How To Build The Universe in a Computer

Thank you to Wren for supporting PBS. To learn more, go to wren.co/start/spacetime.
Take the Space Time Fan Survey Here: forms.gle/wS4bj9o3rvyhfKzUA
PBS Member Stations rely on viewers like you. To support your local station, go to:to.pbs.org/DonateSPACE
*****
Sign Up on Patreon to get access to the Space Time Discord!
/ pbsspacetime
We routinely simulate the universe on all of its scales, from planets to large fractions of the cosmos. Today we’re going to see how it’s possible to build a universe in a computer - and see whether there’s a limit to what we can simulate.
Check out the Space Time Merch Store
www.pbsspacetime.com/shop
Sign up for the mailing list to get episode notifications and hear special announcements!
mailchi.mp/1a6eb8f2717d/space...
Hosted by Matt O'Dowd
Written by Matt O'Dowd
Post Production by Leonardo Scholzer, Yago Ballarini, Pedro Osinski, Adriano Leal & Stephanie Faria
GFX Visualizations: Ajay Manuel
Directed by Andrew Kornhaber
Assistant Producer: Setare Gholipour
Executive Producers: Eric Brown & Andrew Kornhaber
Executives in Charge (PBS): Adam Dylewski, Maribel Lopez
Director of Programming (PBS): Gabrielle Ewing
Spacetime is produced by Kornhaber Brown for PBS Digital Studios.
This program is produced by Kornhaber Brown, which is solely responsible for its content.
© 2021 PBS. All rights reserved.
End Credits Music by J.R.S. Schattenberg: / multidroideka
Special Thanks to Our Patreon Supporters
BIg Bang
David Taiclet
Ben Dimock
Daniel Alexiuc
Nenado763
Pravin Mansukhani
Peter Barrett
Nils Anderson
David Neumann
Charlie
Leo Koguan
Sandy Wu
Matthew Miller
Ahmad Jodeh
Alexander Tamas
Morgan Hough
Juan Benet
Vinnie Falco
Fabrice Eap
Mark Rosenthal
David Nicklas
Henry Van Styn
Quasar Supporters
Alex Kinsey
Alex Kern
Ethan Cohen
Stephen Wilcox
Yogi B
Christina Oegren
Mark Heising
Hank S
Hypernova Supporters
william bryan
drollere
Joe Moreira
Marc Armstrong
Scott Gorlick
Nick Berard
Paul Stehr-Green
Alexander Price
MuON Marketing
Russell Pope
Ben Delo
Nicholas Newlin
Scott Gray
Антон Кочков
John R. Slavik
Mathew
Donal Botkin
John Pollock
Edmund Fokschaner
Joseph Salomone
chuck zegar
Jordan Young
m0nk
Daniel Muzquiz
Gamma Ray Burst Supporters
Lillith Montgomery
Avi Yashchin
MHL SHS
Kory Kirk
Terje Vold
Anatoliy Nagornyy
comboy
Brett Baker
Jeremy Soller
Jonathan Conerly
Andre Stechert
Ross Bohner
Paul Wood
Kent Durham
jim bartosh
Nubble
Chris Navrides
Scott R Calkins
The Mad Mechanic
Ellis Hall
John H. Austin, Jr.
Diana S
Ben Campbell
Lawrence Tholl, DVM
Faraz Khan
Almog Cohen
Alex Edwards
Ádám Kettinger
MD3
Endre Pech
Daniel Jennings
Cameron Sampson
Pratik Mukherjee
Geoffrey Clarion
Nate
Darren Duncan
Russ Creech
Jeremy Reed
Eric Webster
David Johnston
J. King
Michael Barton
Christopher Barron
James Ramsey
Justin Jermyn
Mr T
Andrew Mann
Isaac Suttell
Devon Rosenthal
Oliver Flanagan
Bleys Goodson
Robert Walter
Bruce B
Ismael Montecel
Simon Oliphant
Mirik Gogri
Mark Delagasse
Mark Daniel Cohen
Brandon Lattin
Nickolas Andrew Freeman
Shane Calimlim
Tybie Fitzhugh
Robert Ilardi
Eric Kiebler
Craig Stonaha
Martin Skans
The Art of Sin
Graydon Goss
Frederic Simon
Tonyface
John Robinson
A G
David Neal
Kevin Lee
justahat
John Funai
Cass Costello
Tristan
Bradley Jenkins
Kyle Hofer
Daniel Stříbrný
Luaan
AlecZero
Vlad Shipulin
Cody
Malte Ubl
King Zeckendorff
Nick Virtue
Scott Gossett
Dan Warren
Patrick Sutton
John Griffith
Daniel Lyons
DFaulk
GrowingViolet
Kevin Warne
Andreas Nautsch

[@dcsignal5241]

There's nothing more impressive than a well thought out experiment. Light bulbs? Genius.

[@davidwilliams5497]

You might say it was a *bright* idea. 😉

[@robbowman8770]

Analog supercomputer 😁

[@birbdad1842]

@@davidwilliams5497 Enlightened way of thinking. You definitely shine in it 😅

[@cbuchner1]

How did he compute the effective force acting on each lightbulb per time step? Doesn’t that require the derivative of the gravitational potential function - and how would that be acquired?

[@robbowman8770]

@@cbuchner1 Does the light intensity measured give you something proportional to the sum of all the individual GMm/r^2 ?

[@danieljrea]

As a person with a PhD in computer science, I can say there is finally an episode of Space Time I understand (almost) completely.

[@seanmccrd]

It gave me new appreciation for how well Matt et al. conveyed the scale of the problem and the tricks used by researchers to get runtimes within human lifetimes. Approximation algorithms are amazing.

