If I'm understanding you correctly, you're saying large early votes will deter later voters from coming, but that's already a thing that happens.
It's the next order up in the chain. Large early votes deter later voters from coming, now and later. But right now authors don't care because large early votes benefit them. This change means large early votes benefit no one, but still deter later voters from coming - so authors will want to discourage them.
It might help to think about it from the perspective of a new user who never knew that the current system existed. They just come in and are told that voting on a post in the first fifteen minutes will disadvantage everyone involved with the post, themselves, later voters, and the author. How do you explain that in a way that makes sense? "Well, the system used to be dumb, and we got rid of the part people were taking advantage of, but we left a big chunk of the part where it's dumb just to make it harder for you to understand."
It's like the people who designed this saw the problem as authors getting more rewards but not as voters getting fewer rewards. If one is true then presumably both of them should be.
About the new user perspective: Yeah, I see this as a problem with the early vote penalty in general. Whether it goes to author or pool is just confusing, no ways around that.
Actually since you mentioned it , I'm now thinking that getting rid of it entirely and making curation distribution not dependent on timing at all would certainly be better. But done in a way isn't stake-lopsided. (Sqrt(yourshares) / sqrt(shares)) seems like a pretty good equalizer. I'm sure large stake curators would rage against that but I think I like it.
But about the proposed system: I think authors can discourage all they want, but if they are a popular author, you know that large voters are fighting with each other for curation and will be early voting anyway.
I am 150% behind any proposal that makes curation less stake-weighted. I'd actually like to have a little fillip at the bottom of the curve so that the smallest votes don't get dusted even if it means they get more than their fair share. I think it would help new user acquisition if their votes were visibly making curation rewards even if it's 0.001 SP at a time.
The huge flaw with that proposition (and any attempting to decouple vote value with stake) is it incentivizes the creation of millions of accounts to game the system. Basically, you end up with a system that is still stake weighted, but instead of tokens supplying the value, it's actual accounts that you collect. Whoever controlled the maximum number of accounts would be able to game curation the hardest.
I'd imagine this will be a big issue with ONO and any other platform that attempts to use a more "socialistic" approach. The only solution I've thought of (which is presently impractical because the technology implementation is currently lacking) would be a DNA interface with technology that verified one's unique identity. That way anything from a purely stake-weighted to purely identity-weighted (even) scheme of voting would be possible to implement.
But this DNA recognition technology would have to be mass adopted, which would take a lot of time since so many would object on various grounds. I predict we will be heading that direction in the coming years, but whether it takes 20 years or 100 remains to be seen.
We can move a long way from weighting curation at the square of stake before we get to the point where it's beneficial to spread stake out, though. For that to happen you'd have to get enough below linear to cover the operations cost. If we went to n^1.1 there'd be enough wiggle room to undust the plankton.