3 Counter Intuitive Principles I use to teach myself difficult ideas [Storytime Saturdays]
Insights from teaching myself Machine Learning, Software Engineering, Writing and more.
Hey, it’s your favorite cult leader here 🐱👤
On Saturdays, I will cover stories/writeups covering various people’s experiences 📚📚. These stories will help you learn from the mistakes and successes of others. These stories will cover various topics like Leadership, Productivity, and Personal/Professional Development. Use these to zoom ahead on your goals 🚀🚀
To get access to all my articles and support my crippling chocolate milk addiction, consider subscribing if you haven’t already!
p.s. you can learn more about the paid plan here.
Something that surprises a lot of my readers is that my knowledge is largely self-taught. Even though I break down AI Research online- I don’t even have a Master’s degree in the topic (fun fact- I was never able to take a single Machine Learning class in my undergrad). I don’t have any certificates or even a Bootcamp to my name. 90%+ of my knowledge about the field is completely self-taught.
This is not isolated to just AI. I didn’t take any courses on writing (and most teachers disliked my writing style in school). Most of what I know about software engineering, programming, and tech is what I picked up from reading, YouTube lectures/debates, and conversations with other people.
In this process, I’ve gotten reasonably good at teaching myself difficult topics and ideas. Now there are a lot of resources online that teach you how to learn. However, through my journey into self-education, there are 2 core principles about learning that I don’t hear discussed much. In this article, I will be sharing these two principles, why they have been extremely beneficial to me, and how you might leverage them.
PS- This article is an example of the kind of content that is available to premium subscribers of Tech Made Simple. To ensure you get complete access to the evergrowing library of the world’s most informative chocolate milk cult, consider getting a premium subscription below.
Two Principles that have helped me with self-education
The best way to learn something is to learn it- This might sound really dumb, but hear me out. I often speak to people who are interested in learning about “AI for _”. When I ask them what concrete steps they’ve taken to learn more about this topic, it’s typically, “nothing”. These people have been sitting around- waiting to find the best way to approach this learning, without actually learning anything. This kind of “give me a syllabus” mentality is a hold-over from our formal education system. Ultimately, if you want to learn about a topic, just start learning about it. As you immerse into the topic- you will naturally come to learn about what is important. Trust in yourself and take the plunge into topics. I’m not discounting the value of preparation, I’m just saying that most people undervalue the value of action.
Consistency » Intensity for learning- A common mistake people make is to front-load their learning plans. Whether it’s Leetcode, AI, or any other skill- they decide to spend multiple hours a day for months at a time feeding information into their brains. Not only is this approach incredibly taxing (causing most people to burn out), but it is also not ideal if you seek to immerse yourself in a topic and develop any real mastery of it. Ultimately, people who make this mistake generally treat skill as a static object- once you reach a certain level, you have attained mastery and you no longer need to do anymore. Your skills are a sliding scale, that will continue to evolve as you grow (here we’re considering both atrophy and improvement as evolution). Easy/moderate effort over a much longer time frame will enable you to develop more connections with the topic you want to learn and enable a greater degree of fluency.
We will now go over these principles in more detail, including examples of how they have played out in my life. As always, take whatever I write here with a critical lens. Take what you consider useful and discard what you don’t.
If you want to learn about something, learn it
A bias towards action and willingness to take a plunge will you help you learn as well as any learning plan and syllabus. As long as you start investing in an idea, iterate upon your learnings, and stick to the process over a reasonably long time frame, you will eventually learn the topic with a fairly high degree of fluency.
“There is more wisdom in your body than in your deepest philosophy.”
-Thus Spoke Zarathustra (Nietzche)
Above is one of my favorite quotes. It encapsulates the principle of this section perfectly. Our brains are very malleable and adaptable. Sitting around, waiting for someone to tell you what you should read and what ideas are important deprives you of the ability to use your brain’s capabilities of finding what is novel and useful. If you want to learn about a topic- don’t wait to enroll in a course or some other external event to start. Pick up your phone/laptop and start searching away. You will be surprised how quickly you will start finding work that is useful to you.
I’m not asking you to ignore guidance or not take information from other people. A lot of my learning has been through inputs and suggestions from other people. However, waiting around and waiting for the perfect game plan will not help anyone. Instead, use the following process-
Start looking into a few ideas.
Make associations/connections and focus on understanding the idea to a degree that is immediately useful. In the beginning, your understanding of the core principles will be fairly high-level.
Continue to iterate on your process.
Your iterations will keep taking you deeper and deeper until eventually, you are interacting with complex topics at a fairly high level of resolution. This was how I taught myself Machine Learning and AI. I stumbled into the field at 17. At that time, I didn’t know much beyond the basics of programming, some level of calculus, and fairly basic probs and stats. In the beginning, I understood almost nothing about the field. By constantly interacting with the field, I eventually pattern-matched my way through the field enough till I was comfortable enough to start breaking down research papers online.
Once again, I’m not discounting the value of courses, learning plans, etc. Those guided resources can be a great augmentation to learn something in a focused way. However, every pro has a con. Their structure also makes them very narrow. If you wish to understand a topic/field, you will eventually need to make unstructured learning the main goal. Learning is ultimately a deeply personal activity, and beyond a beginner level- you will be your own best guide. Let others be great resources, but your results will ultimately depend on your iterations and the action you take- not by sitting around and not taking the wheel.
