Live data from Hacker News

Ask HN: What do you want to learn in 2017?

news.ycombinator.com

271–280 of 513 posts

Re: Ask HN: What do you want to learn in 2017?

#271

I'd really like to pick up a functional language, but I'm faced with the paradox of choice. I'm presently torn between: 1. Haskell: all the cool kids are doing it, and I like its formalism 2. Hy: I write a lot of python, and really like the idea of working with familiar data structures.

If you're interested in web development (and perhaps even if you're presently not), you may want to take a look at Elm. It's a functional language that compiles to JavaScript and is almost entirely developed for the purpose of building web components and applications. The way the functional features and other language features are explained as directly benefiting the developer experience is pretty cool.

https://www.gitbook.com/book/evancz/an-introduction-to-elm/d...

Re: Ask HN: What do you want to learn in 2017?

#272
I have plans to relearn undergraduate mathematics (upto Functional and Complex Analysis) and learn the basics of computer science (just Algorithms, Data-Structures and Discrete math) in the first half of 2017. Being able to do this successfully would mean a lot to me.

Finnish and Hindi are the two languages I wish to learn in the next year.

Re: Ask HN: What do you want to learn in 2017?

#274
* become an expert in PostgreSQL

* improve my knowledge of Erlang

* catch up with ES6 (not really useful, but, just in case)

* learn how to scale my side projects from 5-10 initial users to 100-500 (some seo, marketing, sales)

* how to be efficient with my time

Re: Ask HN: What do you want to learn in 2017?

#275
Math-wise, well, it's a matter of where I go to college, I figure it'll have a huge effect on basically everything that happens thereafter. I definitely want to continue on my path towards understanding the Weil conjectures and that kind of thing, so I plan to be able to compute sheaf cohomology for schemes by the end of the year and learn more algebraic number theory. (And maybe look at representation theory.)

CS-wise: learn enough ClojureScript to be able to make convincing mathematical (etc.) visualizations -- I recently got a taste of Figwheel+Reagent and I'm hooked! -- and be able to manipulate Haskell transformer stacks better.

Other than that, I'd like to learn some French (marathoning Engrenages was the best decision ever!) and not forget the little German I know.

Re: Ask HN: What do you want to learn in 2017?

#276

I have been reading alot about Machine Learning and I want to get into the practical application of it. So I will begin with learning Mathematics and then some Machine learning code for training a basic model for NLP or Facial Recognition :) Any suggestions on how to go about learning Mathematics requires for Machine Learning is more than welcome

Assuming you know the basics of calculus: learn lots of linear algebra, lots of probability theory, mathematical statistics, and optimization. But you don't have to wait to learn all the math to get going

He wants to get into the practical application of machine learning, not machine learning theory.

This is a common mistake people new to the field make. You can be very successful by learning how to use machine learning frameworks, and that doesn't require lots of probability theory, mathematical statistics, and optimization. Not that it hurts.

vayarajesh, start using Tensorflow, you'll reach the ability to reason about problems to which machine learning can be applied, and how to apply it, much more quickly than starting by starting at the root of the tree of knowledge. You can always learn as much math as you want in order to dig as deeply as you want, but first get a sense of what you're dealing with.

Re: Ask HN: What do you want to learn in 2017?

#277
post #195

Earlier quoted context omitted.

Couple of quick pointers wrt Japanese. First, human languages take a long time to get good at. If we think that children learn faster than adults (which I think is not actually true, but it's a widely held belief), then it will take you 5 years to talk like a 5 year old, 10 years to talk like a 10 year old and 15-20 years to talk like an adult -- minimum. If you study very effectively, I think you can double this spe…

> If we think that children learn faster than adults (which I think is not actually true, but it's a widely held belief), Anecdotally, I saw my two children learn a foreign language in exactly the same place and over the same time as me. They learnt it to fluency in the same time I could barely communicate. Actually google 'scholarly articles on how children learn languages'. It's probably not just anecdotal.

I'm not suggesting that adults will always learn faster than children. You will no doubt notice that you did not spend your time the same way that your children did. Did you spend 24 hours a day thinking and struggling in that foreign language? There is a big difference to being exposed to language and concentrating on it. Restricting your self to thinking only the thoughts that your language enables you to think is difficult for an adult (well, impossible really).

I'm not entirely unaware of the literature in the area. I spent 5 years of my life teaching English as a foreign language. A large percentage of that time was spent doing research in language acquisition. As far as I am aware, ever since Chomsky the general consensus has been that there is no different mechanism for language acquisition in children than in adults. There are still some researchers who disagree, but that happens in every field and is healthy.

But, you aren't going to beat a child in learning a language unless you do some very specific things. That's because they spend 24 hours a day, 7 days a week learning language. Even when they get to school, their school work is primarily surrounding language activities until they get to be about 10 or so.

An adult's secret weapon is the ability to read. If you use that to direct your learning, then you can go quite quickly. Like I said, though, getting to an adult level of proficiency in less than 10 years is pretty difficult, but it's not impossible by any stretch of the imagination. The same level proficiency for a 10 year old child is very, very unusual.

For things like accent and getting tones correct all the time, that's a bit of a different story. Only about 2% of adult learners achieve native level proficiency in that kind of thing. But it's not something that will hold you back from normal communication. Event then, I'm aware of some specific language training you can do to help, but most people are not really interested in going that far.

Re: Ask HN: What do you want to learn in 2017?

#279
- A different hello world every 3 days - Write better - Think outside of the box - Embrace exposure - Docker - How to make a VR app - Understand inverse kinematics - C - Crystal - API Gateway microservice - Concurrency patterns - Understand FPGAs, and an HDL - Applied machine learning
Post reply on HN