Electron Dance
28Oct/15Off

Talos, 2: I Am The Words

This is the second part of a three-part essay on The Talos Principle which includes commentary from writers Jonas Kyratzes and Tom Jubert. The first part was posted yesterday.

Talos-P2-QR-2014-12-12 01-43-36-50

“I see now that none of us are yet ready. The cycle exists so that we may improve ourselves. But the one who reaches the summit is not our superior, for they stand on our shoulders to reach it.”

-- The Shepherd v82.6.0174

When I first saw the QR texts, I sighed. Jesus, not another game with it’s-not-exposition-honest-guv graffiti on the wall! I initially paid little attention to them, reading them out of a completist need to do everything. But gradually I realised the walls held stories.

Spoilers for The Talos Principle follow.

“They began I think mostly as Jonas' domain,” says Jubert, “because he's great, like I said, at beautiful and expressive writing, and we wanted the QRs to bring the world to life and draw attention to its beauty. We always intended there to be a bunch of characters, but for a long time all we had were a handful of somewhat random lines for each.”

So my fear about the QR graffiti was actually the truth for a long period of development. What changed? “Towards the end we had enough main story content that we knew loads more about how the world worked, and what sorts of experiences the other AIs would be having. I was probably free and Jonas was probably busy, and on playtests I'd noticed the world still felt very empty – it could comfortably take a lot more texts.

“So, I sat down and I developed this main QR story arc with Sheep/Shepherd, who is there mostly to comment on things the player is experiencing, to sometimes elaborate, sometimes add doubts, and generally make it all feel more real and well-trod. I added arcs to each of the characters, along with a tombstone for most of them, so there is some closure. It was late in the day so I didn't know if it would work, but the team all said they liked it so in it went.”

There is ambiguity about how these QR conversations work. Does the simulation only run one child program at a time or several in parallel? Early in the game you see brief replays of other AI in the simulation, but are they actually replays? Are the child programs communicating on a “real-time” basis or only adding to the graffiti of their failed ancestors? The simultaneous presence of Samsara and The Shepherd during the endgame seems to complicate things.

Talos-P2-Palms-2015-04-09_00001

“This is Tom’s area of expertise,” says Kyratzes, “but I always figured some aspects of the simulation are running in parallel (or at least used to), while others are consecutive – basically it's all a bit of a tangled mess at this point, much like evolution in the real world.”

Jubert: “That's the thing, right – once you establish that time doesn't work for them like it does for us, you can just appeal to Kant in claiming that our perception of time is fundamental to our way of being, so there is no way to represent this parallel processing, and anything goes.”

The saddest aspect of the QR texts is that as you move further towards the end of the game, they thin out, making it clear how many programs never made it through and the world grows more cold and lonely. Rather than the Garden of Eden, the simulation is both purgatory and graveyard – many of the child programs fall into depression or go mad.

This ethical aspect of the simulation is never made explicit, although Gehenna does sail much closer to these particular rocks. Elohim continues to iterate from one child program to another, attempting to find the program that will finally defeat the simulation and thus be the candidate for upload to physical hardware. But many of the AI already seem to be self-aware and these programs are being put through mental torture. I’m not sure the creators of the Process, Alexandra Drennan and her team, thought this far ahead, as if they assumed only the successful AI would feel anything at all. But they were trying to save humanity and these abused children were the inevitable price.

I found it disquieting that all those programs who had tried and failed were not given a chance of a proper life, because only one was allowed to transcend. The simulation appears to be wiped at the end of the game – a moment I found gut-wrenching for this reason and also the fact that I’d played Talos over nine months and considered it a sort of home – which suggests thousands of stored AI were lost in the final moments. In particular, the Process could not complete without The Shepherd sacrificing itself to assist you.

Talos-P2-Shepherd-2015-08-15_00001

“The honest thing with The Shepherd,” Jubert explains, “is Croteam basically laid down the law and said ‘There will be a good robot and a bad robot helping and hindering you on the final level’. It seemed like an idea we could use, so we just had to find something to do with it. We already had this running theme about co-operative self-sacrifice for the greater endeavour, so Shepherd's final story arc was a foregone conclusion.”