[@rowill2968]

I have no idea what you're talking about 🤣

[@tuneboyz5634]

@@rowill2968 its ok lil buddy :)

[@blackrack2008]

I understand this without phd. Your degree is overrated.

[@runnersclub3685]

As a person who passed ASTRO 101 I can say there is once again an episode of Space Time I (almost) can begin to understand

[@Mercurius314]

Correction: the room-sized computer used to calculate the Apollo trajectories did not have about the same computing power as your smart phone. The IBM 7090 which was used by NASA at the time among others ran around 100 kflops. The GPU of a modern smartphone ranges in the hundreds of gflop up into the tflop range. While there is some debate to be had about sustained load performance, your smartphone outclasses anything NASA had during the Apollo program by at least 5 orders of magnitude.

[@PendragonDaGreat]

Yeah. I remember back in physics 102 my professor said something along the lines of "If it had the ability to access external RAM your Texas Instruments graphing calculators could run multiple concurrent Apollo missions and probably still have clock cycles left over to play snake." After that class I went and looked and figured that yeah, it probably could.

[@HeavyMetalGamingHD]

Yeah, I immediately doubted that. Computer performance had exponential growth. No chance, that they were anywhere near that in the 60s

[@erik.dahlberg]

Yeah, I had my doubts about that statement as well. So in terms of old "smart phones" and the Apollo trajectory, what would be the oldest phone that would still outclass the 60ths tech? Are we talking Nokia N71, or perhaps Sony Ericsson P8, or even older Ericsson 688H!?

[@HeavyMetalGamingHD]

@@erik.dahlberg Hard to find the data how strong the cpus in those phones were. I am not an expert, but my guess would be, that the phones in 90s were probably stronger than the computers in the 60s. 5 orders of magnitude means 5 zeroes. or times 10^5. I really can't find any proper data to the cpus used in the phones you suggested.

[@andrewharpin6749]

Modern high end smartphone GPUs are around 2.5 teraflops, they're more powerful than a supercomputer from 1999, let alone 70 years ago.

[@Cyberspine]

Having programmed gravitational simulations before, the thought that someone could achieve a simulation of that scale with nothing but light bulbs, a luminometer and calculations by hand is mind-boggling to me.

[@Christopher_S]

A verified account, called "Z" has stolen your comment.

[@Cyberspine]

@@Christopher_S As they say, imitation is the sincerest form of flattery.

[@bipolarminddroppings]

Who would expect that a dude with some lightbulbs and a notepad could correctly simulate a galaxy merger? Thats the power of science and mathematics, I guess.

[@andreubotella6464]

The power of inverse-square laws too.

[@chriseffpunkt4333]

@@andreubotella6464 they make things so much easier, especially when teaching student analogies between the topics

[@PMA65537]

I don't get how he measured the light from the dark matter bulbs.

[@jamescollier3]

assumptions and rounding still provide great estimates

[@l1mbo69]

@@PMA65537 why not?

[@IanGrams]

Fascinating stuff. That light bulb experiment was wonderfully clever but it's pretty cool to see how the development of computers has given us the ability to simulate so much more and with higher fidelity. Props to the people who render that simulation data to make such pretty videos, too. Also dang it's already been 3 years since the Computing a Universe Simulation video. I guess spacetime flies when you're having fun learning :]

[@jersey-dude]

I remember, when I was young, my mother read me a book about how someday, the sun will die. I was devastated. I just had that same dreadful feeling watching the formation of the eventual milkdromida galaxy. Always enjoy your content PBS space time. Good work.

[@mark9294]

Yeah, but the good thing is, this is pretty much inconceivably far out in the future

[@jamessunseri346]

This was my favorite video over the past 7 years of following this awesome channel. This is the exact subject I’m going to do during my PhD in astrophysics when I go to grad school next year!

[@PunmasterSTP]

That's really exciting, and I hope that grad school has amazing things in store for you!

[@FlameWaffle]

Come up with a new one

[@enriquegarciacota3914]

Enjoy!

[@Vetrivel.Shanmugam]

How is it going on?

[@xw591]

​@@Vetrivel.Shanmugambro went to princeton, hes probably strrssed but winninh

[@fyang1429]

Simulations of biological molecules have the same issue. Normally a long simulation of a protein (without full quantum mechanical details) would be around 1 microsecond, which is unfortunately short since some of the proteins have turn over time (the time for one reaction to complete) in scales of seconds. At the current state, it is quite ridiculous even to think about a simulation of a mitochondrion with quantum mechanical details, let alone the universe.

[@Mandragara]

Can you just discount hydrogens quantum fuckery in those MD sims?

[@KalebPeters99]

I wonder if including quantum properties would actually have an effect? I know there's a burgeoning field of "quantum biology", and apparently photosynthesis has some weird quantum shenanigans going on. Who knows what proteins might be doing on that level 😯

[@normangeist890]

@@KalebPeters99 At least in a sense of having electrons relocating according to their environment to allow for chemistry to happen or at least allow proper charge polarisation.

[@thealexanderbond]

Consider the fact that if you did program a universe like ours, you would certainly make it a universe where it's impossible to simulate a universe, exactly so you don't run into the infinite loop problem. You're making the mistake of picturing the programmers of this universe living in a universe exactly like the one they created. Almost certainly not.

[@tybaltmercutio]

@@KalebPeters99 It does, depending on your system and the properties you are interested in. These problems not only arise in proteins, i.e. biological systems, but also everywhere in quantum chemistry and molecular dynamics (and QM/MM). Finding ways to make accurate computations faster is a very active area of research.

[@cadmean-reader]

"My computer isn't powerful enough, so why not just represent masses collapsed below a certain radius with a nil pointer" sounds exactly like what a programmer for a universe simulation would think

[@l88ch3r]

Love the Aperture Science background. I've learned so much from you and this channel. I wish I could support with more than my eyeballs and a subscription. Please keep the good stuff coming!

