Marketing is a fact of life, I think. Of course, there's good marketing (making people who might enjoy your product aware of it) and bad marketing (annoying people), but if you hope to reap the benefits of having your work in front of a lot of people, you do have to reach them.
What OP says, and you miss, is that some people are not interested in reaping benefits. They are interested in doing work.
They are constrained by time and money, so yes, sponsorship can help their projects, but they are not in it for the reaping. If anything they will be happy to see a lot of people reap benefits from their work, not the other way around.
How many are like that? I don’t know. Maybe not a lot, but some for sure.
There are tons of people who like to just go do stuff, and that's great! But if you don't promote it at all, certain avenues are not open to you.
I was more thinking about contributing to other projects. What incentive do I have for fixing an issue in someone elses repo?
With this sponsorship model, the benefits will only go to the dev whose name is on the repo.
It's going to be very hard to market yourself as the guy who fixes random issues in a bunch of unrelated projects.
"Unfortunately, since it's opt-in it didn't move the needle at all. I think it directed a few hundred dollars to Sindre's Patreon. But that was it."
I am currently looking at TideLift to see if there that is a more viable option. But to your point, it might all be "way too little for the effort required".
EDIT: adding my link since I guess that is what others are doing in this thread. https://github.com/sponsors/wesleytodd
In my opinion, it's essentially "gig economy" at scale. GitHub should start publishing raw data on how much money developers are getting... median etc and how close that maps to a developer salary etc
It probably helped that I watched his video demos on YouTube and such and saw there was a person behind the project.
I'd imagine the project being something exciting or something that makes programming more fun helps a lot as well. An extremely useful but boring project probably won't inspire as many people to donate, even if that's not really fair.
If you're a person that prefers to just do some work and not talk about it much, this kind of sponsorship model wont work.
To me, something like bountysource seems like an awesome model: people could all chip in to support a feature request or an issue, and anyone who is familiar enough with the subject matter can do the work.
(Unfortunately the reward mostly seemed way too little for the effort required -- it would be a nice bonus if you would do the work anyway, but I don't think you could make a living by implementing issues from bountysource)