Software continues to eat the world. There are so many new sectors and the demand for software is insatiable. Mobile apps, drones, robots, ai, self driving, medical devices, blockchain... so much more... there are so many fields and applications for software still...
And new sectors... an example: If I were starting in college today. I would major in cs and genetic engineering. Programming living cells...
For context: Andreessen, why software is eating the world. http://online.wsj.com/article/SB1000142405311190348090457651...
There will always be work for people who are good at writing software.
As to your last comment? Maybe but I doubt it. I think that things will morph a bit to the point where there will be better "bootcamps", maybe not called that exactly, that offer more of the real CS education that's necessary for a good software engineer. But we'll see.
They really just learn to tell the app what to do. They tell it to open a camera, save the photo, what format to save it in, submit it it which server, how to store it, when to resize, a balance between quality and size, how to inform another user when the photo has been submitted, how to politely inform the user that it's not going to abuse camera permissions, how to gather user data while staying within legal restrictions, and so on.
A good developer with the best tools might do this in half an hour. A newbie developer could take half a week. The good developer can easily demand more money.
Sure there are things that will simplified to a $25 action. When I started programming, we were building our own push notification services and chat servers. A lot of this is now $25 or less, but a developer can still get paid $1000 to assemble it - the code is there but you still need to pay someone to read documentation.
And maybe through some miracle, AI can do all this work in the future. Then you still need people to manage AI, to communicate with clients/customers, understand what they want, and tell the AI to build it. AI will likely be as smart as some cheap foreign labor who can barely speak English, so someone needs to slowly explain to them what they mean, possibly in their own language... which is a lot like programming.
So let's fast forward 1000 years, where we can reach 90% project estimation accuracy and all this assembly is just blue collar work. We'll get something very similar to the construction industry now, which still hires engineers with a degree, to build complex megaprojects like getting a skyscraper done in a couple years. Maybe a CS certification and the word "engineer" might actually mean something then.
Technology is eating the world, but it’s not a high security field at the same time. Whatever you do today is legacy is 5-10 years, and you can be easily rendered redundant if you’re not in the right place/skillset.