The point of distributed computing is to welcome more computing power, not to turn it away by artificially restricting the number of available workunits.

Absolutely agreed! However keep in mind this is just a single project out of dozens. The point of the project is explicitly stated as a project which provides a mag boost for new users while helping new projects complete work and gain visibility. This means that anyone with a lot of processing can offer some up to the GRCStarter@home project with the intention of helping new users instead of gaining Mag. Once new users learn enough from GRCStarter -- or perhaps there is a time limit for low mag CPIDs or some other mechanism -- they are encouraged to move on to a project or projects they wish to crunch.

There are many details to work out, but a handicap project does not mean that WU will run out. It will constantly be crunching. What @personthingman2 is describing holds potential as a protocol. Hard to say right now if it will work, but it's definitely worth exploring as we build protocols.