In 2024, I started generating music with AI models. I've generally found that the latest models are now capable of outputting any sound that is possible. The limitations of AI music are now generally caused by my selecting the wrong inferences (it usually takes 600-1000 runs plus editing), not understanding song structure, or not understanding what listeners like and dislike.
Every time I publish a new release, which I've been doing about once per month, I will post it here and link to its version on Soundcloud. If any individual song receives more than 10,000 listens before the end of the year, this market resolves to YES. Otherwise, if January 1, 2026 arrives before that occurs, the market will resolve to NO. If Soundcloud ceases to exist, its metrics will be saved and added to a new site, where the music will be copied to.
The songs can be posted at any time, and the listens can occur at any time. Copies of the song posted elsewhere don't count. If even a single song's counter rises above 10,000 before year's end, the answer is YES. I will pay for newer models as they become available.
So far, in order of release, the songs have been achieving an increasing number of views (and higher Manifold ratings) as they improve in quality: the five songs posted so far have 41 -> 81 -> 208 -> 374 -> 3684 (as of January 17).
Current URL: https://soundcloud.com/steve-sokolowski-2
RELATED MARKETS:
1,000If "SPINNING PLATES OF MEANING ON A NEEDLE MADE OF LIGHT!" does not resolve one of the markets to YES, then I don't believe there is anywhere I can go from here.
The song is by far the best "I" made, it's original, and it's the first of its kind designed by a being completely alien to us. If it does not at least spread across the Internet due to its significance, then it's difficult to imagine how any song possibly could.
New market available, because I'm confident in this one. I reserve the right to bet YES.
I find it interesting that the "Manifold rating" market is rated higher than this one.
I personally would bet a hard NO on the Manifold rating market (if I thought it right to bet NO) and am currently betting YES here. All of these polls have responses with reasoned comments until the very last minute and then hours before they close the result flips with people voting 1 - this happens to the AGI and NHI polls too.
I don't know why that is, but the timing leads me to suspect that somehow these polls are being manipulated or gamed and aren't valid.
@4fa I think that song will probably hit the mark eventually as a "back-catalog" release, but I'm not promoting it myself at the moment. I haven't had any problem getting people to listen to it where it is promoted, but it's possible to exhaust the number of extreme claims one can make (as everyone knows, extreme claims are the only way to get attention.)
The issue is that I created it to be technically impressive and showcase what can be done; however, it's not in a genre that is currently popular on mainstream radio. So, I will take the numerous comments that people made about that song into account.
Then, for the next one, I will create a song that is more formulaic and is more likely to appeal to people who listen to the radio and promote that with extreme claims, and I think that will easily hit the 10,000 mark.
My goal is to either get a song on the radio by the end of the year, or to learn enough such that I hit the current models' wall, and therefore am in a position to do so immediately when the next model is released.
So far, "Six Weeks From AGI" is around 2000 listens, with 1500 in the past six hours. The listen rate seems to be accelerating exponentially in the past hour; I can't figure out who is promoting it, as I haven't paid anyone.
Edited to point out that the number is now 2700, for people in NO positions.
I put my own mana into YES, since this market is based on a quantifiable threshold.