If anything, The Shepherd’s role makes the Process seem more arbitrary. Elohim apparently tried to undermine the Process, to prolong his own existence. Did he deliberately break the final challenge of the tower so that a single AI could not ascend alone? Or was this requirement designed in? Consider the simulation supports QR code-based discussion between AIs which potentially makes it easier for later generations to progress. Also, the messengers to help out struggling child programs. The Process appears to be a collaborative effort thus it seems heartbreaking that the transcending AI is just in the right place at the right time to collect on the jackpot. Jubert cites the running theme of co-operative self-sacrifice, but the child programs were just children who barely had a grasp of what it was to be alive. The successful AI is the sum of everyone’s efforts, the sum of their pain. Why is there just one chosen one?

One cynical answer would be there is only one body out there in the real world to host an AI and any program who completed the ascent would be “good enough”. Except, well... The Road to Gehenna. We’ll come back to this.

I also noted a child program must defy and overthrow “God” to complete the Process, so is the Process designed to spawn an atheistic species? “Well, I'd like to think that it's not so much about completely rejecting Elohim,” says Kyratzes. “You do solve the puzzles like he tells you to, and much of what he says is true – as it is about questioning him, doubting him. Finding a path that's neither Milton's blind nihilism nor Elohim's blind faith. Of course, faith does take a different connotation here, one more related to faith in the project of civilization than in a literal God.”

It’s made clear in The Road to Gehenna that Elohim was undermining the Process, but I’ll admit I didn’t really get that from Talos, just that he feared his own end. As far as I can tell, in Talos, there’s no evidence Elohim did anything wrong. I wasn’t sure if this was a glitch in the writing, a narrative sketch the writers forgot to colour in.

Talos-P2-CrackleEnd-2015-08-15_00002

Apparently not, says Kyratzes. “I always intended for a certain amount of ambiguity there. It's the messiness of the story – how the simulation was made out of disparate parts, how it malfunctioned in odd ways, how maybe the very malfunctions ultimately helped it succeed – that makes it human to me. Elohim is part of that. He's supposed to be challenged, yes, but he takes that too far, out of desperation, out of fear. To the player's character Elohim is just a test, but to himself the whole God thing is more than an act, it's his whole reality. Maybe he endangers the entire Process by his actions. Maybe he actually does a better job than intended by accident. Maybe underneath it all he always knows how it's going to end, no matter how much he denies it. To me, that messiness and ambiguity is a more realistic reflection of how we deal with these things in the real world.

“I should probably also point out that there are deliberate references to Jesus and his moments of doubt in Elohim's lines, another syncretic element of the story, another retelling. There's a reversal of roles at the end, God in his doubt submitting to man: let your will be done. The end of the game, in many ways, is about Elohim's moment of humanity, a kind of precursor to the humanity about to be realized.

“I could go on but I'm afraid it would get even more pretentious.”

I do not doubt these words.

My experience with Talos was particularly long. I played in a short bursts over a period of nine months and this meant the already fragmented narrative hints were difficult to keep track of. Although I could go back and review every document I’d uncovered, there was a mountain of them and I was loathe to waste time retreading through old text. I became quite lost in the story, more so than someone who had completed it over a couple of weeks. At Talos’ heart is a mystery: what is the purpose of Elohim's garden?

Talos-P2-ElohimSecret-2015-05-01_00002

This is where I turn to another character I haven’t mentioned so far: Milton, the AI in the archive. I could easily squander many paragraphs on discussing the nature of Milton; is he a glitch or a deliberate part of the simulation? Without going into too much detail, I lean towards Milton having a deliberate role to teach scepticism as a counter to Elohim’s authority, the grit that makes the pearl of an independent AI. But Milton did such a good job of undermining all of my theories over my nine Talos months that I never settled on an explanation for the game until he was finally banished! (I later discovered that some players have other options concerning the fate of Milton.)

This meant I was a little disappointed in a seemingly-obvious resolution of the plot – oh it really is a virtual test where the final challenge is to defy God – because Milton kept me off-balance for such a long time. But the main obstacle was the long duration of play, so the recall of earlier narrative discoveries was always sketchy. It was evidence that a complex story for a game that encourages lengthy play would be a difficult ask for the average player, especially one who might not be willing to commit to reading each and every text available.

“Honestly, I think it's impossible to fully predict how players will interact with a game,” says Kyratzes. “Some people play it in long sessions, other stretch it across months, some pay attention to one type of detail, others to another. And even while you're writing, stuff may change, puzzles may get moved around or cut, and so on. So I felt the best thing to do was just to make the world rich and interesting and consistent and hope for the best. That's pretty much the case with every game.”

