I have no issue in principle with labeling music created with AI.
If someone types a prompt into @suno , downloads the two minutes it spits out, and uploads them as-is to a streaming platform - sure, I get it.
But I also know creators who take AI seriously (not just in music) and use it to make incredible things they simply couldn't make before.
They don't take the first output. Or the second. Or the 20th. They use judgment, taste and intuition. Take a piece from here, combine it with something from there, iterate, edit and polish in external software until it feels right. That process can take days - and quite a bit of money, by the way. I know because I do it too.
So should that simply be labeled "created by AI"? I'm not so sure.
There was a human creator making dozens, sometimes hundreds, of creative decisions along the way. Why shouldn't that contribution get credit?
And the connection between AI and "low-quality content" bothers me too.
I've heard plenty of terrible music made 100% by humans - clashing keys, vocals that don't sit right, awful mixes and various other musical disasters. I'm not convinced those are somehow higher quality than some AI-assisted tracks I've heard that absolutely blew me away.
Earlier this week I read that DJ HUGEL got the inspiration for his latest hit from an AI-generated track. So being inspired by AI is fine, but music created with AI suddenly belongs in a category we need to warn listeners about because of "low-quality content"? That doesn't seem quite fair.
And if we're really doing this in the name of transparency, let's go all the way:
Let's label tracks that use software synths (VSTs) instead of old-school analog synths.
Let's label tracks where the drums came from a sample library instead of an actual drummer.
And while we're at it, let's label the thousands of tracks released under the names of famous DJs and producers that were actually made by ghost producers.
Because ultimately, I think the question shouldn't just be: What tool did you use?
It should also be: How much of you is in the work?
@Repbyronin@johnsummit I used MP3s for the mortals, but since weโre being pretentious: 44.1 Kilohertz is a sample rate, not a format. If youโre going to be an aggressive audio nerd, the least you could do is be correct
Some sounds you'll recognize if you were alive in the 1980s. All of the sounds are created by interacting with these then-everyday objects.
[๐น rerunthe80s]
@theRayW@AlexFinn Honestly I don't see a huge difference between this result and what can already be done using Opus 4.8 with the Higgsfield MCP. If the text in the app wasn't gibberish, I'd be impressed
Claude Fable 5 is just insane.
Since it was released yesterday, my company's ARR jumped to $100,000,000, things that used to take me months now take a few days - a week max, and we've already completed our 5-year roadmap ahead of schedule.
A lot of people think AI creation is all about the tools.
New model drops.
New features.
New workflows.
Those things matter.
But after spending time around hundreds of creators, weโve noticed something:
The creators improving the fastest arenโt obsessed with tools.
Theyโre obsessed with making things.
They ship ideas before theyโre perfect.
They experiment more than everyone else.
They arenโt afraid to make something weird.
The biggest difference isnโt access to technology.
Itโs the willingness to create consistently while everyone else is still waiting to feel ready.
Thatโs what weโre trying to build with Anthum.
Not just a platform.
A place where creators can learn from each other, challenge each other, and push the limits of whatโs possible with AI.
The tools will keep changing.
The creators who keep building will be the ones who matter.