Literally the first place my brain went is "Would it make sense to buy Kwedit from a business in return for cash money." For example, if Kwedit is generally ~5% collectible within 60 days, and you're a startup who is not yet ramen profitable who has rent to pay but Kwedit to burn, I might make a very speculative investment and give you US$1,000 for KW$33,000, expecting that to result in future revenues of US$1,500+. Essentially it is factoring for B2C receivables.
(And, ahem, such a stupefyingly bad deal for you that it is probably illegal in some jurisdictions. Proving once again that charging money is an excellent business model...)
You can then imagine a business selling the Kwedit of customers who they believe are not likely to be good for it (subprime Kwedit?) while keeping the Kwedit of customers they believe are likely to actually pay, which would roughly mirror the activities of a collection agency.
This is, of course, all a flight of fancy: I think the idea is likely to be massively unworkable in practice or, in the alternative, regulated out of existence.
As the article points out, and I'll highlight here, right now it's Kwedit in exchange for virtual property. That's unlikely to crash the economy no matter how large it gets, if we don't start extracting second-order derivatives from it. If you can start taking out mortgages by putting up Kwedit as your collateral or somehow start involving real assets, then yeah I think this is a recipe for more trouble.
Oh, and connecting a story from a few weeks back where the spelled-with-a-C credit agencies want to start using social networking to gather credit rating info, Kwedit is credit rating info on a silver platter. How long will it take for failing to pay your Kwedit to show up on your credit report?
It's an odd thing where it only mostly works if they don't extend it at all, but the pressure to extend will be enormous.
Think of all the FarmVille foreclosures.