"Hard to detect false info" seems like maybe you're using the current paradigm and not really opening up to what a mature, decentralized rep system might offer. There's really no way to know how hard it would be, since we don't even know what it looks like let alone what tools and services will emerge to help sort it all out.
It's easier to defend something that's true than it is to defend something that's false. So over time as techniques improve and more energy gets put into it, I don't see why false information would have much chance.
Anyways, whether it's hard or easy to sort out false info you still want whatever rep system does the best job of it.
Judging people and people carrying reputation (and being regarded as honest or dishonest) is a part of life. How to do it (and whether it should be a centralized mechanism backed up by violence or a decentralized mechanism backed up by ostracism) is the question here.
To just challenge the usefulness of reputation.... I'm not sure what your point is. I'm sure you use reputation in your daily life, and I'm sure you wouldn't want a community where there's no way to differentiate the rapists from the heroes.