Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

DeepMind releases Lyria 2 music generation model

deepmind.google

Post reply on HN
Print view View original

371–380 of 444 posts

Re: DeepMind releases Lyria 2 music generation model

#371

Post by BobaFloutist »

tasuki wrote: post #274
nsteel wrote: post #177

Earlier quoted context omitted.

The scary thing is while a construction labourer knows how to properly lay bricks, an AI's output is reliably unreliable. Does the foreman have to go round checking every little thing on the site? That would be a waste of time. Using these AI tools for anything important is too high risk.

There exist many problems for which it's easy to verify the inputs-outputs, but much harder to write the functions to convert the inputs into the outputs. Just write an executable spec and have the AI generate the code that fulfills it. Where is the risk?

I guess the risk would be that the cost of compute would explode out of control long before you got a viable solution.

To give a bit of an absurd example, I imagine there's a reason it's not currently worth it to generate the code with a random character generator, even though hypothetically that would get you there eventually. If we consider AI a much, MUCH better version of a random character generator (let's say it's a million times faster. No, let's say that it'll get you the solution in quadratic time instead of factorial time), that doesn't necessarily mean it's actually worth it now.

Re: DeepMind releases Lyria 2 music generation model

#372

Post by giraffe_lady »

haswell wrote: post #329
giraffe_lady wrote: post #320

Earlier quoted context omitted.

I was basically unemployable due to health problems for several years before covid made work from home normal. It's not really theoretical for me either. It was, all things considered, great. I have never been more involved in the communities and connections that I find valuable and fulfilling. I learned several complex skills that continue to benefit me and the people around me, I taught and mentored young people so…

> I don't believe either of our individual experiences are really a good predictor of universal human experience in this area. Do you? No, and I said as much in my comment. My point was not that my experience is universal, but that I have direct experience with the failure mode of such an arrangement. Am I 50% of the population? 75? 5? I don’t know. But as I went though it, I met more people who’d gone through someth…

> No, and I said as much in my comment.

You had that disclaimer but most of the rest of the comment was about your prediction that most people would be affected in the same way. Some friendly feedback for your future writing on the subject, I guess.

Re: DeepMind releases Lyria 2 music generation model

#373

Post by iamleppert »

The problem with all these AI companies and products is that they all over-value the content itself. The content is not the thing that is valuable, its the person who creates it and the community and culture that forms around it. These are things that require people, a lot of time, luck, and can never be automated. You see this all the time with people who spend huge amounts of money and effort making things that may even be great and high quality but never gain any real value. Companies do it all the time too. It's very hard to predict what will become valuable, and what will not. Just because AI will let you create more of something, or do it faster or cheaper -- it doesn't matter as long as you are missing some external factor that makes something high quality popular or valuable. And yes, there is always an external factor at work, a lot of times many factors. This is further evidenced by the fact that quality itself of something valuable sometimes doesn't need to be there (although it is more likely to be there, than not).

Just by magically dropping the content price or effort to $0, doesn't matter because the content itself has no or little to no intrinsic value. There isn't suddenly going to be an expanded market of people who will listen to your AI generated music or buy your AI-enabled product.

If tomorrow, I could make a random kid on the street sing just as well as Taylor Swift, or even if an exact perfect copy of her emerged somewhere, it doesn't mean she has any relationship to the value of Taylor Swift.

Re: DeepMind releases Lyria 2 music generation model

#374

Post by ndriscoll »

fredoliveira wrote: post #214
dist-epoch wrote: post #88

Earlier quoted context omitted.

I will also be honest, if you expect to live off your art, you are doing it wrong.

Absolutely ridiculous to assume that only some careers allow the makers to live off them.

That's the point: for almost everyone it's not a career. It's a hobby. Like some people have a career researching physics because they're extremely good at it and society has decided it makes sense to have a few. Then there's people like me who learn what they can of it in their free time, but I do something else as a career because realistically very few people have need of someone who's familiar with the Dirac equation or whatever. Among the general population I'm probably in the 99th percentile of math/physics knowledge/ability, but I don't do that for work because we don't need 1% of the population working on such things. And that's for a skill that causes most people to get anxiety; the demand mismatch is probably even greater for things that average people actually enjoy.

Re: DeepMind releases Lyria 2 music generation model

#375

Post by XenophileJKO »

tgv wrote: post #264
XenophileJKO wrote: post #227

Earlier quoted context omitted.

I mean you say that.. but the counter point.. is maybe I don't like your lyrics and I want to make my own.

Then write your own. Maybe you're no good at it, but you're not going to improve by just copying them from the chatgpt console.

Right, I write my own and plug them into a song generation service. With as little trial and error, I get a song I listen to.

Re: DeepMind releases Lyria 2 music generation model

#376

Post by haswell »

giraffe_lady wrote: post #372
haswell wrote: post #329

Earlier quoted context omitted.

