I started reading Passionate Programmer yesterday, and it talks about this. The idea is to either pick a relatively new technology that you can bet your career

I started reading Passionate Programmer yesterday, and it talks about this. The idea is to either pick a relatively new technology that you can bet your career at and to pick a dying tech that might survive for a while (like COBOL). Picking the first gives you immediate coolness points, but learning the latter guarantees you won't accidentally become obsolete. Another categorisation (especially for languages) is to pick a paradigm that is different from your day job. Picking a functional language is the most common example.