
You're a programmer. Your cat is a better one. Solve visual puzzles built from real machine learning ideas, from decision trees to neural networks, to create a cat-to-human translator. Take freelance jobs, upgrade your rig, gamble on startups, dress up the cat. No coding required, curiosity is.
Modeled estimate, not a reported figure. How we estimate →
Prices
Each in the currency that store quoted — not converted: the gap between two rows is the publisher's pricing, not an exchange rate.
Discounts
Every price cut of more than 20% we have observed, in each market that quoted one, with this game's chart position on either side of it. These are episodes, not a model: a sale is one event, and one event cannot tell you what the next one will do.
Reviews, worldwide (the counter carries no country): 1/day in the week to Sep 10 → 5/day in the week after.
For scale, across Steam · Top Sellers · Brazil over the last 30 days: the 26 chart moves that followed a price cut of more than 20% went a median of 7 places towards #1, and the 2,400 moves with no price cut behind them went 2 away. One store, one chart, one country — it is the only place we have enough observations to say even that.
Prices outside the United States have been collected since 6 August 2026 and chart positions since 19 July 2026, so a cut older than that has no market to compare and a market that joined later has no history yet. Nothing here is converted between currencies.
About
Add-ons
Each DLC is estimated with the same model as a game, then summed — a modeled figure, not reported DLC sales.
Sums Steam + Google Play only; PlayStation, GOG, Xbox, Epic, Nintendo publish no volume signal, so the true total is higher by an unknown amount.
| Store | Price | Audience | Score | Chart | Est. revenue |
|---|---|---|---|---|---|
SteamYou are here | $7.79−40% | 8.2Kreviews | 91% | — | $1.7M |
| $4.99 | 1.2Kratings | 96% | — | $135.5K | |
Updates
Build ids and update times come straight from Steam's PICS — the same source SteamDB reads. Patch history grows as we record new builds.
Engagement
Reviews
Themes across 843 recent reviews and how positive each mention is. A keyword signal, not full sentiment analysis.
Reviews
编程风味连线游戏 现实中越难的知识点,游戏中就越简单,因为作者也不知道怎么表现出来,干脆一笔带过了。 整个游戏玩下来,唯一和机器学习沾边的是每一章节开头留的知识点超链接,点击后他会通过外部浏览器打开维基或者油管页面。 难度很低,即使半夜把你摇醒了让你去玩你都能通关。 如果你是抱着学习的心态来买的,可以离开了。 如果你是抱着消磨时间同时烧烧脑的目的来的,可以去玩玩"人力资源部机器"和"70亿人类"。 如果你是抱着放松放松心情,同时想了解点机器学习发展史,那可以买个玩玩。 不过他画面挺好看的,史低搞一个来玩玩不亏。
Потрясающая игра. Прям лучшая. Совокупляется с мозгом на 10 из 10. Что я поняла? - Многие стартапы - лажа полная. - Круги, квадраты и треугольники - фигуры от дьявола - Программисты - святые люди.
Media

Related
Related
Price dynamics, follower growth, revenue and sales trajectories with interactive tooltips
This and every other Pro feature is free until October 20, 2026. A free account is all it takes.
Every market we collect.
Genres
Stats
All-time low is the lowest price we've recorded for this game since we started tracking it.
Languages: English, Russian, German, Hungarian, Korean, Greek, Portuguese - Brazil, Simplified Chinese, Polish, French, Italian, Spanish - Spain +20 more
| $14.99 |
| 63ratings |
| 80% |
| — |
| not modelled |
| $12.99 | 34ratings | 62% | — | not modelled |
| $14.99 | 26ratings | 94% | — | not modelled |
| $12.99 | — | 90% | — | not modelled |
| $12.99 | — | — | — | not modelled |
Matched by title — a reused name can pair the wrong games. * is a critic aggregate. Prices and positions are each store's United States storefront; Epic charts worldwide. Audience and Score too, except Steam's, which is worldwide.
