Unless you are unable to do something and need a new technology to solve a problem, but then, probably be best to stick with things that have a track record of just working.
If it looks cool do some side R&D with it, but until you see expected results I'd take any new technology with a grain of salt - especially if it's proprietary.
I got out of the JS ratrace but if I had to I would read the helicopter view of the top JS frameworks and then talk to people at a meetup about facts on the ground of using those tools. For a small amount of work you can gain enough knowledge to know when to delve deeper and when to not bother.
I read http://webplatformdaily.org/ daily, and I subscribe to a few newsletters (PonyFoo, HTML5 weekly, JS Weekly). I read the changelogs for new versions of browsers. And, occasionally, I browse all the pages linked from chrome://chrome-urls. I've learnt about all sorts of old-but-new-to-me browser technologies doing that.