> I don't believe either of our individual experiences are really a good predictor of universal human experience in this area. Do you? No, and I said as much in my comment. My point was not that my experience is universal, but that I have direct experience with the failure mode of such an arrangement. Am I 50% of the population? 75? 5? I don’t know. But as I went though it, I met more people who’d gone through someth…

> No, and I said as much in my comment. You had that disclaimer but most of the rest of the comment was about your prediction that most people would be affected in the same way. Some friendly feedback for your future writing on the subject, I guess.

That’s fair, and I edited “most” to “many” shortly after I wrote it because the truth is that I don’t know what exact proportion of people will have the same issues. I had hoped the extensive context throughout the rest of my comments made that position clear.

I do feel confident that the number is higher than some of us think. Certainly I did not expect this to unfold in my own life, and through the experience I became aware of the many others who’d gone through something similar and were similarly caught off guard.

All of this will hinge on personal upbringing, background, support systems, life experiences, locale, etc. At some point I hope to write in more detail about the factors in my own life that I believe led me here after I’ve gotten fully to the “other side” of the experience.

Re: DeepMind releases Lyria 2 music generation model

#377

Post by waffletower »

eucryphia wrote: post #7

You know what the biggest problem with pushing all-things-Al is? Wrong direction. I want Al to do my laundry and dishes so that I can do art and writing, not for Al to do my art and writing so that I can do my laundry and dishes. - Joanna Maciejewska You could add music

With all due respect, even if you are Thom Yorke (actually definitely if you are Thom Yorke) I would probably be bored by your music if it is unwilling to expand and grow with the nascent possibilities of technology. While AI music creation tools tend toward vapid imitation at worst and noble democratization of hit songs at best, Google is providing a more sophisticated interface here which may allow the truly creative to discover viably unique and new music. I live for that. While I support those who want to grab an acoustic and sing at open mike, it would be a loss if all music had to sound like that and it never further evolved.

Re: DeepMind releases Lyria 2 music generation model

#378

Post by ndriscoll »

wavemode wrote: post #33
chii wrote: post #15

Earlier quoted context omitted.

Nothing stops you from doing art and writing, just like nothing is stopping someone today from riding a horse, running a marathon or rowing a boat.

Presumably, the idea is that AI drives down the cost (and thus, value / career prospects) of art and writing.

But it also has the potential to make the experience of creative pursuits better. e.g. have it listen to your playing of an instrument and give feedback on how to improve your technique. Or have it be an always available multi-instrumentalist partner for a jam session. You start playing and it just rolls with it and maybe inspires you in a way you wouldn't have thought of alone.

People are so weird about how to view ML/advanced signal processing. Don't look at things thorough the myopic lens of "prompt ChatGPT and it responds poorly". Look at it as an auto-complete, or a better form of on-the-fly procedural generation. Remember e.g. Audiosurf creating levels from your music? Make it happen on the fly. Maybe you could even create an interactive game where one person plays an instrument and the other does some kind of beat-sabre or dance-dance-revolution thing based on it analyzing and anticipating what's going to be played. The game scores you on how well the group was able to get into a groove together or something.

It feels to me like people get upset about ML encroaching on creative endeavors because they're not sufficiently creative to see how it could augment those fields and be a tool to make these things more interesting instead. Corporations will use it for cheaper slop, but slop was already what they wanted from humans anyway. For people that are actually interested in the artistic or social side, they'll have new tools.

Here's an example of something entirely off-the-wall/fun/creative enabled by these ML tools: Plankton covering Tool https://www.youtube.com/watch?v=Kc-PymIfv8M

Re: DeepMind releases Lyria 2 music generation model

#379

Post by taylorius »

waffletower wrote: post #377
eucryphia wrote: post #7

You know what the biggest problem with pushing all-things-Al is? Wrong direction. I want Al to do my laundry and dishes so that I can do art and writing, not for Al to do my art and writing so that I can do my laundry and dishes. - Joanna Maciejewska You could add music

With all due respect, even if you are Thom Yorke (actually definitely if you are Thom Yorke) I would probably be bored by your music if it is unwilling to expand and grow with the nascent possibilities of technology. While AI music creation tools tend toward vapid imitation at worst and noble democratization of hit songs at best, Google is providing a more sophisticated interface here which may allow the truly creati…

I would dispute that what is happening with AI music is well described by the phrase "expand and grow". In my view "democratzation" s a weasel word, meaning the removal of skill and effort from the process of production. The operator of such a system inevitably becomes more like a client, choosing between options served up to them. You could plausibly replace the human with a random number generator, and an external observer would not tell the difference.

Re: DeepMind releases Lyria 2 music generation model

#380

Post by amanaplanacanal »

otabdeveloper4 wrote: post #363
SequoiaHope wrote: post #169

Earlier quoted context omitted.

The reason you don’t have more time for music is not technological, it’s social. We have enough technology to offer a basic life for everyone for free, but we have not agreed that doing so is worthwhile.

> We have enough technology to offer a basic life for everyone for free Not literally "everyone". Someone still has to make the food you eat and the house you sleep in.

That's what automation is for. And there will always be people who want a much more expensive lifestyle than a "basic life".
Post reply on HN