Hours played at review time, across 844 reviews with recorded playtime.
positive critical· bar length = how often the theme comes up
Gameplay-wise, WT:L is a watered-down Zachtronics-like, with nearly every level being some variation of the same "sort these inputs into those outputs" flowchart puzzle. The training process for the "ML-driven" tools you're given for these puzzles is distilled to a tedious "let it run for a minute to fill the error-rate progress bar before you run it again to see if the solution is valid." Once that's done, these "advanced" tools fail to distinguish themselves as being meaningfully better than the "basic" non-ML algorithms used early in the game -- every one of the ML-driven tools sacrifices accuracy in favor of throughput and low flowchart-box-count, and the game is only able to push you to use them by moving the goalposts to match. Once the game has tutorialized "fast but sloppy is better :)" a few times, the non-ML options are taken away entirely any time the intended solution to a puzzle uses one of the ML boxes. Very little "real" machine learning is actually done outside of one minigame where you "train" a genetic algorithm to run a self-driving car, and even that turns into a repetitive RNG-fest -- for the first task the game makes a big deal about the need for the player to provide a good set of training data (driving the car yourself for a minute), but all four of the other car-training tasks take that option away. The player is left watching helplessly as the car fumbles around rear-ending other vehicles nonstop, praying that the algorithm's mutation settings will luck into a setting that passes muster before it runs into the totally arbitrary epoch cap. Tonally, while True: learn() is a relic of the past, from the pre-COVID times when the general public had an overall positive view of machine learning. That sentence alone should clue you in where I'm going with this ... it hasn't aged well. While the historical context the game provides is interesting, the rose-tinted "ML is the answer to everything" message comes off as tone-deaf in the current era, where the corporate world has thrown away billions of dollars pushing generative AI to no one's benefit except their own (and the net detriment of many). In keeping with that corporate mindset, the "invest in startups" side hustle is pointedly and egregiously cynical when viewed alongside the rest of the game. You provide both code and capital, under the explicit assumption that these ventures will be briefly profitable and then crash-and-burn into nothingness after a few weeks or months (and indeed, that arc is hard-coded into every one of them). There is no interest in building anything meaningful or long-lasting, only in pumping up your investments and riding the hype until the time comes to cash out, all so you can afford ... $30,000 cosplay costumes for your cat? Yech.
i like programming games. i like puzzle games. this game seemed like it would have some appeal to it. after chugging through some annoying stages that required "non-optimal" sorting favoring throughput over accuracy... i'm a bit disillusioned with what it's teaching. i kept waiting for more complexity and logical progression. instead, it feels like i'm just beating my head against a wall to find "the most average result that [insert unknown metric for "a good score"] until i finally happen upon a solution that is "best" (high subjective, in my opinion).... the unlocked upgrades are not clearly defined on "when" you get them, and the startups just feel clunky and overly punishing. you can't redo them if you "fail", unless you start a totally new save. not my ideal way of playing... repeating the same thing over and over. i'm really dissatisfied with this game, even after really giving it a good try. DON'T BUY
Завтра иду работать в сбер благодаря этой игре, говорят надо вопросы решать какие-то, тихо мирно по телефону или подъехать и балансировщиком порешать. 10/10 бабок отдающих свои деньги мошейнникам.
[h1] DO NOT WASTE YOUR TIME ON THIS IF YOU REALLY WANT TO LEARN [/h1] If you are considering paying money for this, let's see which is your case. Case 1: Kids(6-12) who want to learn about coding basics or enhance your logic. My suggestion: Scratch is free and better. Case 2: Teenagers(13-17) who are enthusiastic about CS and ML. My suggestion: Watch some free videos on YouTube, solve math problems and do some actual coding is more helpful. Case 3: Students who are major in CS and want to learn, practice while playing. Which is me. My suggestion: Go to Kaggle because the game also ask you to do so. There're much more fun than sorting circles, squares and triangles here. Case 4: You want to play games without being accused "wasting time". My suggestion: Playing games is not wasting time. Play the game you like if that won't influence your life and career. It will turn out to be wasting time if you pretend you are learning from this game. Case 5: You are a chill person, having a bunch of time to kill and you wan't some basic concept about ML without getting deep into it. If money isn't the problem, this game is for you. [h1] What's wrong with this game [/h1] [h2] Lack of Knowledge [/h2] Take expert systems as example. The game will tell you that once upon a time, there was a method called expert system, which didn't work well. And all you need to know to finish the game is that it can distinguish colors. Want to know more? Click the link that will lead you to Wikipedia and a YouTube video! Which does not help you solve the puzzle better. So why must I pay for this game and spend time using an outdated algorithm to sort colors and finally I still have to learn from Wikipedia and YouTube? [h2] Lack of Effective Challenge and Fun [/h2] The process of solving the puzzle does not make your brain active but simply consumes you time and energy. Players win the golden medal for good sixth sense instead of logical thinking. The challenges of the game focus on the meaningless and repeating adjusting process, instead of finding a valuable solution. [h2] Weakly Related Background Information [/h2] Our main goal in the story is to understand our intelligent cat. It could be a very interesting story if I actually learn how to recognize my cat with graphic data and understand it in some way, even in a sci-fi way. But in the game, I don't see in what way I am understanding my cat. Each level is a request from some company which has a background and it will be cool if I actually learn how to solve these problems. But in the game, I don't see the relationship between the puzzle and the background information. Actually, you won't miss anything by just skipping the background information, not reading them at all. I can't say these stories are completely unrelated to the puzzle and the knowledge. But in this game, they are not helping. [h1] CONCLUSION [/h1] It's not the right place for either learning or playing. I bought it 35% off and I still don't think it worth the price. If you are having fun playing this game, that's nice. If you are expecting to learn from this game, I strongly recommend you to think twice.
