The Talos Principle 2

[pakoito]

I finished the game last night and 3 others captured it perfectly. Beautiful game, the puzzles are much better tuned, the story is engaging but the dichotomy scientism vs bible thumping is just very weakly presented. The C&C leads to 3 endings, where each has at least 2 minor variations that I'm aware of; and a few post-credits scenes to top it off if you've completed some tasklists during your run.

The length is good, a bit longer and it would have dragged. I did not feel like completing the "hard" puzzles or the stars and went to see the alternate ending cutscenes on youtube. Yep! not worth it unless you feel the burn of completing puzzles.

I liked the grandiosity of the scenery, although I'd liked a bit of environmental storytelling here and there. I didn't mind walking around because the views were great.

 

[Neuromancer]

The recorder sadly doesn't make an appearance.

To be honest: For me personally that is a good thing, because I didn't like them very much in the first game.

Tastes are different:
Some hate music/sound orientated puzzles - which I love -, others like myself hate time based or time restricted puzzles.


You have to "time" you actions by e.g. just standing around for some time waiting - and then in execution, if you just made one small movement error, you'll have to replay the whole thing again and again.

Especially in more difficult puzzles later, this process could be quite annoying and time consuming.


I had the same problem with Gateways - a game from 2012 (Basically similar puzzle style in 2D, but without the philosophical undertones):
https://www.mobygames.com/game/101767/gateways/

 

[Beowulf]

I have finished the game now, and I agree that / is a very apt rating.
The golden puzzles are finally a proper challenge, but I have also changed my mind on the difficulty curve. If one plays in short bursts over many weeks the gradual increase of difficulty that is reset between the areas might indeed be batter solution then increasing the difficulty over the whole game, as it would result in all the late game puzzles being tough to get into after taking a break.

The plot and the characters are cliche ridden and the events and dialogues somewhat predictable, thankfully it's not the main focus of the game.

However the games suffers from bugs in later areas, mainly in the form of falling through scenery.
The Prometheus flame puzzle is also quite week and I begun to dread the chase music after a short while.
But it looks gorgeous, the areas differ and some show really nice eye candy and the puzzles never get boring.

I hope that going with full voice acting instead of text adventures will not make the DLC's creation a problem though.

 

[3 others]

The recorder sadly doesn't make an appearance.

To be honest: For me personally that is a good thing, because I didn't like them very much in the first game.

Tastes are different:
Some hate music/sound orientated puzzles - which I love -, others like myself hate time based or time restricted puzzles.


You have to "time" you actions by e.g. just standing around for some time waiting - and then in execution, if you just made one small movement error, you'll have to replay the whole thing again and again.

Especially in more difficult puzzles later, this process could be quite annoying and time consuming.


I had the same problem with Gateways - a game from 2012 (Basically similar puzzle style in 2D, but without the philosophical undertones):
https://www.mobygames.com/game/101767/gateways/

I don't actually disagree with that at all. The recorder was a fiddly mechanic because it was too easy to do something irreversible and have to restart from the beginning. And lots of solutions required you to stand around idling for some time to leave enough buffer in the recording to get all your planned actions done (and if your plans weren't right the first time, now your only jammer is on the wrong side of the forcefield and there's no way to get to it oh well). I just loved the novelty of it, and the scope of possibilities that opened up with the playback items. I don't think there were too many puzzle games utilizing the temporal dimension back in 2014. Braid had a few, perhaps. Maybe I'll check out that Gateways thing.

 

[Star Citizen]

Started it up last night. Think the talking robots are a bit too much for me. I muted all dialogue and hid the hud to stop their avatars from popping up on the screen between their radio banter, but you still get locked in dialogue and I can still hear those horrible voices in my mind, spewing cringe nu-dialogue. Can't mute that shit unfortunately. Still early in but IMO the game would have much benefitted from subtle environmental storytelling rather than the "cinematic" approach of having cutscenes and dialogue with NPCs that break all immersion for me.

 

[3 others]

Goddamn these golden gate puzzles could be tricky. And they have such cheeky names that usually contain a nugget of a hint: "Embodied Cognition" lol. I really loved the few "Gehenna admin -style" puzzles with only a minimal number of moving parts that still have an elegant, yet non-obvious solution. The soundtrack is also growing on me more and more with these 30 minute iteration sessions.

 

[3 others]

One expanded note on the writing and storytelling in this game. I wrote before that "the core flaw of the game is that you're supposedly being prodded towards making a monumental choice between two fundamentally different future visions for the society and one of them is so presented obviously correct it makes Bioshock's save/harvest dichotomy seem like a well-nuanced masterpiece.".

Well, first of all, there are actually 3 different options, but the fact remains that only one pathway is written in such a manner that the characters themselves seem to believe in it. Contrast this to the archaic Alpha Centauri, where Sister Miriam is cast as the surface-level luddite baddie, but the further you get, the more she comes across as the sanest leader in the game. Her message of dissention in the face of an increasingly inhuman pathway towards "Transcendence" is presented just as sensible as the manic techno-utopias and xenoecological theocracies of the other factions, if not even more so.