[@PunmasterSTP]

I know! And it took me awhile to notice that it was from Aperture Science.

[@ptousig]

The Cake is a Lie!

[@Dudleymiddleton]

I still play it! Great puzzles in the community test chambers, there are literally thousands upon thousands of them in the workshop browser!

[@PunmasterSTP]

@@Dudleymiddleton I heard that with all the community-made chambers, there’s essentially a Portal 3. I haven’t checked into it too much yet, but would you agree with that sentiment?

[@Dudleymiddleton]

@@PunmasterSTP Yes I would. Instead of eagerly anticipating and waiting for "Portal3" and "Half life3" for that matter, the source engine code was released - unintentionally, I believe, and modding began, using the hammer world editor and other apps like Beemod to create almost endless maps and mods out there, so valve didn't really have the need to release another episode or version of the games. That's what I think, anyway - I may be completely wrong! The good thing is, they are mostly free to play as long as you have the original games installed.

[@MaryAnnNytowl]

Finding ways to experiment and test things we can't step outside of to see as a whole is a fascinating subject - very much so! Thanks, Matt & PBS Spacetime, for being a bright star in my otherwise very dark world. I appreciate it more than you can know.

[@sebastienpaquin4586]

The Problem: How to simulate the universe inside something much smaller than the universe you are trying to simulate. How physic answers: Assume the universe is a sphere. And ignore atmospheric drag.

[@Slayer666th]

Sounds like just implementing a force you call „dark matter“ and not caring enough to program in, what it actually is and does.

[@Vvopat96]

by optimizing the universe, just like how we optimize games, when it's not important, it doesn't exist and it's calculated as sum of group, like he explains in the video :) It's probably also how universe works, when electrons are not calculated they don't exist and they are just probability wave in space. Looks like that anyway that universe optimizes itself all the time. When cat is in the box and nobody reacts to it, it doesn't exist and only the box is simulated.

[@Josh-iv2bw]

The engineer asks why it has to be a sphere and not a point like object.

[@kaseyboles30]

It has been mathematically shown in several ways you don't need the entire universe to hold all the information about it. For a given amount of 'stuff' (matter and energy) you could encode all the information about that stuff on the 2d surface of a sphere equal to the Schwarzschild radius off a black hole created from all that suff.

[@martinhorvath4117]

@@Josh-iv2bw wow i literally spent like half an hour writing a reply and it just got deleted.... tldw, the universe is not a conventional sphere, it's more like a doughnut, since space itself is expanding at the same rate at every point in the Universe. but it isn't like a doughnut at all, we'd just simply could describe it that way. 1st dimension is space itself 2nd dimension is time 3rd dimension is space time continuum (how the first 2 dimension build upon each other to create our perceived reality and the 4th dimension is where every point in space is at every possible point in time. something that is unfanthomable to our mind... and.. any kind of mind really, since the 4th dimension is either the Singularity or the literal end of this Universe. (the flow of time getting reversed, and the Universe playing itself "back" in some weird non-scientific way.

[@RME76048]

Gotta love analog simulations. The lightbulb scheme sort of reminds me of the salesman's shortest path problem and how it can be solved using appropriate lengths of string and two quick moves.

[@RME76048]

@@hyperduality2838 and exactly what strain of weed were you indulging in?

[@JeffMoody]

I'm going to get groans, and I hate to perpetuate the Dyson sphere conversation, but I keep waiting for someone to address the ridiculous amount of material needed to even build the first ring. They would have to pilfer material from all surrounding solar systems, which would leave a noticeable void of activity around neighboring stars. I'm not seeing it as a much more than fascinating imagination fodder... unless they built wormholes to rob the rest rest of the universe of its riches! Buahahaha! But they'd need a Dyson sphere...?

[@chbrules]

Nice to seem some computer science and Big-O notation in a physics vid.

[@bluemamba5317]

Big O would be a great nickname for a fat scientist that also likes to rap.

[@tybaltmercutio]

Nowadays, there is a lot of research in many/all fields of physics to make good approximations which don‘t sacrifice accuracy but have a much lower computational complexity. Actually, I have always wondered why not more channels talk about that since this can also be a pretty exciting topic if presented like here.

[@egg-iu3fe]

computer science has it's hands in literally every field out there. There is very little that doesn't involve computer science in some way.

[@Hailfire08]

Absolutely love this video, I've been getting into simulations recently so I am very hyped by this

[@venil82]

Big O notation, explained on universe scale! Epic.

[@ModernDayGeeks]

Having the Aperture Science as background truly tells me that we'll be doing some real science here. Glad I stayed and learned a lot, awesome!

[@lorekun]

6:17 This method reminds me of video compression algorithms like x264/265. The way they mostly focus on the changing pixels between frames and not the whole picture, allowing them to compress an huge file into a small video while maintaining high quality. GIFs are always way bigger than mp4 files for that reason.

[@tybaltmercutio]

Actually, many recent algorithms in quantum chemistry use the same idea which is based on a singular value decomposition. It is then used to extract only the most important correlations between electrons and neglect others or to treat them on a lower level of theory.

[@Mirality]

Perhaps quantum uncertainty is just because the simulation model we live in doesn't simulate down to that level, so most of the time it just generates a random statistical answer. It's only when you try to actively measure it that the simulation resolution locally increases and it gives a definitive result.

[@EverettWilson]

There's a sci-fi short story about an AI that effectively abused a simulated population by having them live in a universe with incomplete laws of physics.

[@AkaliWolf]

@@EverettWilson what I need to read this

[@EverettWilson]

@@AkaliWolf I can't remember the title, unfortunately. I think I read it in one of the annual sci-fi short story collections.

