Buckle up!! Of course it's not a trend and it won't go away, but neither will the public's need for 'real' entertainment. That means live performances by living breathing musicians that they can relate to on a personal level. There's no question, technology has decimated millions of jobs over the last 200 years. I for one am grateful that I don't have to ride a horse to get to work to pick cotton in field for 12 hours a day. When tech does something just as well and more efficiently than humans, I say bring it. Yeah, there's pain with progress, but nothing can stifle the human spirit. People won't stop playing instruments or writing music just because computers can do it better or more efficiently. People will create because they just have to. It's in our DNA, and technology will be there for us to make it better. As for an example with AI, already a decent Chess (or Go) program can beat the shit out of the best players in the world playing their A+ game 100 times out of 100, yet the (human) chess world is enjoying a renaissance of popularity. Sure, there are tournaments where computers play computers, but guess what? Almost no one cares. They prefer to watch smelly meatbags playing the game. Me? I always wear a seat belt. My personal perspective tends to focus on making a living in music - as that's what I've done my entire adult life. Of course you are right - music made by humans will continue. Perhaps we'll even head back towards the early 20th century scenarios of "sitting around the piano singing songs in the living room", but I kind of doubt that with virtual reality beating at the door. However, the opportunities to do music on a significant professional level, with the time and effort and growth that full time efforts directed towards the love of your life - music - will wane. It's sad, and what's sadder is that leaders COULD control this, but greed and power seem more important at the time. I feel exceedingly lucky that I grew up when I did and had a career at the time period I did. Blessed and grateful. |
|
|
|
Post by ninworks on Apr 17, 2024 11:05:02 GMT -6
They’re working on an AI that will listen to music for you. It will let you know if you liked it or not and whether you want it to listen to more music in a similar genre or try new music. It’s going to free up a lot of your time for other hobbies and interests. Like watching cat videos on your phone. |
|
|
|
Post by thehightenor on Apr 17, 2024 14:10:27 GMT -6
They’re working on an AI that will listen to music for you. It will let you know if you liked it or not and whether you want it to listen to more music in a similar genre or try new music. It’s going to free up a lot of your time for other hobbies and interests. Like watching cat videos on your phone. If that’s your hobby, sure who am I to judge  |
|
|
|
Post by ironinthepath on Apr 17, 2024 14:40:35 GMT -6
This had me laughing, but was it AI somehow or just good makeup artist? (Beavis/Butthead SNL)
|
|
|
|
Post by chessparov on Apr 17, 2024 14:48:54 GMT -6
This A.I. algorithm Guy walks into a bar...
"I'll have what everyone else is having".
|
|
|
|
Post by lowlou on Apr 17, 2024 17:02:53 GMT -6
Good one Chessparov !
----
AI Depeche Mode
|
|
|
|
Post by drbill on Apr 17, 2024 18:10:34 GMT -6
Yeah, no worries guys. ^^^^. The above could never make it in the real world. <<rolleyes>>
If y'all get tired of your fav songs by your fav artist, all you have to do is have AI generate more new hits for you.
|
|
|
|
Post by chessparov on Apr 17, 2024 18:58:32 GMT -6
Maybe some Milli Vanilli would cause it to implode?
Unless it's thoroughly modern Milli.
Chris
|
|
|
|
Post by M57 on Apr 18, 2024 7:49:34 GMT -6
This all begs the question. Is AI becoming sentient?
..or let's turn it around. Are we just evolving biological computers?
