clearly you need to download more RAM

give Curtis my regards

I like it but in general it seems to work better for closed/gated communities

genuinely wish it were easy to just let yourself dream again, especially at this age.

there's a number of ways I think this could go and I'm not confident that any of them will happen. for me personally, I've always found certain "imperfections" in amateur music interesting, sometimes even enjoyable. maybe in the future, those that value human-created music will feel the same

final note - to bring it back to the original point, the post also mentioned that the remaining 20% of students were FAR above in quality than suno (pareto principle at work I suppose). more than ever, musicians need to hone their craft and develop a unique voice to even have a chance at making it

again, maybe it's my lack of knowledge in prompting/genre. or, even just not enough time on the platform - there's a lot of music there, so maybe there's some hidden stuff there that would meet my standards of what constitutes genre fusion. more testing will be required

genre fusion seems to be a different matter. between my own limited testing and looking through the explore section, I think the AI leans too heavily on certain keywords in the prompt, leading to a very lopsided output being dominated by elements from one genre and few of the other

out in the wild, I expect most people would not really care or notice common AI indicators, especially if the music is used in support of other content (i.e. bgm). this will only become less distinguishable as models continue to improve, especially for genres that have enough data to train on

I enjoyed reading this series of posts about sunday - perhaps they might be interesting to consider (link to pt1):
sequence-trotter.tumblr.com/post/7644537...

if I don't see either of these by the time I get back from travel I'll do it myself

lia...