[@AmyDentata]

If we were in a simulation, the universe running that simulation would use different laws of physics-laws capable of simulating a universe with the complexity of ours. We couldn't reverse-engineer the laws of that universe any more than a Minecraft villager could divine quantum mechanics. The Minecraft villager couldn't even reverse-engineer the code running its world; all it could do is observe the results of that code. The villager will never know about the code, the computer it runs on, the person operating that computer, or the laws of physics that define that person's universe.

[@kaseyboles30]

What you want to do is look at what the symptoms of a simulation are. Resolution limits for one and the rounding errors when near said limit (QM anyone) as well as perhaps rules that don't always clearly connect to other rules and seem arbitrary in each other's context. GR vs QM, dark energy and dark matter, etc. could fit that bill. Though such a rule inter-inconsistency speaks more of a hack to get 'good enough' results out of the simulation. Also while some might run a simulation of different physical laws, there are reasons to run simulation that mirror the simulators physical laws, perhaps with minor variations or differing initial conditions, to learn more about one's own reality.

[@davidp.7620]

@@kaseyboles30 Those are the rules of what a simulation would look like in our unvierse. Simulations in universes with different laws of physics may look completely different.

[@douglasharley2440]

none of your conclusions follow from facts, they are pure speculation.

[@james10pl]

I think this is a very good point. In this scenario, science is actually the process of reverse-engineering the rules that govern the game/simulation. But that doesn't mean that we couldn't imagine additional rules on top of that, which exist in the higher reality and allow them to simulate something as complex as our own reality.

[@Josh-iv2bw]

I think this is only sorta true. We could logic out the requirements of a universe needed to simulate our own, and deduce quite a bit of how that framework could behave mathematically. Visualizing it may be hard to impossible, but mathing it out? We technically do that with string theory right now.

[@AdrianBoyko]

In nested simulations, the “hosting” universe that runs the simulation could be much more computationally capable than the universe it simulates. But, as you drill into the simulation stack the simulated universes become less computationally capable. At the bottom of the stack you eventually arrive at a universe that can’t simulate an interesting universe. Maybe we live at the bottom of the stack!

[@Cythil]

Naa.... We are not at the bottom. Sim Earth and The Sims are still fun and interesting simulations. ;)

[@jamesdriscoll_tmp1515]

The grass is always greener on the high side of the gradient

[@AdrianBoyko]

@@Cythil True, but The Sims are not “interesting” in the sense that they will eventually program simulations that run in their own universe. If we can’t create a simulated universe that can create a simulated universe then we are the bottom of the stack.

[@lucasnoetzold]

the bottom of the stack are our simulations, then

[@spiritual84]

@@AdrianBoyko If we're truly at the bottom of the stack, then chances are that we're at the top of the stack as well.

[@k.ctandy5255]

This is the best channel that has ever existed. I've been watching for 5 years, and it never gets old.

[@p5rawQ]

There is a core problem with the simulation downward hypophysis. That experiment with the light bulbs shows this very nicely. In order to work, it uses a physical effect (light attenuation) to calculate the next step. And this, using a physical process within the world to simulate something, will always be the case! Even so you go digital, the information is stored and processed based on physical laws provided by the universe you live in. If you try to create a simulation where some intelligence can build a simulation them self, you need to simulate also the way the simulation itself stores and computes this. But there is always a downgrade otherwise the amount of data will overflow your system (outnumber the particle count available in your universe). Another argument is often: “Well 'they' only have to simulate our perception.” As a game developer and programmer I can tell you that you need to simulate much more than our 'perception' to create that universe we have. In a video game you have to make crazy short cuts and assumptions that are highly optimized to the limits of what the player is actually able to do (and reduce the amount of freedom). It doesn’t matter how advanced your tech is, as I said above, there is always a downgrade! What you can't do in a video game for example is doing real science. You can't grab a microscope or a telescope or build a particle accelerator (if this is not part of the game). If you want to build a simulation that would foresee all of those curious avatars, of which you might have no idea what they come up with … to challenge the universe you've created … with more and more detailed questions that still doesn't contradict. Moreover those discoveries will actually form a world where you can use laws to predict outcomes, to find an even deeper understanding. This would be something you would not allow to happen as a simulation coder, because your simulation would break at some point … always! If you give your avatars too much freedom you can't get away with only simulating the proximity because science does not only look at what is close and easy. It will ask for the cause of life, matter, atoms, stars, black holes and the beginning of the universe ... if you do not simulate all of it at once you will get contradictions … fast.

[@p5rawQ]

The question is: What is the difference between a simulation and the real thing? In nature a process unfolds because of intrinsic properties and laws of the universe. Light behaves the way it does because that is how photons interact with electrons (it is not a calculation!). In a simulation you need a model and a medium. The medium is something where you have a good control over its states and behaviors to encode something onto it, but itself has most likely nothing to do with the thing you want to simulate. For example ‘arranged’ Silicium. It should be obvious that you use properties of the medium and physical (or chemical) laws to encode something else and in this process you always need some surplus on material and energy (therefore the downgrade). But even more important is the model. This will determine the rules and update cycle of your simulation. But a model always has a perspective, something with an intention. You want to simulate a specific process, behavior and so on. Therefore the model can’t simulate everything, even so you might be so arrogant to want to, because it is just a model and not reality. The updates can’t be infinitely small, the calculations and algorithms have to avoid NP complexity. And there is always more, more as we see, more of what we know, more of what we experience (but maybe still causal in the process). The theory of everything (world formula) will not help you here, because it will not create life nor consciousness, it is “just” an abstraction of the physical processes (would be still a huge achievement). In reality emergence will create new properties, which will create new domains, which will have their own laws.

