I'll field this one. The short answer is "yes".
The retail-release development model isn't great for us or for the community. It's not like we stop work for "three years" and then suddenly do a few weeks work and release the results. Instead, we're constantly listening to what the community needs, developing new features, and then.. delaying their release until the next product. We can't simply release everything as a free patch, or we'd never make any money, but simply "holding on to what we've built", though necessary from a sales perspective, isn't something that we like to do as fans of the product.
With the subscription model, we are able to release features far more frequently. I won't go into specifics (that's Tony's prerogative) and I'm sure that part of this will come down to feedback from the community, but we're looking at smaller but much-more-frequent releases. Offering a hybrid subscription/retail model allows everyone to make their mind up about whether they prefer the old release model or the new one.
Of course, some features take much longer to develop than others. Support for parallax heightmaps took quite a while to get right, for example. In some cases, that may mean that we offer a technology preview to our users for early experimentation and feedback, rather than jumping straight to a finished, fully-polished release. For other features, it really is a case of spending a month developing and testing the feature and then it's ready for release.
I should probably add that comparing the cost of the Gold subscription straight across to the cost of the retail product is quite misleading; the Gold subscription includes a lot of other benefits which have to be purchased separately under the retail model. It's up to the individual user to decide whether those benefits are unnecessary inclusions, or a whether they amount to a massive increase in value.
chris