This idea has been around over 20 years. It dies every time because major lab PIs, usually in US, HATE the idea of not being able to gatekeep research publications in journals of "high impact". This impacts how institutions are assessed, because, God forbid people actually have to read the papers. This feeds back to Editors, so the number one factor that influences Editors now is zip code.

If we went to a simple repository archive, with transparent peer review, then no one could imply their research is more important because of where it was published. We would let citations determine impact. Science publishing has always pushed the idea that if Einstein drove a Honda, everyone who drives a Honda is a genius.

Meanhile, The Lancet (JIF 105) took 12 years to retract a paper linking autism to vaccines, when it was clearly fraudulent from day one. Nature, Science, CELL, just stopped retractions, at best, they have "statements of Editorial Concern". This high JIF model is why Alzheimers research has stalled behind a flawed hypothesis only reinforced by fraudulent work not retracted for 25 years. Some people, like the President of Stanford, rose to the top tier on fraud and journal gatekeeping.

2020 saw the world arguing over ivermectin based off a paper "reviewed" overnight, with the journal Editor as an author. The journal 5 years later refuses to prove the paper was peer reviewed at all. 3,400 citations.

Then we have predatory journals that will publish literally anything for page charges. Examples:

Get me off your fucking mailing list.

and

Chicken, chicken chicken chicken, Chicken? chicken. (Cited 35 times)

The difference is the peer-review process. Without a good peer-review process, those journals won't have a strong cite score and so they will be considered unreliable.