[@Georgije2]

Contradictions such as "Are photons particles or waves"? To me the whole of quantum physics just looks like it was hacked together in a hurry when the universe's creators noticed that some characters were snooping around details that weren't supposed to be snooped around.

[@p5rawQ]

@@Georgije2 I know what you mean, but in spite of the fact that we have difficulty to create a mental picture of it, quantum mechanics can be calculated very precisely. Lots of its laws are already known and do not contradict. We are used to the macroscopic world and first we thought that this was also how the world of atoms and particles looked like … spheres on trajectories like planets … but it turned out to be naive, with a strong bias to what we used to see. Translating quantum mechanics to the macroscopic room is already a model, which has an intention, meaning it explains one thing but not another. Today I actually think the opposite is true, if the world of Atoms would look like spheres on trajectories and so on (I would probably code a simulation like that), it would create contradictions (which were actually already been seen in the beginning state of quantum mechanics). That it is not like that, makes much more sense in the long run.

[@p5rawQ]

@@Georgije2 And also, with your assumption you’re implying that there are creators that were able to create a simulation with all this complexity we know of, with life and consciousness beings that do science … but they could not foresee that we would look into the microcosm? Even so they gave us the freedom to do so?

[@doodelay]

As a student of theoretical physics, would like to hear a lot more about the clever tricks of computational and experimentalists

[@mitchellwilley7208]

Here's 1 to blow ur mind. A great way to run a simulation of our universe and save massively on computing power would be to make a 2½D + depth universe. This would also allow us to enter the universe via virtual reality.

[@Farmfield]

As someone who works a lot with simulations, my approach to simulate a whole universe would be simplifying stuff based on proximity to any intelligence evolving in it - thus a black hole 100 light years away would only be an approximation while stuff close by would very detailed. And if any of the evolved intelligences tried to figure stuff out for real, I'd create a layer as failsafe, where looking too close would yield a glimpse into a fundamental framework so weird it would almost be impossible to figure out. 😂

[@MarlonMuthiani]

The Architect decides to share his wisdom and we all think he is mad man on a street corner because of the limited computational resources he has as to not simulate one light year tall humanoid to convince us!🤣

[@Slayer666th]

I would impliment something to prevent the evolved intelligences from figuring stuff out and therefore needing too much details by something like the uncertainty relation. And to be real, dark matter basically sounds like someone being too lazy to actually program in the stuff that causes the gravitational effects, lol.

[@David95111]

Being truly lazy my solution would be to build in a trigger for such circumstances which would send some huge asteroid, or a big solar flare or whatever towards earth. That’ll keep them busy. But as long as humans are just looking from afar, and not physically going there, it should be doable to relatively low res simulate things

[@szamszatan]

That is the DMT experience

[@alant.6820]

@@hyperduality2838 pass the joint bro

[@dabeste6163]

0:50 Excuse my question, but how can we simulate "Dark Matter Particles" in a multigalactic scale, if we don't know their mass, their properties or if they even exist?

[@diphd549]

You're right we don't know what exactly they are! But that doesn't stop us from modeling dark matter type effects coming from 10^9 solar mass type particles. It's definitely not perfect, but can be a simple approximation for galactic scale processes, considering our galaxy has a mass of ~10^11 solar masses, so maybe a few hundred of these 'particles'.

[@gatocochino5594]

It's a simulation of the distribution of mass and gravitational effects similar to what we observe in real life galaxies.

[@Rauschgenerator]

@@gatocochino5594 That would be a good explaination, if the behaviour of DM wasn't such contradictory, because in some cases it doesn't behave like a particle but more following the MOND predictions, which themselves fail in many cases to explain what exactly is going on.

[@MrSigmaSharp]

Great topic I really enjoyed it. The examples that you provided were all particle simulations with little information per particle. Are there more detailed simulations maybe with less particles but more information per particle? Is it even possible?

[@tybaltmercutio]

What do you mean by „information per particle“?

[@MrSigmaSharp]

@@tybaltmercutio current simulations track position and velocity of each particle. Other properties would add to the detail of the simulation for example angular velocity or collision properties for each particle. This is what I mean by information per particle

[@tybaltmercutio]

@@MrSigmaSharp Okay, thank you, got it. I was thinking in terms of ab initio quantum simulations where people usually use the analytical gradient of the energy instead of a parametrized force field to get the forces.

[@alivateRocket]

"Which is kind of what the universe is anyway." (quantum interactions of the real thing) - Thank you!! This is what I keep thinking to myself. The universe is a physical system of quantum particles interacting. It would take more energy and space to simulate than it would to have the real thing.

[@bluemamba5317]

You don't have to simulate every particle in the whole universe all the time. Just render the ones that are currently observed by a conscious entity within the system. Like we "simulate" worlds in computer games.

[@DundG]

@@bluemamba5317 That is partly wrong because our universe seems highly predictable on large scale. Meaning even if no one is looking you still need to simulate movement of bodies or a function capable of predicting the outcome with the right input, if the system is not to complicated. Movement of planets are relatively simple if they are highly periodical, but many things are ridiculously complex and follow the physics, like the surface of a planet, meteors, supernovae, everything way simpler to simulate than to create an ridiciously, unimaginable large function.

[@bluemamba5317]

​@@DundG Yeah, kinda hard to imagine but think there are ways around that too. Don't really need "real" physics to change things when no one sees or records it for a long time. You could have a large number "pre-computed" outcomes. Or auto generated random landscape changes over time depending on some variables, such as density of asteroids in the area, planet composition etc.

[@AanandBajaj]

@@bluemamba5317 There would be discrepancies in that though and we don't see any, and I don't see how they would define a conscious entity "observing" a particular part of spacetime.

