Ask HN: Should I learn Blockchain (smartcontracts) or AI (in general)?

[esseti]

Hello, I'm late-mid 30, I'm a tech/dev by study and work. Every year I add a side topic in my learning list, 2020 was the year of finance and investment. Right now I've two topics that gained my interest and I would like to explore more, but they seems too vast and I can't really decide what to tackle first.

One is Blockchain topics, most specifically smart-contracts or a way to play and make something out of the blockchain. I already studied the blockchain , I implemented it (more or less) and I've been in the cryptocurrency world for enough time to have missed big opportuiniteis for retirment (late 2016). Another thing that I'm interested is AI, not with a specific goal, I would like to learn how AI works and how can one can create a SW that is able to write miningful text, or summary of text or stuff like that (note, when I went to uni we did ML and a bit of that stuff, still, it was not called AI).

What would you suggest me to do follow? and with what material to start? the smart-contract topic seems more narrow and more easier to target, since it's a very specific topic. the AI i can spend therer years.

[simplecto]

AI is pointless without the underlying skills to acquire and clean data.

If all you ever plan to do is take off-the-shelf data from some repository and teach yourself the mechanics of tensor flow or the others then by all means go ahead.

But if you expect to use AI in novel ways in novel applications truly new ways then you must understand how to acquire and create clean viable data.

That is a real grind for many. Unfortunately that is what many “data engineers“ are doing in practice.

It is a valuable skill but you might even have to get good at Excel.

[esseti]

learning a framework is something in the list, but I was not sure if it make sense to learn tensor flow without having a bigger picture. I don't even know what's the "best" framework and why

[ftruzzi]

"AI" means all and nothing. I'd say just learn what you like and go with the flow.

Maybe you should think a bit more about why you like those topics (and start from a subset of them once you have a good answer, ideally the subset you like the most).

I don't mean to be rude but if your only goal is to add a side topic to your learning list I'd say either skip it or pick a random one as they're both technically interesting.

[esseti]

Thanks, make sense. Just to clear it out, as side topic I mean something that is not linked with my day-to-day job right now.