WITHOUT corporation

I beta-tested this game, but I never got very far. I did complete it this time! (Possibly through exploits).

This is a difficult game to play. It combines in one several features at the start that make it extraordinarily hard to get use to.

1. The online version has no way to scroll back. The game now suggests using a transcript which is good. I downloaded it and made 6 transcripts which I'll upload to the hint thread.
2. Compounding the above, the game frequently prints several pages of text at once and clears the screen.
3. You do not control the main character of the game; instead, they perform random inputs while you observe and help.
4. You cannot WAIT for the main character to do things, since the game will yell at you and disable WAIT. So you have to come up with different actions each turn.
5. The game can randomly crash; I had 4 different crashes on Gargoyle, so saving often is useful.
6. The game still progresses if you save, and at the beginning there are several help menus to navigate that are required to progress that also move the game forward, and
7. The game yells at you if you are too slow and calls you stupid.
8. The game is a surreal game-within-a-game with multiple voices and POVs, and the game-within-the-game is intentionally buggy and underdescribed.

So, the average experience of a new player may very well be trying out different commands, missing chunks of text, having the game yell at you, and being booted out to try again.

If they are able to push past that and continue, they can find some amazing concepts.

The idea of the game is that Max Fog is dead and an AI trained on his consciousness has finished a game based on his notes. Years later, you and an in-game AI have to work together to make the game run for a real-life player. So, you move objects around, queue up routines, etc. in order to get the game running. Meanwhile, the in-game AI is annoyed with how buggy the game is.

More happens, but I'd rather not spoil it. It has much in common with the ending of A Mind Forever Voyaging.

I'm in the game! A neat trick is that whenever the game crashes/reboots you read a review by Me, Mike Russo, or JJMcC, with various scores depending on how well you did running it. I laughed at mine because it was extremely similar to what I'd write, down to a typo in the first paragraph.

The game talks a lot about AI, and is set in a future where AI has caused lasting damage to the earth.

I did find the ending hopeful.

The game is very ambitious, with using Inform menus and keystrokes and regular text to progress, random commands, writing to external files (I think; I never triggered the text that I think does that), and text animations which I've rarely if every seen in Gargoyle. It doesn't 100% pull it off, since it crashed a few times, but it 90% pulls it off, which is great. I did decompile it, which is funny as the fake me in the fake review also decompiled the game (good thinking, future fake me).

I was expecting this game to be harder than it actually was based on Mathbrush's review, so I was pleasantly surprised to find it was very stable, didn't crash when it wasn't supposed to, and in general behaved as I expected it would, with some nice surprises along the way. I played on Gargoyle, not online, and took copious notes along the way, which helped.

There was a popular indie game going around some time ago called Dr. Langeskov, The Tiger, and The Terribly Cursed Emerald. In Dr. Langeskov, you come in expecting to play an ordinary 3D video game, but are instead roped into the behind-the-scenes action and forced to help set everything up, making sure all the props are in the right place, and all the light switches and weather dials have been properly tuned. It's a betrayal of the promised concept, but what it delivers is much more interesting. WITHOUT Corporation is just like that too; instead of playing the game yourself, you're roped into the behind-the-scenes work of manipulating game objects and routines so they run when they're supposed to, all so the actual player can have a good experience. You even get to see the actual player's review when you're done, and the score they assign to your game.

The game combines several ideas I've seen in other places in a stylish way. Dr. Langeskov is one, but there are others. One key difference between Dr. Langeskov and WITHOUT Corporation is that WITHOUT Corporation's fictional game-within-a-game, Climbing Up the Walls, is made with the assistance of generative AI, and you're tasked with fixing things up behind the scenes so the bugginess of the result isn't apparent. An echo of how real-life generative AI systems are backed up by humans who step in to intervene when things go haywire, such as with Waymo driverless cars.

I've been seeing more and more sci-fi stories about generative AI lately. Not stories written by generative AI, but stories by humans about its impact and the potential future it could hold for us. Unsurprisingly, most of these stories are bleak, and this one has some bleakness in it, though with a surprisingly hopeful ending.

I've always had a thing for AI and stories about AI even before real AI started spreading like kudzu, and the AI boom has led to even more stories about its various real and potential impacts. I'm thinking of Everything in Their Hands, which I played and really liked the first act of, and other short stories about AI like Turing Test (warning: contains slurs) or Emotional Intelligence Amplification. All of these stories have a common theme of making the reader question how human they are, if they might also be playing the role of a machine. Which everyone is, in a way, if they have to go to school and work a job and so on. People work themselves into a specific shape to get approval from the algorithms that determine success or failure, slotting themselves into the machine, pass the blunt etc.