“I think it's non-ideal,” adds Jubert. “I wouldn't want to take it on on my own. Talos is huge. With the expansion we've basically written a novel, and I am by nature a short story writer. I like to get in, make my point, and get out and do something new. Most games are ideal for that because a ten hour game fits, very roughly, the same number of plot points and words as a long-ish short story, or an episode of Star Trek. But with Talos, you're likely there for 20 hours. That kills our ability to pace properly. It means we either have to flood the game with narrative content, or risk it feeling empty. A player like me simply wouldn't finish a game like that. Jonas, I think, is more comfortable with the longer form.”

There’s one other narrative problem: the Easter eggs. I hated them. That feeling of narrative whiplash I had when I unlocked the “secret ending” of Kairo (Locked Door Puzzle, 2012) to discover it was a fourth-wall breaking experience? This, again and again. We must save humanity! The AIs must thwart Elohim! Aaaaand then Fork Parker is playing with a pile of money and Aperture Science is hidden behind the Moon and let’s fly around with a jetpack. You’re playing a game, kid, never forget it.

Talos-P2-EasterEgg-2015-08-14_00001

Kyratzes was okay with them. “The setting can explain them away and most people really enjoy them, but I would've been happier if we’d been given a chance to properly integrate them. It’s... I don't know exactly what to say. In Gehenna, there's an Easter egg that is simultaneously brilliantly deranged and totally destroys the ending. I have to admire the inspired madness behind that, even though it's not something I'd do in one of my own games. Sometimes you just say "videogames!" and get on with your job.”

“I think Jonas yells 'videogames!' in his head at least a few times a day,” says Jubert. “I always say about the Easter eggs that if they devoted half that time to doing story-related level design we'd take this game up a notch – and that will be something on the table for the future. However, they're often brilliant, and they are Croteam's hallmark, so what are you going to do?”

Kyratzes and Jubert made some explicit references to the Talos Easter eggs in Gehenna, such as the sheep dream and the cat, which is an interesting attempt to normalize some of Croteam’s excesses ex post facto.

Let’s now turn to The Road to Gehenna which, to my complete surprise, was the superior game in many ways. I also like to think of it as the most commercially successful Twine game.

Next: The Road to Gehenna... and Talos 2?

Talos-P2-Mountain-2015-05-02_00005

Download my FREE eBook on the collapse of indie game prices an accessible and comprehensive explanation of what has happened to the market.

Sign up for the monthly Electron Dance Newsletter and follow on Twitter!

Electron Dance Highlights

Comments (4) Trackbacks (0)
  1. “Gradually I realised the walls held stories.”

    I’m surprised that the characters and their stories in the QR codes were such a late addition—although it’s clear from the nature of their integration that they were designed to avoid impacting any other game systems. But they added much richness to the world, so I am glad they made the cut.

  2. Yes I agree the QR codes were an important part of Talos. I can’t imagine the game without them. You don’t miss them in Gehenna, of course, because the forum picks up the baton there.

    I find it interesting that the AI in the simulation actually have no way of communicating except via terminal or paint. You can’t talk to The Shepherd or Samsara during the endgame. Obviously part of the reason is technical but the game is confident enough to not even bother explaining it. It just is.

    (There was so much stuff I could have chatted about but these articles would have badly frayed at the edges.)

  3. “You can’t talk to The Shepherd or Samsara during the endgame.”

    It would definitely have helped me if I could. I spent what felt like ages trying to figure out what this mute, unmoving robot was there for—and later, what it wanted me to do next. And at first I thought Samsara was the same robot that had been working with me before and couldn’t figure out why it wasnt helping again.

    (That ending sequence! Whole new puzzle techniques to figure out under time pressure, and having to restart it all on failure! So not good.)

    Incidentally—and I’ll be circumspect to avoid accidentally spoiling anything—one room and its commentary in The Beginner’s Guide made me think of Talos’s QR codes.

  4. You’re right. The timed nature is the endgame did step up the “urgency” of the story but I don’t think it was really thought out that well. Fortunately at least there were some checkpoints. I had meant to note with disdain the sudden injection of fast puzzle solving at the end but alas I forgot. When I ascended on the last fan I was relieved I didn’t have to do any more of *that*.

    I guess another problem is that it felt extremely scripted, almost trying to figure out hotspots to move the action forward. It didn’t have the feel of an epic showdown and there wasn’t enough sense of kinship between you and The Shepherd. Altogether the final gauntlet doesn’t make as much sense as I’d hope it would.

    I know the room you mean in The Beginners Guide!


Trackbacks are disabled.