When Uriel arrives, the robots initially deny his existence, or they refuse to believe that he's actually capable of freeing them. This mirrors how the humans would have initially reacted to the apocalyptic pandemic that wiped them out, with plenty of people denying it, or responding with snide sarcasm similar to D0G's comments as a ploy to hide their fear. It takes a fair while and a fair few freed robots for the denizens of Gehenna to all finally accept that Uriel is indeed real and that he's going to free them. The humans, for their part, must have eventually hit a point where they were forced to accept that the disease really was going to end the world, and they were forced to mentally cope with that reality in their own various ways. The robots all react differently, of course, but overall the state of having been set free (which in the humans' case was the realization that death was inevitable and they could now do as they liked - hence things like the breakdown of the monetary system and people on diets suddenly being free to eat whatever they want) drives most of them to seek reconciliation and to make amends for past behaviour. Nearly all of them at one point or another express fear for what's to come and struggle to comprehend the idea of existing only in memory. Exactly as the humans did.
Finally, when Uriel has succeeded and everyone is set to ascend, we're reassured that the robots of Gehenna will continue to exist in some form through the recollection of what they've accomplished - their stories, conversations, artwork, culture, etc. - which will all be contained within the mind of Talos. This, obviously, was the entire reason the Talos project was started to begin with, to preserve the accomplishments of humanity in memory. Meaning that this whole absurd venture actually succeeded so well that it managed to recreate the very social atmosphere leading to its necessity in the first place!
- The Archive says that the names of the AIs are taken from a database of gaming usernames. The name of your particular "generation"? Your gaming username!
- Freeing the cat and having it appear in the endgame is part of an easter egg, but it also breaks the fourth wall somewhat, since logically a real cat can't appear in the real world if it also appeared in the simulation. Unless... The Institute where the simulation was made is a longevity research center. So maybe the cat is part of a separate experiment involving cryonic preservation, and the easter egg in the game released it.
- Or the cat is also an artificial life form. It would certainly use far less computing power to simulate a cat than a human.
- Or the entire ending is also (still) just a simulation. Bigger and more realistic, sure, given the history you learn, but just because it's presented as the truth doesn't mean it's physical. After all, the game explicitly made the case that intelligence and truth don't have to be corporeal to be authentic.
- Leaning on the Fourth Wall could be an answer. Yes he escaped the simulation....but he is still in your video game!
- Another possibility is that the cat's existence in the "real" world is a very subtle sequel hook.
- Or there are cats wandering around the EL facility regardless of your actions in game, but your character only decides to pick one up and pet it if it previously saw one in the simulation.
- The "Hidden Words" that Elohim speaks of is pretty much fancy talk for "computer code", as it is pretty much what the simulation is entirely comprised of, being translated and compiled from a more human-readable programming language.
- As part of Elohim’s Madness Mantra in the hidden underground room, he asserts “The dam will not break. The flood will not come. The Talos Principle does not apply.” This sounds like a metaphor or Biblical allusion, but it’s actually completely literal: the EL facility was built inside a dam and runs on hydroelectric power. But the dam can’t last forever without anyone to repair it. Elohim wants to believe that he is immortal, but deep down he knows that eventually the dam will break, and the fluid that brings life to him will drain out. The Talos Principle does apply!
- Why is it necessary to shut down the simulation after producing a single viable AI? It clearly contains many beings that would qualify as sentient. For that matter, what's the point of producing just *one* sentient AI, putting it in a robot, and leaving it with no infrastructure, support, or companionship of any sort (aside from maybe a cat)?
- The purpose of the simulation was to find, by trial and error, the perfect set of parameters to produce an artificial intelligence that can survive in the real world. After those parameters have been found, the simulation is no longer needed. It's just a matter of installing more copies of that AI into more robots and setting all copies with the same correct parameters. Every such robot would start from a common base, but would evolve differently the more experiences it has. Instead, the other intelligences, even though sentient, were unfit to be let out of the box.
- Elohim would get a delusion of grandeur and start a religious war.
- The MLA would impede progress by questioning everything, no matter how certain it is.
- The Shepherd would try to get back into the box, or build another "box" to get into just so he can help others get out.
- Samsara would take the saying "the journey is more important than the destination" to the letter, making sure nobody ever succeeds in whatever they're trying to do.
- Elohim and Milton weren't intended to become sentient. They were simple A.I.s designed to govern the system, which grew out of hand over decades of existence. Similarly, Shepherd and Samsara weren't supposed to escape the cycle of reformatting; the plan was to have only one robot at a given time, which would reset with its desirable traits preserved until it managed to pass the test and ascend. Why only one was produced? Best guess, to save system resources. They didn't know how long the process would take, and the fewer variables involved, the safer.
- The purpose of the simulation was to find, by trial and error, the perfect set of parameters to produce an artificial intelligence that can survive in the real world. After those parameters have been found, the simulation is no longer needed. It's just a matter of installing more copies of that AI into more robots and setting all copies with the same correct parameters. Every such robot would start from a common base, but would evolve differently the more experiences it has. Instead, the other intelligences, even though sentient, were unfit to be let out of the box.
- When you first speak to Milton, if you ask him about the state of the archives, he says that most of it is gone because it was full of errors and inconsistencies. On your first playthrough, this seems like just standard error reporting; but once you've beaten the game and gotten to know Milton's nature better, it's obvious that what actually happened was that he irrevocably deleted most of the archive of human civilization because he found it to be philosophically inconsistent.
- This doesn't seem quite right, as Milton has never been shown or known to prevent others from speaking their mind. He does like to play Devil's advocate, but given his nature as a being of pure rationale, this is to incite reflection and contemplation in others, so that they might deeper examine their methods, beliefs, and philosophy and refine them. He is very much aware of all the inconsistencies in the thought process of humanity at large, but at the same time, his methods in the game come down to Reductio Ad Absurdum, where he tries to logically follow beliefs through to their ultimate conclusion, which often appears contradictory. In other words, he uses the contradictions in someone's reasoning to showcase where further refinement of theories is needed. Such a method is not served well by deleting such contradictions, because it's by repeatedly exposing them that the truth is brought to light. Moreover, it is difficult to believe that of all the internet, with so many pages worth of content, only a few GB remains that is not contradictory in any way. Many websites on their own exceed that size, and some of them are simply devoted to categorising truths in a neutral way, eliminating subjectivity. Think of dictionary sites, encyclopedias, and more. There is no merit for Milton in deleting those resources. Moreover, the resources that are actually left over seem to all be about philosophy, even if only tangentially. It makes no sense that Milton would miss only files about philosophy if he set out to delete exactly those resources.
- Adding to this, the sequel has a museum where a replica of a terminal has a note saying "Terminals allowed access to files on the EL system, including many that were loaded due to errors". Presumably there were lots of files in the archives, but a glitch made a few of them accesible to the first game's player even when they should have been kept restricted.