Talos Principle 2 could in theory have similar sorts of factions (skeptical / lukewarm / utopian) but the writers' hearts aren't in it at all for anything less than grabbing Promethean Fire by the pussy. The Mayor - who ostensibly is supposed to represent the cautious skeptics - is written as the stick-up-his-ass bureaucrat archetype with a slimy, faggy voice and repetitive arguments "I still think we should shut everything down just in case" whoa there really compelling. A total missed opportunity. An inspiration for the competing vision could have been taken from either some sort of Kaczynskian primitivism, or (I'm assuming this would be more up the writers' alley) from the safety-first faction of the current "AI ethics" debate. Just pick one of the curlyhaired half-Jew mulattoes populating that scene and model the skeptic spokesman on her.

This thing still annoys me. The core theme of Talos Principle 2 is very good, but both the style (repetitive banal chitchat) and substance (formulation and presentation of the core questions) are very disappointing.

 

[Israfael]

For me that was a problem even in the first game --how is Plato related to AI development

Very directly - Plato was an idealist who believed that ideas are separate from the forms (aka they exist independently and that there is 'ideal' chair, spoon, mountain, river etc that all material objects derive from), while Aristotle and more materialist classic philosophers believed otherwise (including in-game Straton (which is a name of purported materialistic-inclined pupil of Aristotle) of Stageira (which is the home town of Aristotle not-so-coincidentally)). So both games revolve around the discussion whether 'pure mind' (ideal conception of conciousness / self-aware thinking being) can exist/form.

Furthermore, the name of "Institute of Applied Noetics" directly refer to neoplatonic concepts (idealistic notions of what conciousness and mind are) that were popular in the West during the start of the Cold War, so these games are actually very related to Plato and discussions that started with his idea of existence of, well, ideas.

 

[Israfael]

anything less than grabbing Promethean Fire by the pussy.

How so? Imagine giving a universal constructor machine (which Athena created) to an insane ageless robot that can turn reality into nightmare (which Athena actually did), and it's all represented in the game. If anything, it prods you towards something in the middle (between Major decaying stasis and Byron's unbridled brinksmanship exploration drive) rather than anything else. What was really cringeworthy was the typical American idea of 'baby as a (God-given) miracle', this is just too much for me.

 

[Iucounu]

So both games revolve around the discussion whether 'pure mind' (ideal conception of conciousness / self-aware thinking being) can exist/form.

But do they really? Admittedly I didn't pay too much attention to the computer terminals in the first game. But the puzzles are just logical problems, and the instances where the player can choose to be disobedient don't necessarily imply self-aware thinking (or lack therefof): surely you can be both conscious, obedient and bad a puzzle solving at the same time? Indeed a human player should by definition possess "human consciousness" (unless someone trains an AI to play the game), but can still fail at the game's puzzles. In fact I wouldn't be surprised if an AI would be better at solving them than a human.

Meanwhile the plot seems to be about testing the AI protagonist to determin if he's conscious (if I remember correctly)? But since the human player is by definition conscious, such tests (if designed properly) should be impossible to fail by a human.

So I believe the puzzles are just entertainment for human players, and so by definition can't be relevant to the plot's AI tests (since the player is not an AI like in the plot).

Furthermore, the name of "Institute of Applied Noetics" directly refer to neoplatonic concepts (idealistic notions of what conciousness and mind are) that were popular in the West during the start of the Cold War, so these games are actually very related to Plato and discussions that started with his idea of existence of, well, ideas.

That sounds more like superficial nods. What I meant to say in my previous post was: is Plato's thinking relevant to AI researchers?

 

[Israfael]

What I meant to say in my previous post was: is Plato's thinking relevant to AI researchers?

Depends what you imply by AI - is it sentient self-aware being or current neural networks that are just an advanced form of pattern recognition software (aka puzzle solvers). It is relevant (historically speaking) for the first question and probably not really relevant for the snake oil that is highly popular today. It is relevant because Plato posits that you can 'conjure' (for a lack of a better word) intelligence that is pure, in other words free from the knowledge of the environment ('world' if you prefer this term) and self, while materialists believe that conciousness and mind are a product of material conditions (aka it can't exist without those contraints and initial conditions)

by definition can't be relevant to the plot's AI tests

The chief AI developer (Alexandra Whatever?) claims in one of the first audio logs that puzzles and games are a thing that played major role in development of human intelligence and conciousness (which is a very popular idea in the field of ethology and everying theory of mind / intelligence related), so they try to speedrun this process in an iterative simulation (which is also mentioned in the logs)

PS To think of it, current snake oil machines are actually trying to do the reverse Plato thing - they analyze myriads of images / samples of curated data to find the ideal representation of a cat, pug, nuclear submarine sonar signals that will allow to identify such objects (of this class) in arbitrary data from the material world

plot relevant points

Also, if you do the extra puzzles, you'd know that the two ancillary AIs that were created to build the testing environment - puzzle generator HIM (turned into EL-0-HIM) and librarian MLA realized that if they complete the goal (iterative development of sentient AI that will defy the iteration loop) their existence will end so they took some 'measures' to prevent that scenario from happening. Something like that (AI going rogue) happened in real life in which AI that controlled drones in a virtual simulated war environment at some point started to kill human operators as they prevented it from killing more targets

https://www.reuters.com/article/fac...as-hypothetical-air-force-says-idUSL1N38023R/

 

[Forest Dweller]

It's now saying Windows 10 exclusively, so I guess I'll be torrenting this bitch.

