Creating a product is hard enough; getting it discovered is even harder

Thanks. Though I'd argue the wins are the less useful half — they're pleasant to read and almost impossible to act on. The posts I've actually learned something from here were people describing the thing that didn't work, with the numbers left in. Cheaper to learn from someone else's version of it than to repeat it yourself.

Thanks. The most useful feedback actually predates the feature — MV generation exists because people asked for it, not because it was on a roadmap.

We only did music at first. What kept coming back wasn't a complaint about the music itself; it was people finishing a track and then asking some version of "okay, now what do I do with this?" Enough of them said video specifically that it stopped reading as a feature request and started looking like the actual bottleneck. That's really what the post is about — I just hit the same wall one layer up.

On the MV feature itself, it only went live fairly recently, so there isn't enough usage behind it yet for me to say anything I'd stand by.

Good question, and the honest answer is that it's too early for me to have a number — this went live recent.

But I know which number I'd be lying to myself with. "MVs generated" is the easy one and it means almost nothing; people try a new button once. The two I actually care about are what share of finished tracks go on to become a video at all, and whether the same person comes back for a second one. The first tells me the gap was real, the second tells me it's a workflow rather than a novelty.

Fair thing to ask. Right now, badly — it shipped recent, so anything I quoted would be noise.

What I've decided not to count is raw generation volume, since a new feature always gets a curiosity spike that tells you nothing. The signal I'm waiting on is repeat use: one video is someone poking at a button, a second one from the same person means it fit into how they actually work. I'll know in a few weeks, and I'd rather say that than dress up a first-day number.

Music creation itself you can try for free — that part isn't gated, and I'd rather people hear what it actually sounds like before deciding anything.

Beyond that, some things are paid, and the honest reason is cost rather than strategy: generation and rendering both burn real compute, and a feature that's free at ten users stops being free at ten thousand. So the heavier features sit behind the paid plans, while some of the lighter tools stay free for anyone to use. Where MV rendering settles I'm still watching — I don't yet know what it costs at volume.