while True learn() Review

Releasing after nearly 9 months of Early Access, the oddly named while True learn() provides players a look into the subject of machine learning. Guiding the player through the history of A.I. development with a mix of simulation and puzzle gameplay, wTl puts you in the place of a programmer as he learns about and creates ever more complex algorithms.

while True learn() Review


Originally released through Early Access back in March of 2018, while True learn()  finally had its full release earlier this week (Jan 17, 2019). The big question about the game is how it has improved in the time between the original preview I did for the game and the full version. The answer is, well, a bit of both

while True learn() is available on Steam for $12.99


One of the main changes from since May has been the addition of a story to the game. In this case, while the player still takes control of an upcoming machine learning programmer, this time around it turns out that his cat is actually much better at it than him. Starting the game with the furry guy fixing a program for you, the rest of the game seems based on the idea of the player trying to find a way to communicate with your pet to pick its brain for ideas.

The way the story is presented is a bit odd. After completing a certain amount of missions you get a cutscene, represented by a group of still images. Depending on your performance in the stages leading up to it, you'll get to chose between one of 3 scenarios. For the most part, they are the same but with some minor differences, like a slight change in a single panel. The only major differences that come are with the ending images, and even then the opening half of them is virtually the same.

while True learn() Review: Apparently AVGN programs on the side
Another interesting addition to the narrative of the game is the online conversations the player character will have with other programmers after making progress. These will usually be used to introduce new gameplay concepts, such as perceptrons or ARMA nodes, and acts more to educate the player on the topic. It's nothing amazing, but it is a nice touch to the overall narrative.

Outside of that most of what you see that qualifies as the story comes through the various work emails, you'll get throughout the game. They will range from the serious to the ridiculous, my personal favorite being those that are totally from humans and not from cats. Definitely human. Not cat. Though sadly some of the funnier things become less common as you take on bigger jobs.


Gameplay wise the meat and potatoes of the game is still pretty much the same. You build programs with various nodes, trying to g for the most efficient setup possible. As you go on you use the money you earn to upgrade your computer and gain new nodes. The main changes to the gameplay are that of the games structure as well as some additional factors to the game.

Probably the biggest change is that instead of just randomly getting emails, the player goes along a tree of progress. As such the player can either just go straight through the center and complete the main game, or go off on different branches to earn more money and deal with ever harder puzzles. It's not anything revolutionary, but it makes the game feel a bit more organized than it did when you just randomly received emails.

while True learn() Review: Progress tree makes the game feel more organized
Another addition in comparison to the previous version of the game is that now, instead of falt completion time for puzzles, the player will be graded via medals based on their performance, being the standard gold, silver, and bronze. The faster you complete the simulation and the fewer nodes you use to do so, the better medal you will get. The main purpose for getting higher grades is, as mentioned in the story section, to get different cutscenes and endings. One thing I noticed more this time around was that the difficulty curve for the game was a bit all over the place, at least for all the mandatory jobs.

Sometimes you'll have to really rack your brain over things and build ever more complex nodes, and other times you'll be doing what is essentially a slightly more advanced than usual tutorial. The most notable of these are near the end, where you could pretty much breeze through the final jobs with a gold medal. Even the side missions seem to suffer from this a bit.

while True learn() Review: Not pretty, but it works
Two big problems I had before with the game were how investing in startup companies and buying things other than computer upgrades worked. Sadly, this hasn't really been improved. Investments are still unpredictable, and in more cases than not will cost you money, even if you make a functioning program. Cat and office purchases are still just cosmetic, making them a waste of money more than anything else.

Another issue I had was with the new puzzle type introduced to the game. In this case, you program a self-driving car. The main difference between this and the other simulations is that this one allows you to test drive it yourself and record the commands. The problem is that it seems almost uncontrollable when it comes to making it drive itself competently. At several points, the car would just smash the brake pedal until it randomly decided to work the way it was meant to. It could just be that I'm dumb and missed something critical, but from my experiences, it seems like a crapshoot as to whether the program will work or not.

while True learn() Review: Driving sections are at best confusing
Overall the game is more organized compared to its previous releases but doesn't really change a lot from those versions. This is a bit of a double-edged sword. While a lot of the good things that were present are still there, so are the negatives. The new additions are also a bit of a mixed bag. The medal system and progress tree make the game feel a lot more organized and complete, but the driving sections are frustrating and, in my opinion, poorly executed.

graphics and sound

Graphically the game is very much the same stylistically as it was before, going for simple designs with mostly flat colors. The addition of cutscene images is nice, though they seem to clash with some of the other art a bit. It does its job, but won't be anything you write home about.

while True learn() Review: Very simple art style
Probably the biggest issue I had with the game before was the sound design or lack thereof. Sadly, this is one of the things that was left the same. All you get in the way of sound is your cat occasionally mowing, fiddling with some of your purchased cosmetics, and a looped music track that, while relaxing, is the only music in the game. This is definitely a game you would play with your own music or a podcast playing in the background.


Overall, not a lot has been changed since the initial release of while True learn(), and what is new is a bit of a mixed bag. On the bright side, what was there previously was far from bad, and can make for an interesting experience. On the downside, this means that a lot of potential the game may have had seemed to go unfulfilled. In its current state, I would say that unless you are really interested in the subject matter, wait till wTl is on sale to pick it up.

 Pros Cons
+Feels more organized than the previous versions-Driving section is incredibly frustrating
+Medal system adds a bit more challenge to gameplay-Erratic difficulty curve
+Machine learning concept is interesting-Very lacking in sound design


Do you like the review?

0 0
Notify of
Inline Feedbacks
View all comments
Would love your thoughts, please comment.x