As far as 'playing the role of a machine': though I've never played Type Help, I have played (without finishing) Dunnet, which is a very programmer-ish parser game where you have to use actual terminal commands to win, necessitating a basic understanding of Bash among other things. This game is similar, very technical. You're not playing the game but behind the scenes of the game, setting things up for the actual player (who in a twist of irony, is of course a construct of code in our real world, where 'WITHOUT Corporation' is just Max Fog's 2026 IFComp entry). The game is very metafictional in that sense. It claims to use generative AI, but all the actual 'generative AI' is just human writing and action, specifically you and your sometimes-acerbic AI assistant. In fact, I'm familiar enough with generative AI and its writing style that I found the story's insistence that it was produced by generative AI to be very funny, given none of the telltale signs are there. Though I also don't like AI writing, so we all win here.

I think that part of the reason I had a surprisingly easy time with this game is that I spend a lot of time with various command lines, thinking about the outputs of various commands, so moving to a quasi-fictional command line and having to interpret those commands wasn't a big deal for me. I liked the general setup, though I think the backstory, for both WITHOUT Corporation and Climbing Up the Walls, was a little vague. Probably the AI backstory could've been tightened up by having more focus on a vastly reduced cast of characters, and (Spoiler - click to show)more detail on the personal life of the employee whose diary we get access to - as a form of humanization and characterization - instead of just talk about the AI project. Similarly, I think the story for Climbing Up the Walls was kind of vague, and despite the postive 'reviews' the game got on our fictional Intfiction, I felt like the ending was really abrupt and didn't answer any of my burning questions, which made me wonder if I was missing something or if there was a second part of the game that I never got access to. I did upgrade using the ZILF update.wit -A command, but I'm not actually sure what that accomplished. I thought an upgraded interpreter might mean the game would crash less, but maybe it was just a necessary milestone for progressing the meta-story that had no effect on Climbing Up the Walls itself.

Having to go through the same setup for Climbing Up the Walls did get tiring, and I think I figured out the right sequence of actions within the first one or two playthroughs, so if there was a faster way to skip the intro, that would've been appreciated. I played Climbing Up the Walls maybe three or four times in total to get to the true ending, I think four times; (Spoiler - click to show)I got kicked out the 2nd time by the AI for failing to move Felipe, and once the player did the UNDO thing at the end that made the screen glitch out, but my first and last runs went surprisingly smooth. I also really didn't like that there was no 'wait' command for you to wait for the player to do things. I ended up using 'x sky' instead, which doesn't print much text, but.

I thought the metafictional aspects and AI aspects were cool, but since I'd seen them before, I wish there was more bite to them, something that went above and beyond what those other stories did that I'd already read. On reflection, I think another part of it is that a big issue with modern AIs is that they really don't have interiority, they really aren't conscious or sentient, they just act like they do and in that sense 'hook' humans into relating to and depending on them. An emotional hyperstimulus. But despite this game using a lot of the language around modern generative AI and LLMs, it still plays with more old-school scifi tropes like (Spoiler - click to show)'accurate digital recreation of a human mind' and 'AIs with genuine emotions, feelings and interiority', something I believe is firmly out of the reach of modern LLMs - despite the proclamations of many to the contrary - and furthermore something I believe will always be out of reach of modern LLMs, as long as there's no drastic change in the technology. Mixing those oldschool tropes with modern LLM terms uncritically, without ever delving into or playing with the possibility that the AIs featuring in this story are totally nonsentient, nonconscious philosophical zombies, bothers me since it implies that modern LLMs, too, could be conscious or sapient on some level, which is an idea that annoys me. End digression.

If you're interested in AI or metafiction and don't mind screwing around with some heavy programmer-type command line stuff, playing this will be a fun two hours.

Walkthrough for Climbing Up the Walls, from what I figured out:

And the ending command, which took me around three failed attempts to figure out because I didn't realize what I was supposed to do and ended up just brute-forcing the (Spoiler - click to show)ZILF command with every combination of target files and letter flags until I found it: (Spoiler - click to show)ZILF MFog.zip -C

Also, the password for each secret file is (Spoiler - click to show)the first three letters under ROT13, and then the month and day it was edited. Though I could never figure out how to get to the .docx file, maybe I just typed in the password wrong for that one?

I think Josh Grams got to a lot of this stuff before me, but writing it down anyway for people who check IFDB in the future.

It also occurred to me that the description of each Climbing Up the Walls command could be useful, so here they are: