I feel that reducing curation rewards would discourage people from powering up.
It could definitely impact this, as I mentioned in my post. It's really hard to say how much impact there would be, given curation payments really aren't that high for the average curator.
Ultimately, they can't be but so high, because all curation rewards are paid from 2% of the coin's inflation. The only reason it rises above 2% for individual curators is that some are curators are getting paid more than others. Predominantly, this is the probably those using auto-vote bots, which is unfair and not really sustainable in the long run, IMO.
Auto votes are mostly done for sniping purposes. Most of the auto votes are in such a way that the snipers pick authors who get rewarded with big votes and vote them as early as possible at the 3-minute mark.
This is true, but like it or not, that's actually by design. An important point to note is that voting earlier than 5 minutes returns a portion of the curation rewards to the reward pool, so that other curators can get it. Voting at 3 minutes actually returns 3/5 of the curation reward for that vote to the pool. The point of this design is to encourage curation of new authors, rather than everyone just piling on to the same list of existing authors. I don't think this theory was completely bad, but it overlooks one key detail: curation rewards just don't pay enough to justify manual curation on an economic basis. So paying them a little bit more for finding new authors still doesn't justify the extra work. So it doesn't achieve the desired goal in practice, because on an economic level it's still easier to auto-vote and get less rewards, than to do the huge amount of work to find new authors to vote on and get the potentially slightly larger reward.