Some of you might be apprehensive about this. Following such an approach would ultimately lead you to have a lot of holes in your knowledge. Your knowledge will be a patchwork of different ideas and concepts- papered together by your ability to connect them. You won’t really be an expert on the topic. Unlike a structured, tested approach, there is no way for you to know if what you’re studying is useful.
To this I have the following response- You can never be an expert on any topic. No matter how much you learn, there will always be more. Even courses (and entire degrees) ignore many ideas to focus on what they consider important. The belief that any certification, accomplishment, or level of knowledge is the end is erroneous. Contexts are always changing, and as they change- so will the ideas that are useful and what’s in. When it comes to learning, there is no end.
My AI writing (and knowledge of Machine Learning) has received a lot of praise from many experts and luminaries, including highly accomplished AI Researchers, Developers, and even a Nobel Laureate. However, even now I make a lot of mistakes (some very basic). In a recent paper breakdown on Sparsity that I did on AI Made Simple- I made a very basic mistake- I confused exponential growth and quadratic growth. The reason was simple when I started learning about Neural Networks someone told me that neural network growth grows exponentially with parameter size, and I never really questioned it beyond that. In general, I knew n^2 to (n+1)^2 growth is quadratic, but I never put that together with Neural Networks because of this gap in my knowledge. You can catch part of the discussion below (thank you to
for correcting me)-Even now, I make a lot of mistakes and have to learn a lot about the field. And that will continue to be the case as long as I’m involved in the AI. There is no perfection, and there’s always a lot more to learn. The only way to do this long-term is to let the waves take you to where you want. If an idea is important enough- you will learn it eventually. As long as you’re moving towards a goal and taking deliberate actions, you will end up where you need to. Ultimately, there is no “the answer”, there’s only “your answer”. And by constantly following other people for what you should focus on, you will deprive yourself of the opportunity to find your answer.
Now let’s move on to the next core idea.
Focus on Consistency and not Intensity
One of the most interesting lessons I learned about training and athletics was that you don’t actually improve between reps. You improve between sleep. The logic was simple- sleep/rest is when your brain forms the neural pathways that will eventually help you learn a task. Thus, when training for a sport- sleep and rest is as important as the training, because they will determine how well you truly learn the craft (PSA- I’m not a neuroscientist or doctor, just a meathead who knows some bro-science).
A similar principle applies to learning. The more our neural pathways fire, the better we get at the task. By engaging with concepts in a less intense, more long-term manner, we can establish more connections between what we learn without overworking our minds. Too many people try to do “5+ hours of Leetcode every day for 6 months”. This is an incredibly short-sighted way to learn something. Intense study periods work for cramming all the studying before your test, but not for cultivating an understanding of the field. Where you can- take lots of breaks, integrate moderate/light effort towards your goals, and give yourself a longer time-frame. It will do your learning a lot of good.
The best part of this about this is that it won’t take you a long time to catch up with the core ideas. As you continue to learn, you will spot some overlapping ideas and concepts, and this will allow you to accelerate your learnings. It took me roughly an year of reading 1 paper in 3-4 weeks before I was able to break down research papers online (this was in late 2020). At that point, I was still only able to fit one paper per week into my schedule. Presently, I go over about 10 papers/writeups weekly, even though I have a lot less free time now than I did then. The progression of my skills, comfort with technical content, and ability to extract information have all been grown organically- without too much effort. At no point was I mentally overwhelmed, losing sleep, or going out of my way to fit things into my schedule. The greatest challenge with the whole process was staying consistent with it, but it was mostly manageable.
Favoring consistency will allow you to stick to your plans long-term while also giving you a much deeper appreciation for the ideas you come across. The longer time frame will allow the ideas simmer in your mind for a greater time, allowing your brains to build a more diverse of connections related t the topic. The approach of favoring consistency over intensity has been extremely useful in my learning journey when it comes to fitting my learning into an overcrowded space. Consider looking at ways you can build your learning schedules to accomodate for this.
Loved the post? Hate it? Want to talk to me about your crypto portfolio? Get my predictions on the UCL or MMA PPVs? Want an in-depth analysis of the best chocolate milk brands? Reach out to me by replying to this email, in the comments, or using the links below.
Reach out to me
Use the links below to check out my other content, learn more about tutoring, reach out to me about projects, or just to say hi.
Small Snippets about Tech, AI and Machine Learning over here
AI Newsletter- https://artificialintelligencemadesimple.substack.com/
My grandma's favorite Tech Newsletter- https://codinginterviewsmadesimple.substack.com/
Read my articles on Medium: https://rb.gy/zn1aiu
My YouTube: https://rb.gy/88iwdd
Reach out to me on LinkedIn. Let’s connect: https://rb.gy/m5ok2y
My Instagram: https://rb.gy/gmvuy9
My Twitter: https://twitter.com/Machine01776819
Consistency is so huge.
People ask me how to get better at BJJ. I tell them: make it into a boring routine, so you automatically do it.
Nobody wants to hear this! It's not sexy or exciting.
But it's how you get good at things.
Also, yes, very much so on diving in! I share this entrepreneurial spirit, and if I didn't, I would not be running multiple businesses today and writing every day here on Substack.
Good piece, Dev.