My Thoughts

Posted: 24 February 2016

I haven’t really been posting much recently as I was really busy over the Lunar New Year. I’ve been thinking about quite a few things recently: Web Programming, Deep Learning, and Entrepreneurship.

Let’s start with the first one. I just had dinner with a very good friend from mountaineering. He’s also really good (in my opinion) at Ruby on Rails. We really need to revamp the mountaineering website as the club is quite inactive now. We started it and shouldn’t let it die so easily. The first thing we should do would be to have a really nice site with user sign ups, so people can get information and perhaps communicate there. There could even be a “profile page” where everyone gets to see what mountains other people have climbed. I thought that would be interesting. In my very first post when I started this github.io page, I claimed that every CS student needs to know how to program a webpage. Well, I officially do not hold that stance as of today. Web Programming in itself is a complicated field. Setting up the various things on Rails today was tedious. Becoming good at rails requires A LOT of practice. You need to be familiar with the entire system to be really efficient. I’ll be happy with a github.io page from now on. Looks like I’ll need to rely on my friend for most of the heavy lifting of the website. I’ll just do the basic HTML and CSS.

I attended a course recently, and I’m starting to realise how hyped up Deep Learning is right now. It is insane. I’m not a professional at it; I just know how to use it, but the common phrases I currently hear would be “Deep Learning is not proven enough, but it works”. With many nice Deep Learning Frameworks like TensorFlow, Theano, Torch, NVIDIA DIGITS, etc, almost everyone can do some form of Deep Learning. I mean, it’s easy right? Just throw in lots of data, run it for a few days, and you have an awesome classifier. My friend showed me this post, and I think it pretty much summarises it. This quote was particularly pertinent.

This “WhatsApp effect” is now happening in AI. Software tools like Theano and TensorFlow, combined with cloud data centers for training, and inexpensive GPUs for deployment, allow small teams of engineers to build state-of-the-art AI systems.

I’m really interested in Deep Learning as I believe it’s a powerful tool that is worth exploring. I’m kinda in a dilemma now.

Should I just be very good at implementing these tools and visualise the results beautifully? Well, I would be a “Data Scientist”.

Or should I know how to implement these tools, and be really hardcore at mathematics? Well, I would be a “Machine Learning Researcher” in this case.

This brings me to my next point.

I am currently reading this post about Elon Musk. Well, he’s really legendary. I must say he is in his own class. Thomas Edison is akin to “inventor”, Einstein is akin to “genius”, Elon Musk is akin to “I don’t know what they’re going to call him in future, but I’m sure legendary/godly/Tony Stark fits.” This guy is really out of this world. If I could be a fraction of what he is, I would be really happy. I really want to make products that matter in this world. If I truly want to do that, financial freedom is necessary before I can focus on the things that matter. This brings me to the following questions.

Should I work extremely hard on a start up, hope that it gets acquired, get a few million, and then use that to spin off the next business? I could read and understand many things in the course of doing a start up.

Or should I work on getting strong fundamentals in Deep Learning (or some other field if my interests change), get some experience, and do a start up at 35 (and hope that it acquired)?

If I choose the former of doing a start up now, then I think what I could do would be to provide such Data Analytics services to clients, but, the market is really saturated with this. I’m not sure if the start up would even work.

If I choose the latter, then I would focus on becoming very sound in mathematics and becoming an extreme expert in Deep Learning. I could do the start up at 35 (with a higher chance of success…?).

Hmmm… As I was writing this, I thought to myself, why 35? I could technically start earlier. Well, these are my uncollected thoughts. I’ll just put this up here and re-visit it again. I really need to make a decision soon and focus on it.