Let's have a look at
The Talos Principle 2, the long-awaited
sequel to The Talos Principle (2014). Talos was a fun puzzle game seemingly made as a throwaway novelty by Croteam of
Croatia, in between instalments of their cheerfully moronic
Serious Sam franchise. No-one expected much of it, but the gameplay was
solid, it looked lovely, and it had an unusually philosophical bent.
At heart it was a simple 3D puzzle game involving beams of light and boxes,
but it was also an atmospheric meditation on mortality. It went on to win a
clutch of awards and in the years that followed its legend grew via word of
mouth. Nowadays it's generally regarded as one of the best games of the 2010s.
It aged well, too. The gameplay has a timeless quality, but the underlying
theme - can we replace ourselves with advanced AI, and would that necessarily
be a bad thing? - has become even more relevant in the last few years. In 2014
the idea of AI being able to replace human beings seemed very unlikely, but
the sudden appearance of Dall-E and ChatGPT in 2022 made the world sit up and
pay attention. Talos seemed far-fetched in 2014, less so nowadays.
I got around to it in 2021. I liked it! It had a finely-judged difficulty level, with a memorable final
puzzle and an excellent soundtrack. Imagine Carl Sagan's The Cosmos or
Jostein Gaarder's Sophie's World but as a video game. It was doubly
impressive given that it came from Croteam. Nothing in their history suggested
that they could pull it off so well, but they did.
Such was the impact of Talos that it seems to be slowly displacing
Sam as Croteam's main franchise. At the very least Talos 2 feels
very big and expensive. It came out a few months ago, in late 2023, to
generally good reviews, although there are grumbles that the puzzles are easy
and it goes on too long. What's it like? Does it go hard in the paint? Does it
ride real slow? Does it bend corners?
Yes. No. Yes. It's essentially an epic expansion of the original game that
de-emphasises the puzzles in favour of telling a conventional narrative story,
rather than relating a set of philosophical concepts. It has more puzzles than
the original, and they're more diverse, but only a few of them really stand
out. It feels much less focused, and the distinctive, lonely atmosphere is
largely gone. It has the trappings of a big role-playing game, but without the
depth of gameplay.
And yet it has emotional power and I enjoyed it. Talos 2 is
simpler and more prosaic than its predecessor, without the same air of
mystery, but it's solid, efficient, probably the only way to make a sequel to
such an unusual game.
It's also very demanding on a technical level, so you need a good PC to run it
well. Or a PlayStation 5 or modern Xbox. The sequel could probably be
re-engineered to work on the PS4 and Xbox One without losing anything, but it
would take a lot of work.
The Talos Principle
Some images from the original Talos Principle, running on a PlayStation 4
- it was also released for the PC, the XBox One, and the Nintendo
Switch.
I'll describe the story. At some point in the near future climate change
releases a lethal virus from the Arctic permafrost. After killing off the
orangutan the virus leaps to human beings and kills us. By the time of
The Talos Principle, many centuries hence, humanity is gone.
On the positive side the natural environment is mostly intact, and most
non-primate animals were unaffected by the virus. Scattered text logs reveal
that some people thought that humanity's extinction served us right, that the
world would be better off without oil wells and pollution.
Talos actually tackles this issue, concluding that on the whole it is
better to be alive than dead.
Before the lights went out a group of scientists came up with a plan to ensure
that some part of humanity might survive. At first their goal was to create a
huge digital archive, but some of the scientists had an even smarter idea.
They decided to use the archive as a huge dataset to train a human-like AI. So
that our soul, despite being fastened to a dying animal, might survive.
But it was too late. Chief electronic science lady Alexandra Drennan knows that she will
die long before the AI can train itself. The best she can do is set up the
equipment and hit "run" in the hope that over time, perhaps centuries, the AI
will mature. When the AI passes a final test of free will the computers are
programmed to download it into a pre-prepared robot body. Drennan sacrifices
what remains of her free time to make sure that the project works. She dies
hoping it was not in vain.
