Yeah, I see that risk. I guess the hope would be that you’d find some way to incentivise ‘good’ behaviour — so commenting on famous people’s works would be inherently worth less than commenting on not so famous people’s works. And the hope would be similarly trolling would be disincentivised. One way to do that — which is independently desirable — would be to have a real name policy. To the extent that the proposed system would replace the current academic system, which of necessity operates with a real name policy (which attaches to publications, for example) in order to allocate credit (you publish papers to have a line of your CV), you might hope that trollish behaviour would be as unreasonable as, for example, yelling obscenities from the back row if at a conference.

The risk that everybody would just comment on famous people’s work is a real one though. I kind of think it should be possible to engineer a system that can get around that, though.