[@bluemamba5317]

@@AanandBajaj Probably, the system's highest priority would be not to reveal to the inhabitants that they reside within an artificial matrix system. Just think of an open world computer game, it will always know where the player is and what it sees, only loading the parts that are currently relevant. The rest of the information lies compressed and dormant, waiting to be triggered.

[@INGIE32]

7:05 I just had a course about the (discrete) Fourier transform, and I tried to make a python function that did the DFT, but this calculation took for ever (N was roughly 300k, from the Nokia tune). So I look online and it turned out that since it is a too long calculation (also N^2), usually algorithms use the FFT, which also only has N log(N) computations. Are these two related?

[@INGIE32]

@Fk Yu Yes that’s what I’m saying, but I’m asking whether these two things (the part in the video and going from DFT to FFT) are related, since they both have the same speed increase

[@tybaltmercutio]

@@INGIE32 I guess, then you could also ask whether a bubble sort and a merge sort algorithm are also related to this since they also go from O(N^2) to O(N log(N)). ^^

[@INGIE32]

@@tybaltmercutio I am asking…

[@tybaltmercutio]

@@INGIE32 I‘m not an astronomer but a quantum physicist. Hence, my answer may not be 100 % correct. But I think there is no direct relation since FFT is not an approximation to DFT. The main idea of that Barnes-Hut tree approximation is to coarse grain the system of interest to save computational resources by not having to compute all individual interactions for large distances. Basically a way to reduce the number of terms to be computed if two parts are distant from each other. Similar approaches are also used molecular dynamics simulations. Hence, it is approximation to the respective parent method.

[@renderproductions1032]

5:44 clippy!

[@jypsridic]

We might not be able to simulate the actual universe, but we could simulate enough of it to be perfectly convincing to the things living inside of the simulation. The simulation hypothesis is just another version of the brain in a jar though. I'm sitting alone in my room with the door closed and the blinds drawn, if the simulation exists the only thing it would need to be rendering right now is this 100sqft or so of bedroom. That's easily within near future technology.

[@jypsridic]

@@hyperduality2838 verbal diarrhea is very bad at conveying a message. Be clear and concise. I can't tell whether you were supporting or criticizing me, or even if you're just a crazy person screaming into the void looking for an echo.

[@nomizomichani]

For tree code, instead dividing by the volume, you could divide by distance from the observer with varying time steps. Further away the slow time is. It could simulate expanding of universe and save the processing time. With some tweak, you might even able to simulate relativity.

[@kaseyboles30]

Yep, use statistical methods to reduce complexity away from the central portion of the simulation in direct proportion to their impact on the 'core'. This likely works well both with distance and with approaching resolution limits (QM anyone?). Be interesting if DM and DE started as cumulative rounding errors and got left in as 'interesting' to our simulators.

[@johnransom9983]

When you said "tree code" I think you meant quad-tree (2D space) or oct-tree (3D space). These are used in physics simulation code to resolve the problem you are describing and other similar problems. I am a coder with a little bit of experience with physics simulation in games.

[@halfbit]

Does an image of the milky-way take time into account? Being that the milky-way is thousands of light years across, wouldn’t a spiral just be an effect of delay? We are viewing how it looks now but because of its size the image would be distorted in a gradual arc. I would think the actual shape of our galaxy is not visible with out compensating for time.

[@epicsmashman6806]

People have done the calculations for accounting for the time differential of viewing our own and other galaxies and it is negligible. Edit: it’s because stars take hundreds of millions of years to orbit the galaxy, but light takes max ~75,000 years to reach us

[@AverageAlien]

no because we can see other galaxies and the spiral arms can be predicted using our calculations of gravity.

[@AverageAlien]

we can't see any spiral from inside our galaxy. At best we can see the centre and one or two of the arms.

[@cortster12]

Astronomers are smart, they've taken all this into account.

[@epicsmashman6806]

@@ericlaska4748 we define it to be time when moving at zero speed relative to the CMB.

[@jasongg7327]

When we simulate our universe does it make new universes?

[@jimmygeeraets9039]

If you program something you get what you put in the program.

[@KatOliYT]

My thinking is yes, but is it as real?

[@xNMM]

If we simulate our own universe exactly then we become "gods" of the simulated universe and begining the simulation is their big bang. Would be interesting to see what would need to happen to create a simulated universe because it could give us something to look for in ours to see if we too are simulated by "gods". Edit: if our universe is based upon infinite regression then I'm gonna be pissed

[@bolshevikY2K]

i got bad news chief -- the universe might be an infinite regression of simulations

[@relt_]

basically, ddlc+

[@xNMM]

@@bolshevikY2K :/

[@bolshevikY2K]

@@relt_ yes

[@Chris_128]

Start New Game+

[@KaktitsMartins]

The simulated universe with creatures evolving in it doesnt necessarily have to be the same complexity as the the universe hosting the simulation. Perhaps we are a simulation and the host universe is unthinkably more complex than what we have here.

[@westhouse4641]

Black holes are the recycle bin. The big bang a fresh install.

[@justarandomdood]

Isn't that addressed at 11:35

[@kivagalehall4965]

@@justarandomdood I believe what they are trying to say, is that the universe that is simulating ours is more complex. Like if we simulated a universe at the particle level with just one or two forces.

[@ZintomV1]

Martin, this is exactly what I was thinking! You don't need all the low level sim stuff just the "perceived effect" it has on other particles.

[@zenithparsec]

Perhaps it's much simpler: Perhaps the "host" universe is just the "world of forms", and the universe we exist in is just an inevitable result of the concept of universes existing.

[@thedownwardmachine]

I don’t care if I’m in a simulation, pain still hurts, and pleasure still feels good.

