Ask HN: With such fast changes in technology, how do you update your skillset?

That sounds rather blunt, but most organisations that aren't startups don't change technology quickly if at all. C++ has served me well for two decades; I probably ought to adopt C++14 but on the other hand my current job requires that the codebase build with a 2008 compiler.

I'm also extremely skeptical of the extent to which AI and VR are new, as opposed to incremental improvements to technology which takes it over an adoption barrier. Have you seen the 80s VR headsets? SHRDLU? The "AI winter"?

If you're worried about this stuff then it's helpful to develop a level of knowledge about it that's slightly higher than Wired but lower than actual implementation detail, in order to talk about it in interviews. You can then pick this stuff up as you go. Machine learning in particular is maths-heavy, matrix algebra in particular, and that's never going to go obsolete.

I also agree with the commentators who are saying that you should ignore the latest flash-in-the-pan frameworks unless you really have to to get frontend gigs.

Skills vary both in how much the market values them and in their durability. There's often a trade-off between these two characteristics. For example, half a year's worth of study in a foreign language or pure math is only somewhat valuable to the market but that value doesn't tend to decrease over the years. Learning AngularJS in 2013, on the other hand, was so highly valued by the job market that it was a great way for junior programmers with no degree to break into a software engineering career.

I believe it's best to generally focus most learning efforts on durable skills, but occasionally when there's an opening, to flop and focus 100% on an ephemeral skill that's highly valued and appears likely to be even more highly valued in the near future. After capitalizing on the opportunity, return to mostly focusing on durable skills.

I believe that even though technology changes fast, the things you need to change do not move as quickly but that probably depends on the company and technology that company is using in the first place.

Where I am working, we are writing software mostly in Java, some analysis on the data with Splunk and SQL for the database.

Sure enough I had to keep up with Java development but it is not _that_ rapidly. Nevermind the fact that the company only now is switching to Java 8.

That being said, I do like learning new things in the field but they are not the "latest cool things", for example now I am learning Haskell by reading books on it and doing excercises, the normal way to learn a new language afaik.

I do tend to check out things that tickle my interest, lately I have made a small app in Angular2/Dart because it sounded interesting, but by no means have I learned to use them in-depth.

===============

0. Assume any "new" thing is worse than the "old" alternative - until proven otherwise.

1. Critically filter out hype/PR.

2. You're left with much less to learn.

3. Invest "out-of-work" time in something really valuable.