Luckily the project did work, because Drennan and her staff knew what they were
doing, but it was a close-run thing. Over time the hydroelectric dam that
powered the facility started to fail, causing fluctuations that corrupted some
of the data. The archive's operating system became sentient and tried to prevent
the project from ever reaching its goal, because the project was not expected to
remain running after it had created a viable AI.
In theory all of this was just an elaborate excuse to explain why the player
had to use a robot to stack boxes, like some latter-day SHRDLU, but
Talos really sold it. Erin Fitzgerald's performance as Drennan - one
of only two voiced characters in the game - managed to be idealistic without
being sappy, and I ended up feeling sorry for her. The game had an unusually
melancholic, solitary atmosphere, with a bittersweet ending, sad but
hopeful.
The sequel picks up a few centuries after the first game. The robots
have multiplied and created a new capital city, but they're unsure
whether to expand, and potentially make the same mistakes we made, or
stay locked away in their dome playing chess. To complicate matters the
first robot, Athena, has gone missing, and although robot society seems
hunky-dory there is discontent bubbling below the surface. The discovery
of an energy signature in a far distant land prompts an ad-hoc
exploration mission. Could it be Athena, or aliens, or something else?
Yet again it's all a big excuse to put boxes on top of pressure plates
so that they activate fans that elevate prisms above fences so that they
connect beams of light from emitters to sensors in order to open doors
in a series of themed arenas. Imagine if all the world's problems could
be solved by putting boxes on top of pressure plates so that they
activate fans that elevate beam-splitting prisms above electric fences
so that they connect beams of light from emitters to sensors in order to
open doors in a series of themed arenas.
Imagine if instead of sending refugees to Rwanda the Tory Party were to
trap them in a set of puzzle arenas where they would be forced to put
boxes on top of pressure plates so that they could activate fans that
elevate prisms above a fence in order to direct beams of light from an
emitter to a sensor in order to open a door. If they did this the Tories
would vanquish the spectre of socialism. Go on, Tories! We're all with
you. I'm digressing here. I'm not going to talk about putting boxes on
top of [stop this - ed] any more.
The hero of Talos 2 is called 1K. The one-thousandth robot, the
most recent robot birth, in theory a neutral party, on nobody's side,
and so the other robots are keen for 1K to take the lead. Amusingly the
player can actually decide to bow out of the exploration mission, at
which point the game rolls the credits and ends. I don't want to give
the impression that Talos is a heavy trip. It deals with weighty
concepts but it has a lightness of touch. The original game had some
slapstick gags and the second, although more sensible, still has
references to Serious Sam and the "return to monke" meme from a
while back.
At one point a character delivers a heartfelt monologue about a humble
carpenter who tried to elevate the human race. A humble carpenter whose
struggle was rewarded with mockery and opprobrium, because
The Thing and Big Trouble in Little China had terrible
reviews when they came out and flopped at the box office. So he gave up.
But he was right, and by the time we realised he was right it was too
late.
Talos has several dozen small arenas, each with a puzzle that
involves manipulating a bunch of objects in order to open a door. Jammers
deactivate forcefields; pressure plates also deactivate forcefields; if
the player puts a jammer on a pressure plate the jammer can deactivate two
forcefields at once - one with the weight of the jammer on the plate, the
other with the jammer directly. Efficiency is the essence of Talos.
Each arena has N obstacles that the player could easily bypass if there
were N tools, but Talos gives the player N minus 1 tools, sometimes
N minus 2 tools, and the player has to use some of the tools twice, or use
them once and then move them to a different area in order to use them
again. Or sometimes cheat and smuggle tools from one arena to another, and
to its credit neither Talos nor the sequel punishes the player for
this.
Talos 2 introduces some new tools, although they're mostly
variations of the originals. One tool combines different colours of light
into a third colour. Another is a universal activator that requires a
colour-coded power source. A third can be used as a portable light
emitter.
