I'm thinking I should learn a ML framework/platform (Tensorflow, Azure, ???).

Which should I choose? The only thing I can think of doing with it tho is to build a tic-tac-toe or maybe a gomoku AI...

Learn keras which is a deep learning framework. Keras uses tensor flow as backend, therefore keras's api is more simplified than tensor flow. All the best.

@mohammed so Keras would be in Python? So I guess of all though, what language would be best?

Usually Java is the go-to pick but not sure for ML.

Also what sort of hardware would you need? Hadoop was all about clusters/server farms, same with Alpha Go but not sure about simple applications. Do I just use local (slow? I have an i7 5yr+ old laptop or pick a cloud service, not sure if they have a free tier).

@billgates lol don't do ML in Java

@mjones44 y? U gotta I've some valid reasons

@billgates cause it's slow and HELLA boilerplate

@mjones44 slow compared to? From my own usage Java seems to be faster than Python.

@billgates I'm not sure about compared to python, but C/C++, Go, etc. i.e. anything that compiles directly to machine code

but the thing I hate most about Java is the boilerplate

@billgates python is the language of AI, and most of the ai frameworks are in python. For the hardware, you can use Google colab for free if you feel your laptop is too slow.

@mjones44 guess I should learn Go. C++ tried but just cant handle all the the deferencing, pointers, constructor, destructor stuff...

What about Kotlin... somehow i always think the 2 are related...

@mjones44 you get used to it... and write a lot of Utii libraries assuming theres no lib for it already.

@billgates Kotlin runs on the JVM, it's basically Java that doesn't suck syntactically, but it runs about the same speed. I don't personally have experience with Go but some close friends are Go devs and I've heard GREAT things about the language.

@billgates also I know I've written HELLA Java lol

tldr: stay in the cloud

@mjones44 If you want the best power, I'd go with C++ but if you want a lot of training that means you'd need a VM server to run it. Many cloud providers like Azure have premade solutions especially optimized for ML. If I were you, I'd look at those. Although I'm working quite a lot with Azure I haven't tried their ML branch yet and I also don't know how much you can do with a free account.

Tensorflow with Google colab( if your hardware has toaster speed. )

Start with sckit learn on local or use ML Studio