Traditionally, and as is reflected in salary hierarchies, "product testing" is far lower on the totem pole than "development" in Silicon Valley.
I think the AI age starts to level that, if not even inverting the value completely. When anyone can develop an app, especially in the mobile context, the bottleneck isn't the development itself but an *actual* human being with *actual* thumbs using it on an *actual* smartphone out in the real world. No LLM will be able to test that.
So the bottleneck specifically is having real people test it in real environments *and* report back useful feedback to the human/AI developer centaur.
That's a role that I feel I've played in Nostr... Some developers welcome it, some don't, but for every 1 of me who's providing actual feedback, there are 10 users who download your thing, hit the same friction point I did, and give up.
