As someone also practically not contributing at all at the moment, my experience from other projects is that when you’re starting out helping out on any large software repository you just always create more work for the person who needs to integrate/review it, at least until you’re trusted enough to merge others’ work yourself.
It’s a fundamental weakness of the merge/review process, because good review inherently means duplicating the “real” thinking work, + some communication overhead. So for as long as you can’t make up for this work duplication by merging other people’s stuff yourself it just adds additional work for the group of people who have merge permissions.
You can try to get around this by contributing something important that your reviewer would have had to do themselves in the short term anyway, but that is by definition something critical, and then makes the review harder again (and probably more likely to create problems, because nobody reviews code well enough to actually duplicate the work).
That, or you surprise them by solving something tricky that they had no idea how to solve in an elegant way, but that’s a very hard thing to do, and nobody will be able to point you at things like this, because by definition nobody knows an elegant solution exists.
It’s really annoying, but it’s a tricky short-term vs long-term situation. There is rarely a good “beginner” issue, almost every project I’ve been on struggled defining those (and spent more time labeling tickets with it than they got out of them being solved by “beginners”). On the other hand, getting up to speed (and enough trust for merge access) is incredibly valuable long-term, and people realize this, so I’ve rarely been in a situation where a maintainer wouldn’t take their time with me.
It is unfortunate feeling like you’re not really helping despite your best efforts in the mean time, though, and I personally agree it is a hurdle to contributing to nixpkgs. I think the trick lies in making people feel like they’re doing helpful work rather than actually making them do helpful work 
Reducing the barriers to “trust” can be helpful too, of course, but defining appropriate levels of trust is a can of worms in its own right.