|
|
|
|
Post by ironinthepath on Apr 18, 2024 8:34:20 GMT -6
Good one Chessparov ! ---- AI Depeche Mode Dang, this has the same lo-fi MP3 sound thing going on, but it's scary impressive otherwise. |
|
|
|
Post by gravesnumber9 on Apr 18, 2024 11:48:25 GMT -6
Good one Chessparov ! ---- AI Depeche Mode Dang, this has the same lo-fi MP3 sound thing going on, but it's scary impressive otherwise. My take is that, at least for now, this works a lot better with music that is synthesizer and beat based to begin with. I don't mean this in a good or bad way, but in my opinion that stuff is more mathematical to begin with for lack of a better term. Even when it's made by humans, it doesn't sound very organic. So yeah, music that sounds like it's made by robots can also actually be made by robots. Is that a hot take? This doesn't mean that it couldn't get there, it just means that I've yet to hear anything that sounds at all creative to me. And I mean "at all" creative. And from a lot of the AI people I talk to in my day job, the general consensus is that we are very, very far from genuine creative output in AI and some even think we will never get there. I'm not talking about head in the sand folks, I'm talking about guys who are developing neural networks and stuff like that some of whom think that AI will literally never be able to generate genuinely new ideas or ever make creative decisions whether in music, art, or medicine, or anything. The basic problem seems to be that we don't have the slightest notion of how the human brain makes up ideas but we do know that we do it easily. The challenge is that because we "make stuff up" basically from infancy, it's nearly impossible to study how humans do this since it appears we don't learn to do it, we just do. |
|
|
|
Post by M57 on Apr 18, 2024 12:31:52 GMT -6
This doesn't mean that it couldn't get there, it just means that I've yet to hear anything that sounds at all creative to me. And I mean "at all" creative. And from a lot of the AI people I talk to in my day job, the general consensus is that we are very, very far from genuine creative output in AI and some even think we will never get there. Unfortunately the vast majority of human output is also not creative "at all." Think of how many musicians out there just copy what they've already heard ..or who for the most part can only play printed sheet music, or need tabs, etc. There's a lot of "creators" out there who slam together a handful of loops and call it a day. AI may not be truly creative (ever) but it can already put a fork in a lot of what passes for art created by humans. |
|
|
|
Post by ironinthepath on Apr 18, 2024 12:36:37 GMT -6
Well, the first Udio example in this post, to me, is a bit of a counter-example - it's pretty good for being 100% "fake", if it were in the background in a cheesy made-for-TV movie it would probably go unnoticed. I know that AI algorithms are essentially just doing what they are trained to do, it's not like it can "feel" emotion and put into a song, it just mimics it.
The infrastructure and everything necessary to get Udio going is probably quite expensive, but the additive cost of "creating" one new song with this tool is probably boat loads cheaper than paying icky humans to do the same thing - they get tired, need to eat, need bathroom breaks, need an acoustically treated fancy room with fancy mics and expensive instruments. Obviously, I'm on "team human" but I'm not ready to discount the strides made in AI - totally legit. Has me thinking even (non-audio) engineering jobs will be greatly impacted soon enough (the people designing the RF chips for your cell phones, etc.).
In any case, I do hope your friends are right!
|
|
|
|
Post by gravesnumber9 on Apr 18, 2024 14:03:47 GMT -6
Well, the first Udio example in this post, to me, is a bit of a counter-example - it's pretty good for being 100% "fake", if it were in the background in a cheesy made-for-TV movie it would probably go unnoticed. I know that AI algorithms are essentially just doing what they are trained to do, it's not like it can "feel" emotion and put into a song, it just mimics it. The infrastructure and everything necessary to get Udio going is probably quite expensive, but the additive cost of "creating" one new song with this tool is probably boat loads cheaper than paying icky humans to do the same thing - they get tired, need to eat, need bathroom breaks, need an acoustically treated fancy room with fancy mics and expensive instruments. Obviously, I'm on "team human" but I'm not ready to discount the strides made in AI - totally legit. Has me thinking even (non-audio) engineering jobs will be greatly impacted soon enough (the people designing the RF chips for your cell phones, etc.). In any case, I do hope your friends are right! Actually I think non-audio engineering jobs are going to be really, really hurt buy AI. Which is ironic since those are the guys that are inventing this stuff. AI can already do complex code more accurately than people although it still needs humans to direct it and cross check it and all that. I agree, where Udio type thing is sure to crush is in sync placements. That actually really hurts me and artists like me in some ways since sync is one of most reliable ways that I fund my own songwriting. But let's be real, in most of my syncs you can barely even tell it's my song. AI would be no worse at that level of limited audibility. Here's a really great example of what people do instantly (try it with a three year old) but that AI struggles with. Take four unrelated words and make a coherent story out of those words. Not just a story that uses all the words but a story that naturally weaves them together. Here is Chat GPT trying to do that (I literally just did this). chat.openai.com/share/ebf3af11-7a47-4d72-aace-44afc8b6a0af |
|
|
|
Post by gravesnumber9 on Apr 18, 2024 14:07:02 GMT -6
Even giving it a second chance, the AI couldn't make the really fanciful connection that the elephant could be the center of the story and could be picking strawberries near the volcano when it found a secret door that could only be unlocked by typing a code into a hieroglyphic keyboard.
