Because the approach that Shenzhen uses for hardware is rarely the same approach that they use with software, and their software approach is generally not something to be emulated.
We may get a lot of wacky gizmos out of China, but the vast majority of them run unmodified or craptastically modified Android/Linux/vendor BSP.
The software model that's closest to what Bunnie describes in Shenzhen is the plain old open-source model. To get good quality software, you need to put significant time and resources into these open-source projects, and let people modify or fork them as they see fit.
Really, the goals are just different. The goal of the Shenzhen model is diversity - "there are so many products out there, what should I make that's a little bit different?" The goal of most software is arguably quality and reliability. There's no shortage of software with silly features, and it's usually easy to add a silly feature to an open-source project if you really want something specific. This is often the kind of software modification I associate with the Shenzhen model.
What's difficult is forking a project over an architectural disagreement and then fixing it, because you have to understand a lot more about how the software works before you can do that, and whatever documentation exists is likely to be in English.