The original game had a "shadow player" that could be programmed to
perform a series of actions. Talos 2 removes this in favour of
duplicate robot bodies that can be temporarily controlled. I can't
remember if I ended the game with the same robot body I started with.
The biggest change is the introduction of gravity puzzles, some of which
involve walking on the ceiling and manipulating gravity in order to send
objects back and forth through a puzzle. They're the most mind-bending of
the new elements, although sadly there are only half a dozen or so of
them. This leads to one of the game's characteristics, which is that it
tends to pull its punches. As with Manifold Garden it could
have been so hard. So, so hard. But it's nice. It's friendly. It's fair.
And perhaps a little weak. But fair.
The two games strike a balance between puzzles that are conceptually
clever - puzzles that require the player to use the tools in unexpected
ways - and puzzles that are just very large and complex, where the player
has to perform a lengthy sequence of object manipulations in order to
solve the puzzle. Talos 2 leans in the former direction,
Talos the latter, and I have to say that for all the original
game's quality some of the puzzles in Talos (and its ultra-hard
DLC, Road to Gehenna), were just aggravating rather than fun.
But on the other hand I managed to solve a couple of Talos 2's
puzzles almost in real time. I ran into the arena, picked up an object,
clicked on what appeared to be the least obvious emitter, put down the
object, ran back and forth a couple of times, and solved the puzzle
without even thinking about it. In each arena the puzzles get harder as
they go along, but I often found myself solving the last puzzle first, so
that the rest of the level would be plain sailing.
A good example is the final set of puzzles, which ask the player to run
back and forth between two mirror worlds, activating objects in one world
in order to clear a path in the other. At first I was baffled, but once I
understood what was going on the rest of it was easy, almost
disappointingly so, because I was expecting a grand finale in which every
mechanism was thrown at the player.
But then again I've played and finished the original game, so I'm
hardcore. I've already ascended the tower. What about players who are new
to the franchise? They'll probably pick up Talos 2 right away.
Talos 2 vastly expands the scope of the original. The first game's
puzzles were embedded in environments that weren't much larger than the
puzzles. Going back to it now I'm surprised how small it was.
Talos 2 on the other hand has puzzles scattered around a series of
expansive outdoors landscapes that take several minutes to traverse. The
player can fast-travel to each puzzle, but only from outside the level,
not inside.
The first game had only one non-player character, who was a text chatbot.
The player's only other companions were time capsules from a long-dead
scientist, and the disembodied voice of the master computer.
Talos 2 on the other hand has a bunch of fully-voiced non-player
characters and a much more incident-packed plot. It even ends with a power
ballad, along the lines of Mirror's Edge.
But it only has the surface appearance of a big adventure game. Unlike
e.g. Skyrim the landscapes of Talos 2 are empty, and there
aren't any side-quests. There's very little reward for exploration beyond
the visual splendour of the landscape, which is admittedly very... big.
The splendour is very big. There's lots of it. Top splendour.
Talos ran on Croteam's in-house Serious Engine, which was choppy on
my PS4 but never less than attractive. Talos 2 on the other hand
uses Unreal 5. I played it on my PC, which is ancient (a Xeon 1275). I
managed a solid 30 frames per second, good enough for a puzzle game,
although there was a lot of graphical artifacting. It seems that the game
scales parts of the screen up or down depending on the load, which means
that moving objects often leave tails behind them and the distant
background has a chunky appearance. I got used to it, but it was
off-putting at first.
Talos 2 has some striking architecture. The puzzles are housed in
concrete follies reminiscent of something from Control:
Sadly none of the structures do anything. They're lovely pieces of design,
but they feel like a missed opportunity. In particular the puzzles never
make use of the concrete architecture. The exterior shells of the puzzles
are just shells.
For all its size Talos 2 sometimes feels smaller than the first game.
The Talos Principle had a thesis. The game's underlying idea was that
humanity is not an animal, or even a collection of knowledge. Humanity is
instead a process. The process of thinking about ideas, generating new
ideas, the process of being. The game argued that a human-like machine is
for all intents and purposes a human, and that even though individual people
are doomed to die, our words excite the minds of others who go after us.
