My proposal is to equal-weight votes within the first 60 seconds or so (or even maybe 5 minutes, but not so long as to remove the value of being early). 4 hours is much too long in Internet-time (news cycles and attention span). Currently there is a large incentive to vote first, which will incentivize fast-voting by bots (even more so than now -- most of the early-adopter bots self-impose a delay). Even your own post may be upvoted by a bot before you are able to do so! By limiting the early-voting incentive to a human timescale there will still be bots but they won't have a major advantage over humans who are actively curating posts.
I'd also suggest effort into designing an incentive for downvoting. The deterrent to swarm-upvoting that is disconnected from the merit of the post should be fundamentally the threat of being downvoted later by smarter voters, which should then be rewarded (perhaps half of the reward that would have gone to a votes of previous-upvoted post instead goes to the later consensus downvoters). The possibility of being downvoted and burning vote power already exists, but it relies almost entirely on altruism, which works to a point, but likely doesn't scale very well.
This kind of knee-jerk reaction to a perceived problem reminds me of the troubled days of Bitshares, where things like "The Merger" happened very quickly and turned out to be terrible ideas..
On topic: I was thinking more or less the same thing: reduce the "first-vote" advantage by allowing a grace period like you say, the exact length can be tweaked later.
I'd also like to add that I think removing curation rewards is a very bad idea, as it removes the major incentive for most people to come here and vote on posts. Everyone can't be content creators, but if they're the only ones being rewarded we will see a quick concentration of Steem Power in the hands of the top posters. Come to think of it, why should content creators be rewarded with 50% Steem Power? Creating content does not make them good curators, so to me it would make more sense to have a different distribution: 25% Steem Power and 75% Steem Dollar for posts, and 75% Steem Power and 25% Steem Dollars for curation for example.
Before removing curation rewards completely I would highly recommend tweaking the current system, perhaps using the idea above, combined with a tweaking of reward distributions between posters and curators, 25% to curators and 75% to posts for example instead of 50/50.
Agree about knee-jerk reactions (as I said elsewhere there is hardly even a userbase yet, certainly not a representative one). Interesting point about allocating the rewards differently between SD and SP.
If equal-weight votes in first x minutes/counts, the strategy of bots would be just following the whales' or masses' votes before the period ends.
There is far less gain to that than trying to vote first. At best you get a small share along with many other people/bots. May not even be worth the consumed vote power.
I can't see how any system that differentiates between votes a very short time apart does anything other than heavily favor bots over humans. This is independent of other voting factors.
A small share? Aren't bots mostly used by whales?
//Edit: I see no issue on bots used by minnows.
Holdings have a return on investment in this system. That's fundamental in how it is designed, and will apply to any system whatsoever that avoids sybil attacks by weighing rewards by stake. The rich get richer is not, by itself, a problem, as long as it doesn't negatively affect categorization of posts, or appeal to a wider user base.
By "small share" I'm comparing it with the enormous share that the very first votes get in the current system relative to the later votes. That strongly incentivizes bots.
In actual practice all the current bots run by whales delay their votes 30-120 seconds and are rarely first, in fact often quite late in the mix of votes. As hard as it may be to believe, current whales are not trying to wreck the system for their short term gain. This by the way does not include me at all; I don't run any bots.