Or anything like that.
What the AI does is make discreet images and try to tie them together. What humans do is create a new idea. Again, even children do this. There is zero literature for AI to draw from that has strawberry hunting elephants finding trap doors in volcanoes. But I came up with that in 5 seconds of thought.
Interestingly, this is how the current cutting edge white hat hackers defeat password decoding algorithms. What I'm talking about here is battle tested, not just theory.
|
|
|
|
Post by gravesnumber9 on Apr 18, 2024 14:12:03 GMT -6
So yes AI can say bizarre things. And weird things. But it doesn't do creative things. And despite massive effort to figure it out, many are kind of giving up on it. Self driving cars are a great example. It looked like we were getting there rapidly. We were all saying "at this trajectory, we'll have them in a year or two." But it turns out that last 10% (sound familiar??) is the hardest to get. That's the part humans do naturally. And now you have leaders in the industry saying we are actually decades (DECADES!!) away from true autonomous vehicles and it may never happen at all. www.businessinsider.com/fully-self-driving-car-tech-needs-decades-to-develop-2023-10#:~:text=Fully%20self%2Ddriving%20passenger%20cars,of%20an%20autonomous%20vehicle%20company. For the same reason. We can't teach the cars to make creative decisions, we can only teach them to respond to stimuli in the ways we train them to. I see the same outcome for music/writing/etc. The tech will get far enough for light commercial purposes (jingles, interstitials) but may never get to the point of replacing music as a form of entertainment. |
|
|
|
Post by gravesnumber9 on Apr 18, 2024 14:14:54 GMT -6
Last point and then end of soapbox.
Counter-intuitively I think AI is going to increase the value of those of us who work with humans and decrease the value of those who don't. That's good news for all the producers out there. Maybe not such good news for the beat programmers. Great news for salespeople, probably bad news for demand gen marketers. Great news for interviewers, bad news for newscasters.
Etc etc
|
|
|
|
Post by lowlou on Apr 18, 2024 14:45:14 GMT -6
Except baby AI has figured human voice quite well already. So I don't know about this recorded vs programmed music distingo. From what I hear, it's good at doing both already. Moogish square wave bass and 808 sinusoidal trails are the easiest sounds to replicate but I fear it'll reproduce everything with ease & in 5 seconds.
|
|
|
|
Post by gravesnumber9 on Apr 18, 2024 15:34:43 GMT -6
Except baby AI has figured human voice quite well already. So I don't know about this recorded vs programmed music distingo. From what I hear, it's good at doing both already. Moogish square wave bass and 808 sinusoidal trails are the easiest sounds to replicate but I fear it'll reproduce everything with ease & in 5 seconds. The voice replication is really impressive. It's already good at replicating idiosyncrasies of existing voices. |
|
|
|
Post by rowmat on Apr 18, 2024 16:46:52 GMT -6
I’ve noticed a ton of YouTube videos are now being created using AI.
For instance ‘COPS’ style law enforcement videos using AI narration and subtitles that are often incorrect.
Also many videos showcasing famous photographers using the same type of slideshows and music.
Often some of the photos are duplicated especially if they have been cropped differently which indicates the AI was fooled and didn’t identify the duplicate.
