I don't think that people's desire for physical stuff is the real problem. The real problem is mass production, which requires mass marketing as a condition for its survival. Get rid of mass production, and you get rid of the need for a lot of marketing, which in turn gets rid of people's artificially inflated desire to consume products.

Current manufacturing processes are much too wasteful: if you want to build ten products, then you have to build a thousand - that's 99% waste, which is literally swept under the rug by convincing people who don't want it to store it in their basements.

Of course, rewinding the clock to the pre-industrial age is not an option (although it's always a danger as you point out). I believe the next step forward is better on-demand production. We're seeing the first baby steps with digital goods, on-demand production of CDs and books, etc.

What we need are more versatile robots, that can build a multiplicity of components from raw material. There are printers that can build 3D objects layer by layer from an engineering drawing - this is the kind of thing that's needed, but with a variety of types of material that can't be done today. It cuts out the need for factories that produce components - and in turn cuts out the need for all those surplus components to be actually stored and transported all over the world and used to produce more useless goods and marketed to people who store them in their basements. It also cuts out the need to mass produce a product to justify lower manufacturing prices.

Ideally, in the post-industrial on-demand society, you simply visit your local manufacturer and choose some products from a catalog, then pick them up a couple of days later after they've been built - all without the baggage of global trade and marketing.

--
$E(X_t|F_s) = X_s,\quad t > s$