[@jooptablet1727]

Enjoy your steak.

[@drdca8263]

I, on the other hand, care but only a little bit / only conditionally. If this world is a simulation, then there's a possibility that the world outside the simulation could influence what happens within the simulation, making exceptions to existing laws, or possibly ending the simulation. If we knew that this was a simulation, I think it would be worthwhile trying to determine whether we had any method of having a non-negligible chance of influencing the simulating world in a way which we could observe. (So, either getting boons from the simulators, or possibly hacking their computer in order to make backups of our universe on other systems, or otherwise do things that further whatever goals we might have) But, I don't think we are in a simulation (unless something like "the mind of God" counts as "in a simulation".. but I don't think "the universe exists in the mind of God" is probably quite the right theology either, so, whatever). I agree with your overall point that whatever this is, this thing we call reality, it is real in any sense we have reason to care about, and whether or not it is a simulation does nothing to change that. (Though, I think I would give different examples other than pleasure and pain. I don't think those are a good foundations for meaning/value/whatever .)

[@bluemamba5317]

@@jooptablet1727 🤣

[@fkarg10]

What I've been wondering about for a while: Are we somehow a detectable anomaly if we were in a simulation? As in, could someone find us by looking for irregularities in ... entropy? energy? Or any other way? Would life as we know it leave noticeable disturbances in the simulation?

[@christianpetersen163]

There is a conspiracy theory that the numerous UFO incidents relating to nuclear technology can be explained as manifestations of the simulations' search mechanism picking up the signature of fission.

[@AverageAlien]

@@christianpetersen163 UFO "incidents" are all hoaxes or mass hysteria. The classic flying saucer shape was actually born because the media misinterpreted the first ever UFO sighting. Basically, 99.9999% of UFO sightings can be written off as frauds. The rest are all actually misidentified flying objects.....e.g. space debris, planes, ducks

[@CombinatorialKk]

@@AverageAlien that's what an alien would say

[@christianpetersen163]

​@@AverageAlien Yeah, that's why I said "conspiracy theory", and not "THIS IS 100% LEGIT"

[@ASLUHLUHCE]

'Life' isn't fundamentally different from 'non-life'.

[@_BlackSpectrum]

How can we simulate Dark Matter if we don't know what exactly it is and it's properties?

[@alwaysdisputin9930]

We do know some of dark matter's properties: 1) galaxies spin faster 2) space's bent by gravity near to it (which causes gravitational lensing eg ring shapes) 3) it doesn't emit EM waves. Is there a 4th property?

[@code-cave]

Dark matter interacts through gravity, so you can simulate its motion via Newton's laws without knowing what it is phenomenologically.

[@_BlackSpectrum]

@@code-cave 👍🏼

[@_BlackSpectrum]

@@alwaysdisputin9930 👍🏼

[@qzbnyv]

Hi Matt! If we create a simulation where sentient entities eventually ponder “Am I in a simulation?”, there is no rule that says we have to have made that simulation anywhere near as complex or as rich as our actual universe in the sense that the sentient entities of our simulation perceive. I haven’t come across any work that says that entities couldn’t become sentient while being constrained to a “physical” world no more complex than an advanced version of The Sims (or whatever). So aren’t we dismissing the simulation hypothesis too eagerly by making claims about how difficult it would be for someone to simulate our universe? If we’re inside the simulated universe, we have no idea at what complexity scale our simulation is relative to the scale of the parent ‘real’ universe. Anyway, love your work :)

[@NeonVisual]

Somewhere far beyond the stars there is a scientist making algorithms to compress the universe simulation, and just mashes classical and quantum physics together haphazardly because they know they don't need to simulate that much detail and just enters a few sets of probabilities to loosely join it together. Also has to add dark matter without any actual matter to avoid having to simulate every single particle. No one will ever know.

[@stephenspackman5573]

When you simulate a universe, in principle you only have to simulate that part of the result that you actually observe (this-like the hierarchical and frequency domain approaches you mention-is a common transformation in computer science, often called lazification, and it allows pointwise computations involving infinite objects, which are otherwise nonsensical). When you simulate a simulation, in principle you only have to simulate that part of the simulated simulation whose observations you observe. It's thus not _completely_ obvious that universes can't nest inside themselves-there are subtle fixed points involved. So long, that is, as we are sincere that we are doing empirical science, and we are willing to abandon these fairytales about (in principle unobservable) “reality”.

[@DundG]

Not if the simulation needs to be predictable. Simulations (in entertainment or videogames) work for us, because mostly we aim for belivability, not "realism". If we don't look into the sky, the crab nebulae still expands according to its physics, planets still move around and affect each other. You could tone down the accuracy, because we don't know all the details and still could accept the simulation as possible, but it all needs to be simulated at all time.

[@zenithparsec]

@@DundG Predictable to whom? What if the only ones you need to be convinced live inside the simulation? In that case, you could look at the simulation and turn the "simulation quality" knob down until it started being noticeable. If anyone notices, redo any affected calculations at higher quality. Because you're simulating their minds , you could just use "hacks" most of the time. And in the worst case, maybe you could reload from a save recalculate with the "quality slider" nudged a tick or two rightward.

[@DundG]

@@zenithparsec It is predictable to us, who would live in a simulation. Everything you just said is what I said in other words... Still needs to be simulated at all times as things happen, regardles if we look or not, and if we know that it happens or not. A glass wine left in a room still looses alcohol because it evapporates, doesn't matter if we know that it does or not, we notice that it did.

