Getting the product started
The one thing that your article, along with most others that I see, does not address is how projects of commercial interest will get started.
Historically the OSS answer has been that a project gets started when a single person is sufficiently interested to build a working prototype. However a significant portion of business problems are not reliably going to fit into this model.
The standard commercial answer has been that the large potential profits of a commercial product makes venture capital worthwhile. However OSS removes the economic structure that provides these incentives.
Which leaves a gap.
The model that I think will fill the gap is that consultants will be asked to make custom OSS solutions for companies. The company is hoping to get bug-fixes and extra features for of a product that they need to pay for in any case. The consultant is doing what they would be doing anyways, and getting a little publicity while they are at it.
Which boils down to the observation that since OSS reduces the economic reward to the inventor and improves it for the consumer, it makes sense for new OSS products to be motivated by the consumer.