The invisible mountains

Nikita
5 min readJul 3, 2017

--

This is old article, check my wiki to get latest info on Learn Anything, the tool I am building to solve the problem outlined in this article.

I wanted to write about this topic after reading this wonderful article on Research Debt. It’s a really wonderful look on what it is like to do research amidst all that is happening every day in the world.

I have never done any professional research myself and I have a vague idea of what real research is aside from what I have read about it online and the few questions I asked to people doing real research.

But one thing is clear. The big problem with research is really exploring this unknown territory that no one explored before. You are often in the dark and you don’t really know where to go and what the best path for going anywhere is.

Learning, unlike research is easy. Learning new things means learning material that some other human before you understood. What more, he or she probably already wrote a book, an article or made a video on the topic you are trying to learn. The only problem is that the surplus of knowledge on the internet is immense. It’s a shame that so much time is being wasted on finding what the best path for learning a topic is when the best path should already exist in some form.

This is the problem we are trying to solve with Learn Anything. By visualizing all these learning paths in the most clear and engaging way possible.

We also do it in a visual way with user curated topics. Each of these topics is connected in a meaningful way in an hierarchy of dependencies. So for example linear algebra is in mathematics. TensorFlow library is in machine learning libraries and so on. This allows users to explore this graph of links and topics instead of staring at a black box search bar that Google and most other search engines provide.

Everything is also subject to change as all the topic’s content is open source and thus anyone can submit feedback and improvements to the structure and content of these topics.

There is currently ~2000 topics all fully connected and ready to be improved with the help of community.

Another really awesome thing with this approach of crowdfunding resources in this way is that anyone can potentially add his or her knowledge to this graph. If someone wrote a really awesome article on a topic, the current best way to promote the topic is to post it on Reddit, share it with friends and then hope people will notice it. After few weeks though, what that person submitted gets forgotten in this black hole of information that Internet is. The chances of your article surfacing on Google are very slim if you are only starting out and have no following.

With Learn Anything however, we don’t care how popular an article or a course or the video is. What we value is quality. If the content is great, it will be added to the website where it will potentially be seen by thousands of people. You don’t have to do any SEO optimizations, you don’t have to waste time trying to think of how you can best optimize your viewership. What is important, is the content that you are trying to deliver and what message or knowledge you are trying to share.

Doing research is like climbing an invisible mountain. Like in the image above, you climbed this far to the point where there is no clear path in front of you any more. There is nothing in sight that you can get a hold of.

For many though, these invisible mountains come all too early. For many, mathematics is a black box of knowledge. What is mathematics really? Linear algebra, Logic, Statistics, Calculus .. I remember I read something about analysis. Oh and category theory? Or learning about machine learning. What is machine learning really?

Wikipedia is a great way to try and get a better visual picture of what some field of knowledge really is but Wikipedia is dense. To get this visual picture, you have to read through many many articles, try and make the mental connections in your head. Try and construct a logical set of dependencies for learning the topic. All this takes time and all of this is wasted time when ideally you should just learn until the point you hit a wall where there is no longer anyone who really knows what you are trying to understand.

With Learn Anything we try to expose these mountains that you have to climb before you get to the really interesting bits. If you want to learn React.js to make an awesome website because you heard somewhere that React.js is awesome. Learn Anything will show you what the best learning path for learning React.js is. In this case, you should best learn JavaScript itself and to best learn JavaScript, it’s really handy to know some computer science, specifically what a program actually is, what computation is and what algorithms and data structures are. For all these topics I mentioned, there exists a topic with best resources that guides you in learning these topics in the most optimal way so you don’t waste time finding how and what to learn and just learn. And then go on building awesome things.

Another thing we try to visualize is ‘missing knowledge’. Some topics are really dense with material like machine learning. Others are not as dense like brain computer interfaces. We thus give a certain sense of material purpose to research. You know what we collectively as humans already know and you try to add to it or clarify what we know in a more meaningful way. Research becomes fun and engaging. Like a fog of war waiting to be uncovered.

At least that’s what our hopes with Learn Anything are. We still have a long way to go and there are many things that will be improved along our way. Another important thing is that the website is open source. If you for some reason dislike the format of our nodes or how we structured the topics and do not wish to help us make it better, you can just fork what we have and create your own knowledge graph. For yourself and for your own research. Extend it in your own way to use.

These kinds of visualizations are extremely powerful as graphic representation of knowledge if done well is extremely empowering. It gives a sense of direction and awareness of what it is you know and what things exist out there that you don’t know.

--

--