This is not entirely true. It has a low impact, but is still reasonable. Nature Scientific Reports is still published by nature and does peer review. Granted the content is not on the level of nature itself.
They are supposed to do very light peer review just trying to check if the article is technically sound. Apparently even that review process is broken.
Technical soundness is not light peer review. It's the core of peer review. Everything else is editorializing and selling "importance".
As a matter of fact you can get plenty of crap into the higher tier Science/Nature journals if you just tell a good enough story about importance. You can leave the real work to disentangle whether the story is true to others that can publish in lower impact topical journals.
Generally speaking I believe Science and Nature, and the role they occupy in academia, are a major problem for science. They are huge contributors to overselling and hype driven research.
And junk science? Please provide links.