💾 Archived View for pennywhether.xyz › articles › while_true_learn.gmi captured on 2023-04-26 at 13:10:54. Gemini links have been rewritten to link to archived content

-=-=-=-=-=-=-

While True: Learn ()

Generally, I’m pretty terrible at puzzle games outside of things like Tetris and Puzzle Bobble. Stuff companies like Zachtronics puts out tends to kick my butt pretty bad. Nevertheless, I took a bit of an interest While True: Learn () (I’m just going to refer to it as WTL from here on for brevity). It’s largely a game that is intended to teach players about machine learning, but even stripping that aspect away it functions very similar to what I’ve experienced in those Zachtronics games.

What’s probably making me want to stick with this game is the fact that it feels like I’m learning something practical here. Usually I don’t make it far in puzzle games before calling it quits, yet here I am persevering. I’ve had minimal experience in programming over the course of my life. Some simple HTML, playing around in Basic as a kid, and making this capsule is about as far as things have gone. The likelihood of me taking up a CS-related career is remote. Besides, by the time I’d be adequately trained to do such a thing, it would be time for me to retire anyway.

That hasn’t stopped me from taking an interest in this game. Machine learning is a term that has been appearing more and more, especially as the public is constantly reminded that automation is coming for their job sooner or later. So, to some extent I’ve made more effort than I normally would to soak in what WTL has to offer. It seems rather poignant.

So far, it has largely involved essentially having a very large bin of stuff on the left side of the screen, then dumping the items in a variety of bins on the right side, making sure that the items have been sorted by various parameters defined at the start of each level. It sounds simple enough at first, but when multiple parameters come into play things get complicated. This ranges from sorting items by multiple properties, including shape and color to making sure everything operates within certain thresholds (speed, accuracy, volume of items sorted, power consumption, etc.). Currently, I’m close to halfway through, and have quite a bit on my plate right now.

There isn’t any actual coding that I’m aware of in the game, as everything is presented visually (generally colors and shapes). WTL is more about getting the general idea across of how different aspects of machine learning work from what I’ve seen so far. It doesn’t seem to really want to focus too much on the nitty gritty. That being said, it provides a ton of links to learning material every step of the way for people who do want to explore more on a given topic.

How this is all presented in terms of plot is that there is a novice programmer who keeps running into problems with his projects. Whenever he leaves the room, though, his cat hops up to the keyboard and corrects his code. So, now the guy wants to find a way to make a cat to human translator, and this is where the player comes in. The game is a gradual learning process with more and more concepts being added, allowing for increasingly sophisticated forms of automation. Levels branch out on a tree between main story quests and optional tasks, with the eventual goal of being able to talk with the cat.

Each stage takes the form of small businesses emailing requests for the player to complete programming jobs. One’s performance is ranked, so the better someone does in making the program work, the greater their monetary reward. Also, certain parts of the story can only be unlocked by doing well.

As a nice touch, the money players earn can be used for a variety of things. These range from knickknacks to decorate their desk with to better parts for their computer which in turn improve all the nodes used in projects. While improved stats are nice, I’ve enjoyed collecting the decorations more. I’m busy saving up for a model cat dressed in Warhammer 40K space marine armor at the moment.

So yeah, WTL has been surprisingly interesting so far. It’s definitely tough, at least for me, but I’ve been enjoying how it is trying to teach me a bit about machine learning. I’ll never pursue a role in that field, but hopefully this game will at least help me to develop a better appreciation for it.

Here's a link to the game's Steam page for anyone who might be interested

- Pennywhether

pennywhether@posteo.net

March 15, 2021