For iOS, probably last year (2015). Assuming Apple doesn't massively lose consumer marketshare (certainly a possibility in hardware, just ask Motorola), indie development on macOS/iOS/watchOS/tvOS will likely become viable again between 2018-2020.
This is an interesting time in computing history, though, because there are multiple overlapping tech cycles and multiple big companies driving them. During the web interregnum, the dominant tech company was Microsoft, and Microsoft basically completely controlled the channel to the customer. So we just weren't getting fundamental innovations in the browser - there was XmlHttpRequest and hidden iframes that sparked a lot of Web 2.0, and there was Flash from Adobe, but that was largely it. Today, we have innovation in cloud computing from Amazon and Google; we have innovation in programming languages from Mozilla, Jetbrains, and Apple; we have innovation in big-data tools from the Apache foundation and others; we have innovation in machine-learning from Google and a number of startups; and we have innovation in computing devices from Apple, Google, Pebble, Oculus, and a large number of other companies. So while the Apple ecosystem itself may face a dot-com bust, there could easily be a killer app brewing in technologies outside of the pure iOS ecosystem that uses an iOS app to deliver its service to the customer.
This is an interesting time in computing history, though, because there are multiple overlapping tech cycles and multiple big companies driving them. During the web interregnum, the dominant tech company was Microsoft, and Microsoft basically completely controlled the channel to the customer. So we just weren't getting fundamental innovations in the browser - there was XmlHttpRequest and hidden iframes that sparked a lot of Web 2.0, and there was Flash from Adobe, but that was largely it. Today, we have innovation in cloud computing from Amazon and Google; we have innovation in programming languages from Mozilla, Jetbrains, and Apple; we have innovation in big-data tools from the Apache foundation and others; we have innovation in machine-learning from Google and a number of startups; and we have innovation in computing devices from Apple, Google, Pebble, Oculus, and a large number of other companies. So while the Apple ecosystem itself may face a dot-com bust, there could easily be a killer app brewing in technologies outside of the pure iOS ecosystem that uses an iOS app to deliver its service to the customer.