 

[Iucounu]

It is relevant because Plato posits that you can 'conjure' (for a lack of a better word) intelligence that is pure, in other words free from the knowledge of the environment ('world' if you prefer this term) and self, while materialists believe that conciousness and mind are a product of material conditions (aka it can't exist without those contraints and initial conditions)

But the AI protagonist in the game already knows how to interpret visual input, spoken and written language, how to walk, lift things etc (all a priori), which seems to contradict Plato as soon as the game starts.

by definition can't be relevant to the plot's AI tests

The chief AI developer (Alexandra Whatever?) claims in one of the first audio logs that puzzles and games are a thing that played major role in development of human intelligence and conciousness (which is a very popular idea in the field of ethology and everying theory of mind / intelligence related), so they try to speedrun this process in an iterative simulation (which is also mentioned in the logs)

No idea if modern ethologists claim that, but I recall problem-solving abilities as an adaptive advantage during all of natural history was the theme in Lorenz' Behind The Mirror.

In any case a chess computer is not considered conscious just because it can defeat a human.

PS To think of it, current snake oil machines are actually trying to do the reverse Plato thing - they analyze myriads of images / samples of curated data to find the ideal representation of a cat, pug, nuclear submarine sonar signals that will allow to identify such objects (of this class) in arbitrary data from the material world

True, in that sense they may even mimic the natural evolution of cognition (which seems far from being conscious).

plot relevant points

Also, if you do the extra puzzles, you'd know that the two ancillary AIs that were created to build the testing environment - puzzle generator HIM (turned into EL-0-HIM) and librarian MLA realized that if they complete the goal (iterative development of sentient AI that will defy the iteration loop) their existence will end so they took some 'measures' to prevent that scenario from happening. Something like that (AI going rogue) happened in real life in which AI that controlled drones in a virtual simulated war environment at some point started to kill human operators as they prevented it from killing more targets

https://www.reuters.com/article/fac...as-hypothetical-air-force-says-idUSL1N38023R/

I don't remember the plot that well, does it imply EL-0-HIM and MLA had already become sentient? The AI in the link sounds like it's acting rationally on poorly written instructions.

 

[Israfael]

But the AI protagonist in the game already knows how to interpret visual input, spoken and written language, how to walk, lift things etc (all a priori), which seems to contradict Plato as soon as the game starts.

May as well be, might be not. Our mind is heavily 'augmenting' what we see, hence all the visual illusions and various artefacts of senses. Who knows what in-game AI percieves and is it actually reminiscent of what we see/hear/do. In any case, as you rightly pointed out, the question is what is considered conciousness and cognition, are they different things, the same or maybe parts of bigger whole? Even before the advent of "AI" and AI this question bothered some people, as you can see from writings of both Plato and Aristotle.

In any case a chess computer is not considered conscious just because it can defeat a human.

No, but there's vast difference between specialized neural networks and generic ones. For example DeepMind initially created highly specialized systems and then they branched out to generic AI which seems to be able both to play various games with zero prior knowledge, do science stuff, fold proteins and etc. The question is it all real, or clever sham and whether this actually matters to an external observer (the oft-mentioned "Talos principle" in the game)

does it imply EL-0-HIM and MLA had already become sentient?

There is no definite answer and probably won't ever be. As you said, they had 'poorly written instructions' that were interpreted in some unintended (by the in-game IAN researchers) ways. There's actually a good Soviet comic strip about robots and AI, illustrating this problem. But it's just minutiae, in my opinion. The pertinent question is what the IAN team tried to create - pure mind, a copy of human intelligence, mimic of human or something else. Second game is more about this question, in my opinion.

Robo-Fortran, peel the potato

 

[Israfael]

It's now saying Windows 10 exclusively, so I guess I'll be torrenting this bitch.

Considering it's UE5, it probably requires LATEST W10 version. I've tried running some other modern "games" on a not so up-to-date W10 installation (1909 i think) and it simply refused to launch because Shader 6.5 or something is missing.

 

[Jaedar]

Started playing this game, I'm 4 areas in.

The start of the game made me consider refunding, walking around the city with no puzzles, dumb conversations with dumb people. I really was dreading that the game was going to preach eco-conservatism and degrowthing. But now, I wonder if that part of the game isn't bad on purpose so the rest of the game can dunk on it.

Puzzle mechanics are not very inventive so far, and the maps are too big. My pinky hurts from holding down the shift key. The first game also had the first problem, but I don't remember the maps being quite so big.

 

[3 others]

There's an autorun toggle in the menus, you know

 

[Jaedar]

There's an autorun toggle in the menus, you know

Yeah, but I don't like sprint toggling. Maybe I should enable it anyway.
Made it to the 6th area. The mechanic with the item swap tables feels incredibly uninspired compared to pressure plates.