That's a very definitive statement. But perhaps it is untrue?
It would seem yes, that it's good for scientist careers.Are "good" careers versus "bad" careers good or bad for science? In other words, is it useful to rank scientists when allocating grant money? Do good scientists make better use of the money than bad ones? Are there other ways you could suggest to rank scientists?
Perhaps there are other groups that benefit as well? What about people issuing grants? Is it useful to them that there is industry recognised feedback regarding the scientist and her work? Is this useful when allocating limited funds to an unlimited demand?
What about people following the science? Let's take industry. Say I want to make a commercial product. Should I start by paying attention to the field, understanding the science? Or is it ok to just read any unvetted thing?
What about media, and by extension the population? Can a media outlet run a story based on "something someone wrote on the internet"? Or should they prefer credible sources? Should the public have some interest in the truthiness of something? Should the public (via the media) understand the difference between a result published in the New England Journal of Medicine, or what my homeopath down the road published on their blog?
What about say doctors keeping up with "current ideas"? Should they believe everything, every "study" posted on a blog? Or should there be a system of gate-keeping, sifting the valuable studies from the chaff? Presumably knowing that a study is well formed, and not paid for by say big pharma, might lend it more weight?
Of course publishers benefit (financially) from this system as well. But they don't matter right? So we'll ignore that. But even if you remove them from the equation, I'd suggest that "science" does indeed get value from the system, beyond just scientist careers.
Now, could all these benefits be gained in an alternative way to expensive journals? Almost certainly so. But in order to build such a system it's important to understand the strengths of the current one. A "simple" law might solve one part of the puzzle, but at the same time have very foreseeable consequences in other parts of the picture.