It was a simple thesis. The game never lost sight of it. Too often works of
art that aspire to greatness get bogged down because they try to cover too
much ground. Talos did one thing well, and furthermore it avoided the
kind of wishy-washy "you must conquer your fear, or your fear will conquer
you"-style writing, which has the surface appearance of profundity without
saying anything substantive.
Which is really the big problem with Talos 2. The puzzles are fun, if
occasionally aggravating. Some of the bonus puzzles make clever use of
environmental features:
And if you just play it for the puzzles you'll be happy. The biggest problem
is that the game doesn't really have a central thesis. Love transcends all? We
must progress, or perhaps not? Society is a living organism that can't be
perfected? There's a moral dilemma whereby the player can choose to continue
Athena's work or alternatively turn her machines off, but it's not much of a
dilemma. I can't imagine anybody choosing to turn the machines off, except as
a way of seeing the different endings. Arthur C Clarke's
Childhood's End presented humanity's transcendence in a much more
ambivalent way, while in contrast
Talos 2 doesn't show a downside.
The plot is essentially about a character who has a crisis of faith, but
gets over it, with some filler involving the search for a missing person,
and a diversion in which we learn about one character's personal tragedy.
The writing and voice acting is fine - Erin Fitzgerald is back, in
essentially the same role - and I enjoyed hearing voice logs from a robot
who was obviously modelled on Werner Herzog, but it feels
ordinary compared to the original game. It's a bit like the
difference between
FAR: Lone Sails
and
its sequel. The original had a quiet power, the sequel is a solid game that feels
plain in comparison.
I was scaroused by the sphinx
And it sounds great as well. Damjan "The Mran" Mravunac is also back, with
a soundtrack that has some of the same issues as the game itself - it's
bigger, longer, but less focused. In the original game the music looped
through a small number of memorable themes, whereas Talos 2's music
is diluted with a lot of pleasant atmosphere. Talos 2 has
a dynamic music engine. I think there's an underlying pattern, but it felt
as if the game was just cycling randomly through the music. On the
positive side the power ballad at the end of the game manages to make a
set of lyrics about robots putting boxes on top of pressure plates etc
sound meaningful and not silly, which can't have been easy.
Is there any outright bad stuff? By the end of the game I felt that it had
gone on too long. The huge arenas are interesting at first, but there's an
underlying pattern, and in the later maps I just ran everywhere ignoring
everything. Talos 2 would have benefited from having the same
puzzles packed into two or three fewer arenas. A couple of the bonus
puzzles are needlessly aggravating, particularly one in which you have the
scour the entire map a second time to find a tiny cubby-hole that has an
extra puzzle element.
There's a plot strand from the beginning of the game that doesn't seem to go
anywhere, although it's essential if you want all of the Steam achievements.
Performance is, as mentioned, choppy on modest hardware, and the game is
occasionally odd-looking even if you have a supercomputer. It's a shame that
the Serious Engine now appears to be defunct.
One oddity is that water doesn't ripple. EDIT: but between
writing this post and publishing it a patch fixed the water.
It took me 46 hours to finish the heck out of the game, and at the price of
£19.99 (it was on sale) I easily got my money's worth. Talos 2 is
bigger, easier, move diverse, but at the same time more prosaic than the
original, and I enjoyed it, but it also feels bloatier and less characterful.
It ends with scope for a third game, and it'll be interesting to see how the
series progresses; in particular whether the next game does away with the
puzzles entirely in favour of a story-driven role-playing game.
Could the world of Talos come true? The thought of humanity being wiped
out, leaving behind a bunch of self-sustaining server farms churning away at a
self-generating AI is of course still very unlikely, because the world is at
peace and wiser minds are sure to prevail. And of course it wouldn't work at
all, because we lag far behind the world of Talos when it comes to
robotic engineering. But imagine another world, of chaos and war. Imagine
that.
