1. Web devs will start owning performance on mobile
2017 will be the year where 20 seconds time-to-interactive on mid-range phones stops being cool. If you don't think that's an issue, maybe stop testing for mobile with a MacBook Pro and a small browser window.
Good idea that will not happen in 2017: Google search results showing how many minutes a page load will suck out of your battery.
2. IoT hype will continue to endanger the public
Pushed by consultants, every device that does not need to bundle Wifi and a web server will do so in 2017. Hence we're going to see a lot more botnets of unpatched toasters and lightbulbs.
I predict that there will be at least one large-scale DDoS attack caused by unmaintained household devices.
Good idea that will not happen in 2017: Regulations enforcing a best before date for networked devices, until which manufacturers are required to produce security patches.
3. Native apps must stop competing with the web (or die)
With two thirds of users having access to Progressive Web App features like store-less installation and offline support, the sweet spot for native apps must move or disappear.
If you're working on native apps right now, make sure to build something that cannot be done on the web: Large offline media, background audio, low-friction payments.
I predict that in 2017, several high-profile media properties will abandon their native apps in favor of PWAs, lowering maintenance costs and increasing engagement.