Make something useful!
Sure robotics is where computers were in the 70s. They're big clunky things, expensive, high maintenance, and suitable only for use on simple pre-programmed tasks (and requiring heavy-duty programming to do those tasks).
What made home computers happen? They became small, relatively cheap, relatively powerful, and most importantly versatile enough to provide several applications that people really needed in one box: mainly spreadsheets and word processors. Games were a big part later, but they'd never have happened if there wasn't an established market for home computers already - and when games did become a major player in home computing, it was because they came up with innovative ideas for how to play games (or copied those ideas from arcade games).
So until robots are small, relatively cheap, relatively powerful and versatile, they're never going to have any significant impact. At the moment, they're large, heavy, expensive, weak and unable to do anything useful. And most of what robotics places call "robots" are just funky versions of a remote-control car. (You say you can control that robot's fingers with your fingers? And it's using remote-control servos and a radio link? Just like the £10 kiddy toy I got as a present 25 years ago? Wow dude, that sure is progress!)
The only robot to do anything useful to date is the Roomba, and that's more like the robot equivalent of the electronic calculator than the computer. Still, it's a start, but if the robotics industry wants to go mainstream then it needs more focus on what people actually need and how to achieve it, instead of yet another robot head or remote-control buggy.