Top
Best
New

Posted by hmokiguess 6 hours ago

Google Flow Music(www.flowmusic.app)
95 points | 64 commentspage 2
twobitshifter 2 hours ago|
Ok, someone explain the use case for this? Jingles? Making a song about your friend / sig other? Are people thinking they are going to sell these songs and create an AI artist?
kelseyfrog 1 hour ago||
I've summarized Supreme Court cases into Broadway musicals. The thing about memory is that novel input increases retention. So now I know about grouse hunting and explosives that fall off trains and their constitutional implications.

Another was a set of songs that helped me emotionally regulate on the drive home after couples therapy. The lyrics contained grounding exercises that helped maintain awareness and presence and contained mindfulness practices.

Both did their job, but they were also music for utility, not necessarily for artistic enjoyment. So it's not entirely an apples to apples comparison.

_sys49152 57 minutes ago||
turning class notes into songs for study purposes sounds like genius. never wouldve thought of it, but i could definately see value. catchy 1950's style radio commercials advertising highlights of case law to remember for an exam coming up.
mh- 43 minutes ago||
I think you've got a great app idea on your hands.

By the time I finish writing this comment - yours is 10 minutes old - someone will have vibe coded one, probably.

Also feels like an easy feature for someone like Suno to add, to help subscription retention.

But something like NotebookLM emphasizing subtle mnemonic devices set to music..

lern_too_spel 2 hours ago|||
Marketing jingles for video ads.
torben-friis 2 hours ago||
people like to think they're successful but they don't like effort. So this will let them pretend they're artists.

Also, probably someone will game an algorithm to get revenue from a bajillion tracks of lofi slop.

SmirkingRevenge 55 minutes ago||
Yep, I'm going to say the overwhelming use-case will be slop-4-revenue.

Slop is starting to dominate uploads to some music services, so I think it will only get worse from here

ttul 1 hour ago||
Here's my personal take on what I'll call the new realm of "AI art". Whether it's prompting a music model or an image model, there is a huge space for creative output, limited only by the human imagination. Sure, tossing in a single prompt and letting the model crap out something will produce "slop". But if you pour your heart into exploring the high-dimensional landscape of the model, you can find truly amazing stuff. This is no different than exploring the creative landscape of music, photography, and other forms of art in the pre-LLM era.

I find that people who rush to negative judgement of LLM-generated art are not going far enough in the creative process to properly judge just how much juice there is to be squeezed out of those 50-billion-dimensional spaces.

unicorn_cowboy 3 hours ago||
This is their recent acquisition (producer.ai) rebranded.

Before the purchase, the quality of generations had been going down for a while (IMO; subjective and anecdotal). I tested multiple iterations of their chat interface and was never thrilled with its ability to actually understand or adhere to prompts. I had liked their previous (Suno/Udio-like) iteration (Riffusion).

Curious to hear how it performs for people now and whether anything has improved.

jmull3n 2 hours ago||
It can generate something well produced, but it's really bad at applying taste or direction in the way a human does.

The workflow feels wrong. it should be closer to a DAW with chat, where the model outputs stems, samples and arrangement parts instead of one finished track. Then you could target a specific sound, section or idea and actually develop it.

filoleg 2 hours ago|
I agree with your DAW UX suggestion very much. I think the writing is on the wall, and Suno is doing exactly that with their Suno Studio.
mh- 40 minutes ago||
I think this wasn't done earlier because the Suno (etc.) models couldn't output stems.

They could attempt messy stem splitting like all of the other tools have done for a few years now, but those aren't really usable in a production setting beyond small samples you were already going to chop/distort.

florilegiumson 3 hours ago||
AI music generators are not my cup of tea, but it was fun trying to get it to create this fever dream machine. https://www.flowmusic.app/space/2dc0e63f-a27c-4f4f-91c9-60e8...

I especially love the glitchy ui sounds, although I suspect it's hardly intentional.

MrZander 4 hours ago||
Really bad at prompt adherence. Was trying to get it to compose a solo old time banjo piece. Couldn't get it to stop adding in backing instrumentals at all and it sounded too much like bluegrass style.

"solo banjo instrumental, strictly no other instruments" ... ten seconds later: drums, a fiddle, and a guitar join in.

p1mrx 58 minutes ago|
I tried to make an orchestra where they smash household objects, and a synthpop song where all the lyrics are burped. Didn't work. Wake me up when we're in the future for real.
deferredgrant 4 hours ago||
I can see the appeal if this ends up being good at iteration rather than just first-pass generation. A lot of AI music products look impressive for five minutes, but the real test is whether they help someone get closer to a specific thing they actually wanted to make.
pavel_lishin 4 hours ago||
Given how much Google lives to mash their offerings together, and then sunset them, I live in fear of them killing Google Youtube Music (or whatever it's called), in favor of combining functionality with this, and having my music cycle between my actual library, and bespoke AI-generated stuff.
mh- 35 minutes ago|
If I were a PM at Google trying to connect those two products, the far more obvious approach is the end of the creation pipeline having an "Upload to YouTube Music" button.
dabinat 5 hours ago||
I’m a little confused about the pricing packages. In what scenario would being able to create 600 songs a month (20/day) be too few?

I could understand if this was an API that people built products around, but it seems to be geared directly at consumers.

smallerfish 4 hours ago||
If it's anything like suno, it probably takes you 30 to 40 attempts to dial in what you were looking for. (And don't get me wrong, the results can be great with suno - there's just a lot of trial and error, and dice rolling.)
numpad0 3 hours ago|||
I don't know anything about these AI tools, but it seems to me like, the yield rates of all these AI media generators are exactly in the range of that of lootbox games. Kids "pull" it like slot machines for set prompt, keeping no more than 1% of outputs. The rest is just thrown away, only potentially useful as negative data. So 600 per month total is probably like just couples per month usable.
janalsncm 2 hours ago||
That’s a huge amount of messing around to get those handful of songs then. If only 1% are good, you’re pulling the lever 100x more than you should need to.
mh- 31 minutes ago||
Putting commentary about AI media aside:

How many iterations (arrangements and recordings) do you think a typical Billboard pop song goes through before it's ready for a final mix and mastering?

Go find a YouTube of someone doing this work, it is kind of mind blowing. Given how expensive studio time is, you realize why it costs so much for a popular artist to produce a polished album.

999900000999 4 hours ago||
Eminem allegedly has hundreds of songs in the vault.

Odds are for every 200 ai songs you generate , 2 or 3 are decent.

Anyway. UMG will probably force you to sign over training rights in future record deals.

The models still can't rap. Sounds like if you asked someone who didn't know what rap was to read a script

minikomi 3 hours ago|
The descriptions generated from the prompts are almost always great, but the generated music is always terrible. The sound pallete seems so limited.
More comments...