The general sentiment of bootstrap, get traction, get paying customers etc... is generally accepted by the vast majority of serious entrepreneurs, but I think the idea that moving forward you will be able to get to an MVP with no budget will become very difficult. This is largely because the projected technologies of the future will require much more up-front capability to even prove useful to a potential user than ever.

Doing anything useful or reliable with machine learning, computer vision, data processing, image processing, robotics, genetics or nanotechnology takes years of practice engineering and studying. Those are the frontiers 10 years from now and to think you could bootstrap a competitor to SIRI or QR reader etc... without already having a significant background in it is in my mind far fetched.

I would love to see this proven wrong with the creation of an API/framework/toolset that allows people with <5yrs in engineering to create such products.