Thinking about it, and if we want in the next few years for the software industry to re-discover
* Coding by humans
* With well-resourced teams on schedules designed to deliver a quality product
* With requirement-gathering, product-management, and architecture design stages explicitly on the schedule
…then we absolutely *have* to figure out some way to sell it as requiring large, powerful GPUs. As the last five years show, a new technological fad can only succeed if it costs a lot of $ in GPU
Perhaps we can somehow find a way to do traditionally effective software development practices in VR, and thus sell it to the industry as a new technique?