One difficulty here is that karma ranking boards still wouldn't optimally reflect status. Suppose PersonA did a ton of good work on Less Wrong and got 10,000 karma, and then donated all of that karma toward making some great project highly incentivized. This person has done two wonderful things - all the wonderful work and incentivizing more wonderful work - but they have 0 karma to show for it.

This is the difference between us and what.cd. There are so many albums in the world filling a request might mean satisfying the desire of only one or two people. But major contributions to site code are collective goods. Because they are collective goods people will often be unwilling to sacrifice karma and will instead wait for someone else to fund the projects. Unless we're going to name site features after the posters who fund them ("The Lukeprog Lawrence, KS Meetup Group!") it would be more effective to have a public funding model.

Posters should have a set monthly "Karma Vouchers" that they can spend on requests. When the request is fulfilled the person who fulfilled it is accorded in real karma whatever the task was given in vouchers. The amount of vouchers one gets can be a karma incentive as well- say you have to get to 500 karma before you get any and after that it increases occasionally as you level up. You cannot claim your own vouchers. Karma vouchers can also be put against tasks (i.e. you can downvote with the vouchers) for an easy way of disqualifying unpopular site features and preventing people from gaming the system.