In philosophy (which is my academic background and profession), 'moral saint' is used to denote a hypothetical agent whose every action is as morally good as possible. I don't think people need to be that good to make linear rewards work, but clearly more people would need to be 'better' than they are now (otherwise we wouldn't be having this conversation).
Given my interest in non-ideal action-guiding theories, I should know better . Economic systems, systems of governance and government, conceptions of justice; none of these things should be designed for the way you wished people were - you have to design them to cope with the way people actually are.
This isn't an endorsement of some sort of moral determinism (so you are still not off the hook). Humans, and the societies they exist in exhibit and incentivise a range of levels of selfnessness, so I'm not peddling some idea the kind of assholery that steemit rewards is somehow reducible to a pre-theoretic notion of 'human nature'.
Incentives. I ponder John Rawls' theory of justice - he was quite happy to use economic inequality to incentivize desirable behaviour. For example, even though equality is good, inequality in the form of paying surgeons more than average is better. That's because it encourages people to become surgeons - but this ethical advantage only true if there's reasonable equality of opportunity to become a surgeon, and if the benefits of having surgeons are reasonably equitably available to all members of that society (via the notions of income redistribution and universal healthcare that so many people around here find so triggering).
This isn't about progressive taxation or other things that give voluntaryists nightmares. But the principle is the same. We want big investors and whales to exist, but we also want their existence to produce a beneficial effect for smaller users. Actually it's more than 'want' - we need them to produce benefit for smaller stakeholders, otherwise we'll end up right where we are now: heading toward a distinctly uncertain economic future.
So yes, curation does need to be incentivised. I am not sure what exact curve the rewards need to be set at, but I can see the argument for changing it. On top of that, we also need to try and educate and encourage people to not game the system, and for whales live up to the responsibilities of being major stakeholders.
Trouble with all this is that you need to get the witnesses, who are controlled by these large stakeholders, to agree to a change that means these stakeholders either make less money, or have to put more effort in to make the same amount of money. It might not technically be regulatory capture, but IMHO, it's pretty analogous.
One thing that did occur to me would be to change how witness voting worked, then watch the problem solve itself in an organic, 'bottom up', way. We can't have one account = one witness vote, for obvious reasons. And with the ability to use bidbots to quickly raise one's reputation, weighting voting by rep won't help much either. What you could do have the stake-weighted witness voting work in a non-linear way, so that while the ability to choose witnesses grows with stake, it grows less with each additional vest. You could do this with either a log function - tweaked to not allow hordes of tiny account to choose the top witnesses - or maybe a cube root function - something like that, my calculus is pretty rusty. Whatever the case, if witnesses were suddenly more beholden to people in the minnow to dolphin range, then maybe they'd suddenly they would have both the motivation and ability to do more than act in the short term interest of their biggest voters.
Or not - it's just an idea.
So there you go - I agree this needs to be addressed systematically. But I want to be clear on one thing: If a system incentivizes bad actions, this does not magically mean that bad actors have no moral responsibility. For that to be true, it would mean that moral goodness or badness can be purely reduced to what is incentivized or not. I'm not saying that isn't the case, but you sure as hell haven't established that position to my satisfaction. Self-voting, delegating for profit rather than curating, low-effort spam, multi-thousand account voting rings - all of these things are currently incentivized, and all of these things (to a greater or lesser extent) also deserve condemnation.
Unless someone's steemit-assholery is needed to pay the rent or feed their kids, I'm not interested in excuses. I'm glad you are focused on solutions (doubly so, considering how much more influence you have than the average person here). But, incentives or not, you are still responsible for your choices. Don't expect people to let your behaviour pass just because the system rewards it, or 'everyone else is doing it'.
RE: Embracing Linear Equality on Steem: Unlearning the Sucker & Maximising the Arsehole in Me