[JustSomeNobody]

More than anything I would say stay slightly behind the hype wave.

Take two steps back and one can see the waves of hype roll through programming blogs, HN and reddit. 10 years ago people were arguing over ORMs. Then it was dependancy injection frameworks, then SQL vs NoSQL. Now there's JavaScript and it's various frameworks and libraries. THere's always so much hype about what to use that it's easy to lost in it and worry that if you aren't riding the trend wave, you'll drown. But most times it's easier just to float a while and watch from a short distance. Get done what you need to get done with what you currently know. Let the hype die down. Learning new things is a ton of fun. But they can be devilish to deploy in production. Don't feel like you have to use them.

Edit: I'm not saying DI frameworks, or ORMs or NoSQL (or even SQL) are bad in any way. Just that within each topic there was much arguing about what is "the best" or "proper" or whatever. Wait for that to cool off and die down before betting your next project on it.

[Joeri]

They're not waves, they're cycles, with each successive generation rediscovering the lessons of the previous generation. I found it particularly amusing to see javascript front-end developers struggle with model/view binding strategies a few years back knowing that all of this was already hashed out by desktop developers in the 90's and terminal interface developers in the 70's. Still, react came out of that, and I don't think the desktop ever had an equivalent to react, at least not in the 90's. So every once in a while a genuinely novel idea pops up, but they're rare.

Citing Alan Kay on the difficulties of keeping up with tech:

We are fortunate that most of what is "new" is more like "particular 'news'" rather than actually "new". From the standpoint of actual categorical change, things have been very slow the last 30 years or so.

[JustSomeNobody]

Even react is based on 4 decade old ideas; Component based programming (at least the idea of it) has been around since the 60's.

[Joeri]

I wouldn't say components is the novel idea in react. For me the novel idea is continually generating the entire ui by applying functions to data. Previous component-based frameworks were about static declaratively-described ui's that where then bound to data, which is in many ways an inferior approach.