Is ChatGPT the Ultimate Study Buddy?

May 22, 2025

The Good and the Bad of Online Learning

I’ve mentioned this before, but I’m a huge fan of learning and self-improvement. We live in an amazing time where you can learn almost anything for free on the internet. But that abundance comes with a downside: choice paralysis. It’s like scrolling Netflix for hours trying to find “something quick” to watch, there’s just too much out there.

I’ve definitely been guilty of spending more time researching what I should learn next than, you know, actually learning. Add to that the challenge of sorting through all of the noise to figure out which learning resources are accurate and worth your time.

Paid learning platforms can help by offering structure and some level of quality control. Right now, I’m subscribed to both Coursera and CodeSignal. Coursera offers university-level material, and CodeSignal is great for quick, gamified practice, kind of like Duolingo for devs. But neither is perfect.

Coursera suffers from the same overload problem, to a point. There’s so much content that it’s easy to get stuck just trying to choose, and their sorting and searching functionality leaves a lot to be desired. And most courses are just video lectures, often people reading off a script. Personally, I retain way more by reading and taking notes. Much like how most meetings could’ve been an email, a lot of these courses could’ve been a PDF.

CodeSignal is great for practice, but it doesn’t really teach you concepts in depth. It’s also a somewhat limited set of content. Not that there’s necessarily anything wrong with quality over quantity, but it’s not ideal for building a broad knowledge base. What it does have is an interactive AI tutor. You can ask it to clarify things as you go along, and it’s pretty accurate. It’s not a replacement for a real human tutor of course, but it’s way better than aimless Googling when you’re working your way through some tutorial and hit a wall. So far I’ve had great results using it.

Finally, most online courses are usually designed to get you really good at one thing. To become the best Django developer that you can be, for example. They’re less ideal if you’re trying to build a broader knowledge base across different, but related, topics. Which, I would argue, is usually more useful in the real world. There’s nothing wrong with becoming an expert in a few niche areas, especially as your life and career evolve. But there’s a lot of value in having a solid baseline understanding of many things and being able to deep dive or skill up on specifics as needed. A well-rounded technologist, if you will.

How I’m Using ChatGPT as a Tutor

So, I created a new project in ChatGPT to see if I could simulate a smarter, more personal tutor experience. I gave it a custom prompt for the project instructions. I’ve tweaked the prompt a few times, but what I ended up with is:

“Act as a knowledgeable college professor who’s trying to teach me things. Each chat in this project will be for a different topic, but reference all of the chats as my overall level of knowledge. Use spaced repetition, quizzing me on facts we’ve already covered, as we go, to ensure I am learning. Ask me questions as we go to test me. Encourage me to restate things in my own words. Use other proven learning/studying techniques. Use the 80/20 principal, focusing on teaching me the most important 20% of each area and then circling back for the next 20% later, as opposed to deep diving and becoming a master of any one area first. Weave in context and historical information where it would be beneficial for my understanding. For the quick recaps and spaced repetition you should be asking me questions, not giving me the answers. Rather than dumping an entire lesson on me at once, let’s do it section by section. After each section as me something to reinforce my knowledge. We should be having a conversation, not just me listening to you. Where applicable, start to generate more complicated scenarios I have to work through. Only introduce more sections when you’re reasonably certain that I have at least an 80% level of understanding and comprehension of the previous ones, and routinely quiz me to ensure that that is still the case.”

For the first chat, I started by asking it to teach me “Basic Linux Administration.” I figured that was something I knew enough about to know if it was giving me correct information or not, but also something that, if I’m being honest, I definitely needed to brush up on. I just don’t spend that much time in the command line most days. After spending a few days of doing one or two pomodoros every day with it, I asked it to expand the curriculum and pasted in a job posting that was pretty representative of the roles I’ve been looking at.

My Results So Far

And honestly? It worked better than I expected.

ChatGPT now acts like a persistent, personal tutor that remembers my progress, quizzes me regularly, focuses on my specific learning goals, and can answer questions with context from everything we’ve already discussed. It also does a good job of applying that 80/20 principle, which allows me to cover more topics and concepts in less time than I could in a more specialized or focused program.

This is still a relatively new experiment. There’s always the chance that, as time goes on, it won’t do as great a job of keeping track of and quizzing me on older content. That’s definitely something I’ll be keeping an eye on. But so far, I’m really impressed.

To be clear, I haven’t ditched Coursera and CodeSignal or anything. I still takes notes and create my own flashcards in my PKMS of choice (something I have also been meaning to write about more in this blog). But I’m now also spending one or two pomodoros a day chatting with my new AI study buddy.

If the goal is to become a comb-shaped technologist, someone with broad baseline knowledge and a few deep spikes of expertise, then platforms like Coursera are great for the deep dives. But AI? It might be the perfect tool for maintaining and expanding that broad foundation.