我从清华五道口金融学院毕业之后,就躺平成为了一个骗钱...啊不,算命的,顺便干着奶爸和厨师的行当。但当我玩了这个游戏一段时间之后,我在今天拿到了字节的offer。 很开心,所以写了这个评论,强烈推荐
和机器学习最大的关系是简介,没有任何的硬核内容,完全不如《一个人工智能的诞生》,这个游戏就是给小学生的连连看,这个游戏更应该和大数据挖掘分类有关,而不是人工智能
Pretty boring. Gameplay: You're given a random prompt in each level and are supposed to sort colors and shapes into boxes with different modules. The game expands on it a bit but essentially mostly the same dreck for the first 3 hours. The upgrades barely make any difference early game and so are basically useless, just like the decorations you can spend money on, but don't get to see if you're actually playing the game. Not my kind of puzzle game at all. Also, the "education" part is basically slapping a "This is how it is in game, this is how it is IRL" but doesn't really go into much detail, so as to make that information a trivia piece, as interesting as it is.
It's alright.. it's more of a logic puzzle game than a programming game, all you do is connect a few components in a way that makes sense, but this game is way way too easy, and instead of getting harder it seems to get easier, the last epoch only has a few levels and it feels like the developers ran out of ideas, and the solutions are a straight line.. what even was the point of that? You can easily brute force most levels, and the only difficulty in this game comes from the inputs having different amounts of shapes/colors so you need to divide them among the outputs. Even the reinforcement learning levels with the car just boils down to changing a few variables, you aren't programming anything, you aren't even solving a puzzle at that point, it's just brute force. This game has a currency that you can spend on customization and some new nodes and upgrades. The upgrades are basically mandatory if you want to gold all missions but at least it incentivizes you doing well to earn more money. The best levels were the startups where you could make your own soluton to satisfy your clients and make money, and patch it as the demand grows, escaling it to supply it all.. at least in theory, if this wasn't so buggy and escaling actually makes you lose money, there's a lot of complaints about this, I wish it was done better. All this game has going for it are the pretty graphics. You aren't going to learn anything about machine learning, it's not even an engaging puzzle game.. it's a really weak recommend if you want to waste some time, there's way better puzzle games out there.
Active players
Estimated daily and monthly active players (DAU/MAU) modeled from peak concurrency.
Players by region
Where this game's players are, estimated from review languages.
Players also reviewed
The games this title's players most often review too — audience affinity.
Related
| Game | Estimated lifetime revenue | Reviews | Review score | Price |
|---|---|---|---|---|
| — | 34 | 62% | $12.99 | |
| — | 0 | — | — | |
| $109.2K | 158 | 93% | $12.59 | |
| $199.5K | 321 | 92% | $14.99 | |
| $113.4K | 188 | 76% | $14.99 | |
| $23.9K | 31 | 100% | $19.99 |
Related
| Game | Estimated lifetime revenue | Reviews | Review score | Price |
|---|---|---|---|---|
| $1.7M | 3.5K | 93% | $19.99 | |
| $1.7M | 5.6K | 42% | $17.99 | |
| $1.7M | 3.2K | 81% | $29.99 | |
| $1.7M | 1.6K | 85% | $23.99 | |
| $1.7M | 2.2K | 79% | $19.99 | |
| $1.7M | 3.7K | 86% | $19.99 |