It has reached the point where AI can now create a YouTube channel from scratch based upon a subject and then using bots harvest existing content from the internet and re-edit and assemble it including a script, narration, music and automatically create and publish new videos at set intervals.
|
|
|
|
Post by Shadowk on Apr 18, 2024 18:27:45 GMT -6
I’ve noticed a ton of YouTube videos are now being created using AI. For instance ‘COPS’ style law enforcement videos using AI narration and subtitles that are often incorrect. Also many videos showcasing famous photographers using the same type of slideshows and music. Often some of the photos are duplicated especially if they have been cropped differently which indicates the AI was fooled and didn’t identify the duplicate. It has reached the point where AI can now create a YouTube channel from scratch based upon a subject and then using bots harvest existing content from the internet and re-edit and assemble it including a script, narration, music and automatically create and publish new videos at set intervals. I hate to say it but I work in tech and what's about today in the grand AI schema is nothing more than a gimmick compared to some of the big budget stuff held behind curtains. Practically no industry is safe, although it really does make me wonder when we get to the "big bang" what then? Economies fall, we don't have jobs and even the creative side is covered. So what exactly do we do with ourselves? I'm not afraid of not working that's for sure but I gotta fill that time. |
|
|
|
Post by rowmat on Apr 18, 2024 19:08:54 GMT -6
I’ve noticed a ton of YouTube videos are now being created using AI. For instance ‘COPS’ style law enforcement videos using AI narration and subtitles that are often incorrect. Also many videos showcasing famous photographers using the same type of slideshows and music. Often some of the photos are duplicated especially if they have been cropped differently which indicates the AI was fooled and didn’t identify the duplicate. It has reached the point where AI can now create a YouTube channel from scratch based upon a subject and then using bots harvest existing content from the internet and re-edit and assemble it including a script, narration, music and automatically create and publish new videos at set intervals. I hate to say it but I work in tech and what's about today in the grand AI schema is nothing more than a gimmick compared to some of the big budget stuff held behind curtains. Practically no industry is safe, although it really does make me wonder when we get to the "big bang" what then? Economies fall, we don't have jobs and even the creative side is covered. So what exactly do we do with ourselves? I'm not afraid of not working that's for sure but I gotta fill that time. Hence the murmurs of a UBI (Universal Basic Income) a minimum barely survivable carrot-on-a-string wage for all those made redundant. To risk going down the rabbit hole event further are the talking points emanating from some governments (Canada) of legalising euthanasia for those with mental illness. With the direction we are heading in increasing depression seems to a given for many. Depression being a form of mental illness could lumber those along with the elderly and disabled into the category of ‘useless eaters’. Better crank up the Soylent Green plant! 😵💫 |
|
|
|
Post by rowmat on Apr 18, 2024 19:22:43 GMT -6
Arms and legs make for a much more compelling live show typically. Sure do. And they are working on that. . Just wait till they execute the hot babe model - then the final frontier (did you notice my nice alliteration? AI is KILLER at artistic alliterations) is lips moving. <<thumbsup>> Lest anyone doubt or have their head in the sand, music 20 years from now will bear virtually no resemblance to music 2024, and certainly not to music 1979. At the rate of progression, it might take only 10 years.... Glad I'm in the phasing out stage of my career as opposed to the ramping up stage. My sincerest best wishes for those of you just starting out.... To put things in some kind of perspective these publicly available Boston Dynamics videos of various robots performing stunts are what they allow in the public domain. It would be foolish to think that these are the state of the art. The really scary shit that could be used in high tech military applications likely remains hidden within highly classified R&D projects probably deep underground. |
|
|
|
Post by christopher on Apr 19, 2024 0:17:37 GMT -6
Something to consider.. if any of us make a spoof of an artist and say it was AI, then in a few years we might have millions of views. We could make hundreds of these. Thats potential big money. Its easy for us to use pitch shifting to sound like these AI vids. I’ve thought about posting some AI vids shared with me who suppposedly trained with an artist’s voice and made a new AI song.., but then I think.. it could be some dude pretending, so much autotune artifacts
|
|
|
|
Post by wiz on Apr 19, 2024 1:18:48 GMT -6
Good one Chessparov ! ---- AI Depeche Mode Dang, this has the same lo-fi MP3 sound thing going on, but it's scary impressive otherwise. That's the first AI generated thing I have ever heard (that I am aware of). ....... sheeeeitttt... that's really impressive...and scary...know we know how drummers felt when the drum machines came for them....8) cheers Wiz |
|
