Educational content in the era of AI

When I first learned to program I used a platform called treehouse that I can still vividly recall to this day, how their content felt fast, interactive and refreshing. The animations were beatifully made, the explanations clear and you could choose different kinds of paths.

I chose a path of 'fullstack' developer in python and django, a language and framework I have never used since that time but really kickstarted my journey into professional programming.

But, it got me started and also helped me survive my master dissertation where I had to do a lot of statistical analysis. It was a long process of me going back and forth between actively learning through treehouse and "I should start learning to program again". This was all done next to my studies in business and finance that I eventually left behind. My point is, going from beginner to professional developer took a lot of time and determination.

In the era of vibe coding, that apparently means very different things to different people (more on that in another post), why would you struggle through long-form courses when you can just ask claude to do it for you?

I think there are three elements that will drive people to kick up a long-form course now:

1. unknowns unknowns or you don't know what you don't know
2. a competitive advantage
3. taste

To be honest, they could short of be grouped into one but I think they are distinct enough to separate.

LLMs being statistical models means that we are dealing averages. They do an impressive job at coding but someone being new to programming and never having had to debug some random issue until 2 am, it can be hard to intuatively "know" that some code suggestion is good or bad. Unknown unknowns. Like the company that allowed women to shit-talk about bad dates and required a driver's license to signup. That developer vibe coded the fuck out of that app and didn't know that you are supposed to secure your s3 buckets (I can' remember specifically what he did, but it was something like that).

Now, clearly, that is quite obvious to most developers. But to someone who haven't been through the ringer it might not be.

So much of what we do as humans builds on top of what has come before. Like Carl Sagan's famous quote that "we're standing on the shoulders of giants". Learning from people with many years of experience is such a cheatcode and competitive advantage. Sure, it will take long to really internalize certain topics and properly understand them but at least the ideas and concepts are made available to you. Going about discovering things from first principles using generative AI feels like a loosing strategy.

Finally, I think in this day and age, what is starting to truely matter, is having good taste. This was always the case but you could get by without it if you compensated with technical savy. Now, if a proficient developer that has access to AI and has good taste? Your vibe coded app is completely cooked. Learning from someone with good taste gives you a place to start from and further develop your own.

Good artists copy, great artists steal.