"Your code should show that you are able to keep up with the times and you are making an effort to continue learning."
I offer a caveat of especial import to those of use who live and die by the browser-based UI. Keep up, but beware too early adoption of new technologies and standards.
One of the recurring themes in web development is the failure of browser manufacturers to deploy new releases that comply with new standards in an even remotely timely fashion. You MAY be well advised to learn a new standard early and play with it in the sandbox of an early-implementing browser, but take care not to "boldly go where no one has gone before" for deployment to production apps.
I know, there's an ideal of providing the latest technology and failing gracefully backwards to earlier standards in relatively more lame and halt browsers. However, embracing the bleeding edge, when there is only A browser deploying the newest features is ill-conceived. There's too much other stuff that needs doin', e.g., bug fixes, unimplemented TODOs, refactorings, etc. to justify providing the most perfectly new-&-improved feature set when it's only supported in the latest release of ONE browser. You might provide a limited in-house proof-of-concept or a demo for presentation at a convention (or on Hashnode) of the new features but not all-hands-on-deck for all new development.
Be certain the features that are new are truly worthy of the commitment to new learning and corresponding new practices...period. For instance, there are multimedia aspects of HTML5 that resolve long-standing complaints against technologies such as Flash. Without those virtues, I personally would not make the exchange for the formal, inarguable, canonical correctness of XHTML. HTML5 is in some ways a giant step backwards.
Keep up but don't overcommit to deploying new technologies too early. Don't even vaguely appear to be a lightweight distracted by flash-in-the-pan innovations. Adopt new ideas in a timely manner that reflects good judgment.