I'm not sure I agree. The system certainly isn't optimal but results aren't just dumped into a vacuum. Something is only useful if people can build on it. Even if negative results don't get published, even if it isn't optimal, by virtue of future positive results building on past things that did reproduce you get forward progress.

Regardless, I don't think that's at odds with my original assertion that becoming a venue for publishing negative results would undermine the "point" of Nature.

The missing link isn't a venue in which to publish. It's funding to do the work in the first place. Also funding to spend the time writing it up when you find that you've inadvertently been tricked into doing the work while trying to get something that builds on it to work.