Some thoughts on using LLMs for learning
LLM — Large Language Model
I’ve recently come across a blog post (cited below) from Josh Comeau in which he wrote about a strategy to follow when using LLMs like ChatGPT for learning new things, especially programming.
It’s a good strategy, so I decided to re-share it by adding a few of my own thoughts.
I tried learning programming on and off for a few years now. Every time, I start, get somewhere, and abandon the process for some reason. This year though, I had to learn it cos I decided to pursue a Master’s in Computer Science and it’s a strong requirement for me. I love coding, but the learning process can get gruesome at times. One of the hard things is getting an understanding of how things work at a high level. Sure, almost everyone has an idea of what a website is and how it works, but what about the process of building that. That’s what a programmer needs to get a grasp of to build a website or web application. And this is what ChatGPT helped me understand with ease.
I had an idea for a simple web app a few months back, and I wanted to see how far I could get with just ChatGPT. After two days and two nights of work, I built a close enough version of the app I wanted to build and deployed it on to Heroku. And that was the first time I ever built a web app that could actually be used by someone other than me. There were and are still a lot of things that I don’t understand about how the web app works but it helped me get an overview understanding of the process, the frameworks involved, how a model-view-controller framework works, how I can style front-end using TailWind, etc.
I jumped into something without having much clarity over it and I still don’t have complete clarity over it. And that’s okay, cos now I’m learning, I’m learning how things work at a low-level based on my understanding of how things work at a high-level. I’m able to keep up much better with my learning process cos I got an understanding of how things work in general.
Now, pointing back to the reason I wanted to write about this, ChatGPT and other LLMs like that, while they can help you in your learning, they can also make you lazy and point you in a wrong direction. If you truly want to learn something, you have to put in the effort to understand how things work at a low level, there is no other way. ChatGPT is not a quick or easy solution for that process. Josh Comeau provides a good strategy in his article about ‘The End of Front-End Development’ from the point of view of rise in LLMs.
The following is an excerpt taken from his article.
Using LLMs to help you learn
I’ve heard from a few folks who have said that ChatGPT has been really helpful when it comes to learning technical skills. If you’re confused by something in a tutorial, you can ask the AI to explain it to you!
This is a really interesting use case to me. Essentially, ChatGPT is like a pair programmer, someone who can help you make sense of things you don’t understand. You can ask it specific questions and get specific answers.
But I do think you need to be careful. There’s a right way and a wrong way to use tools like this to help you learn.
The wrong way would be to treat it like GPS navigation. When I have to drive somewhere, I pop the address into my GPS, and follow its instructions indiscriminately. I do generally wind up where I need to go, but it requires 0 mental effort on my part. As a result, my sense of direction has totally atrophied. I can’t go anywhere now without a synthesized voice telling me what to do. 😬
Instead of treating it like a GPS, I’d suggest treating it like you’re a member of a jury, and the LLM is the defendant, taking the stand.
You’ll listen to what they have to say, but you won’t accept it as fact. You’ll be skeptical, and think critically about every word.
Instead of blindly copy/pasting the code that ChatGPT generates, go through it line-by-line, and make sure you understand. Ask it for clarification. And double-check things that seem suspicious with an authoritative source (eg. the official documentation). Keep in mind that LLMs are 100% confident, but not 100% accurate.
If you follow this strategy, I think LLMs can offer a lot of value. 😄
