I recently took a look at the Cryptopals Crypto Challenges. It's a series of challenges through which you can learn all about crypto and crypto attacks. They say they'll eventually have solutions but since the site appears to be at least a few years old, who knows it they'll ever publish them.
One interesting thing about the site is that it really doesn't have a lot of content to teach you the concepts around the challenges. At least not for the first few, the only ones I've looked at. You get a bit of sample input, the sample output, a short description and then you're off to the races.
This means that you're going to do some web searching to figure out what the terms they use mean and how to approach the problems.
The first challenge was to convert a string of hex digits into base64. I'd never done this so I figured I should look up what base64 was. I figured it was some encoding but exactly what was it?
A quick search yielded a bunch of online encoder/decoders and also the Wikipedia page: "…Base64 is a group of binary-to-text encoding schemes…" - a group? Uh oh. Which scheme am I to use? It turned out that I was able to use the conversion table on the Wikipedia page but I digress.
I decided to write the solution in Clojure and was about to do so but decided to see if there was an idiomatic way to do this. I did a search. Most if not all of the searches resulted in links either to a Clojure library to encode and decode Base64 or example code that used said library.
If I wanted to I could code up a complete solution and not really learn anything about base64 encoding. I'm not talking about wholesale copying a solution off of StackOverflow. I'm talking about using the included Clojure (technically Java) Base64 conversion library.
This made me think about what I'm seeing some times in CS Ed and lots of time in drop in / opt in programs.
I'd hear about classes where students are doing unbelievable things. The kids made a neural network to do blah blah blah. I'd think to myself "that's upper level undergraduate work, how can they really understand all the math and CS to build that stuff when they're just taking Algebra and their first CS class?"
It turns out, they weren't. They'd basically glue together libraries and use pre-existing tools.
This raises the question: "is this a good thing?"
Maybe and maybe not.
If those kids mentioned above aren't really studying AI but rather intro programming and they're using AI as part of the motivation, it can be terrific. It could inspire them to study AI later and that's great.
On the other hand, if they think they're really learning AI, that's not so good.
I've met kids who were in either classes or programs on cybersecurity. They were every confident in their knowledge and skills but it turned out that they didn't really learn anything in their classes. It was just a curated canned experience. In these cases the programs and classes weren't really helping the kids. I've met teachers who've told me about classes like these where the kids ended up getting jobs but then promptly got fired because they didn't really know anything more than a superficial view of the subject.
In this day and age students can't write everything from scratch but I've always believed that they should have some idea as to what's going on under the hood. A friend would always say "never use a tool that you couldn't' write yourself" I concur. He didn't mean you have to write every line yourself but you should have a basic idea as to the underpinnings.
There are cool and interesting things at all levels of CS and CS education. I know we live in a day and age where teachers brag about how early they can cover some advanced topic with their kids but it's not a race. Let's teach the kids what they need when they need it. They don't need to think they can write a deep learning platform in year one. Right now I'm teaching college first year students. We don't cover the same things these kids will study in their 400 level electives. I'm fine with that. I cover some topics that allude to what's to come and I set the stage for their futures. Let's design and implement CS experiences that make sense for our kids rather than giving them false experiences so that we can say that they're doing graduate level work in third grade.