[@stephenspackman5573]

@@DundG No, I am not suggesting any reduction in fidelity, and certainly not of repeatability (which is arrived at trivially if the simulating computation itself is constructive-classical and synchronised). My point is that if we are sincere that science is the study of the observable, then, theoretically (though I am not at all implying that I know how to perform this reduction!), simulating the universe requires only simulating the inputs to a single physicist, and simulating a simulated universe requires only simulating only those parts of the simulated physicist that are read out from the simulated simulation. Even if this is done with 100% fidelity, it still appears likely to be a much smaller problem than time-stepping a model of an “everything” that includes multiple time-stepped models of several “everythings”, a programme that appears to be doomed to failure by a basic counting argument. From a philosophical perspective this amounts to saying that the solipsism question is, mathematically speaking, a separability question, and comes in degrees. Is the “actual solipsism” of the universe, the isolability of each observer, strong enough to exploit? As far as I know (though goodness knows this is not my field), this is an open question.

[@MC-wh3xm]

We don't need to simulate our entire universe, just what needs to be rendered for each observer.

[@juzoli]

But since every single particle acts as an observer, it is easier to generate one big universe, than separate smaller universes for each particles.

[@MC-wh3xm]

@@juzoli You misunderstand me. One big universe, with parts of it rendered at certain times. Not separate universes.

[@juzoli]

@@MC-wh3xm Those “separate” universes overlap each other. And to render the neighborhood of every single particle, every time they are making an interaction, is actually way more effort than just to render the whole big thing once. Because particles are everywhere, and they are doing interactions all the time, so you saved nothing.

[@leagueoflags]

As a programmer it's quite simple why universe simulations are a problem. The more variables one feeds to each object and the more intricate the interactions, the slower the overall calculations, and the slowdown is exponential in nature. Thus, models simulating billions of stars are either pretty crude or extremely taxing on the hardware.

[@polarisnorth4875]

Not if you set a limit on the exponential growth dependant on your hardware.

[@leagueoflags]

@@polarisnorth4875 Sure.. it's all a question of accuracy. The hard part of modelling isn't the coding.. It's coming up with approximations that are good enough and reasonable to calculate.

[@seraphik]

fastest I've ever clicked

[@shauryakalia3296]

So the all the uncertainty of quantum mechanics and inability to merge it with general relativity could be just because our universe's simulators are using shortcuts to save their system memory. Wonder what headaches we would give to the astrophysicists of the universes we simulate

[@brainwater]

Technically our galaxies colliding isn't as likely to happen as the sun rising tomorrow, since there's a non zero chance of vacuum decay destroying the universe in the next 100,000 years.

[@thebeeemill]

My rationale as to why we are not living in a simulation: we cannot simulate the entire universe down to the subatomic level because simulating one electron would require more than one electron to perform. Thus we have to simplify and use approximations. Because we have access to a limited amount of matter and energy within the visible universe whose limits we will never be able to breach, that means that any simulations we run will suffer the same fate, as will the simulations run within those simulations and so forth. This means that each “level” of simulation decreases in complexity, making infinite simulations from our universe impossible. If we were being simulated, what reason would our parent universe have to limit the amount of accessible energy and matter in ours unless their universe also suffers from the same issue? This itself doesn’t dispel the possibility that we are being simulated, but greatly decreases the probability to a level where I feel comfortable saying that the cost/benefit ratio of running finite limited simulations of our world’s complexity is low enough that we are likely not being simulated.

[@brainwater]

What if the simulation *is* the universe?

[@juzoli]

What if the computer they are using to simulate us is bigger than our universe? But it is still tiny compared to their universe? Anyway, this is not a scientific idea to begin with, as it is not falsifiable. We cannot disprove it, as whatever we find, can be just part of the simulation. But if we find an evidence for it, it can be just a weird new natural law of the universe we just don’t understand yet. This is not a bit better than other creationist stories.

[@birbeyboop]

The thing that makes it unfalsifiable is that there's no way to know for certain whether it's impossible, because what we see in real physics could indeed be optimized approximations of a universe that's strikingly similar but more complex at the most fundamental levels. We really have no way to know for certain. It certainly does make an infinite recursion of simulations seem pretty much impossible for sure, almost entirely discounting the overwhelming odds in favor of being in a simulation that would otherwise emerge.

[@juliasophical]

Indeed, simplifying and approximating would need to be done, and it would be noticeable to the inhabitants of a simulated universe if they looked closely enough. In an old game I used to work on, there was an issue with the simulated physics where moving objects would sometimes pass through walls instead of bouncing off of them, due to the simulation not getting enough physics ticks to register the collision. So in the simulated universe, you'd see effects like objects jumping from one side of a barrier to another from time to time. Also, because there are a limited number of bits available to store any value, if you looked too closely at small objects, their attributes would not vary continuously over ranges, but would appear to be quantized at a fundamental level, varying in discrete steps instead of continuously over a range. So instead of infinitely divisible values and continuous motion, you get tunneling and all kinds of weird effects when you try to look at things at the smallest level of detail, due to this simplification. Oh, wait...

[@thebeeemill]

@@juliasophical interesting point. I’d have to give it some thought to offer a rebuttal

[@breezyx976]

All arguments about our universe not being simulatable go out the window when you realize the ones simulating us might not be bound by our laws of reality...

[@erikziak1249]

The only downside of the "universe sized computer" is that it renders everything only in real time. Not faster. At the same time it cannot be stopped and reverted to a previous state. If you ask me, that is a pretty lame computer.

[@ApiolJoe]

Yeah, but at least it's also a time machine. Granted, a very lame time machine, allways going in the same direction at the same speed. Universe is crap, must be very outdated technology :D

[@gooblepls3985]

but that's just what *you* feel like is happening

[@nellyx8051]

For all we know it could be stopped, reverted, or sped up. We are observers within the universe so we couldn't tell. You can only tinker with a simulation you created not the one that you're in. If you ever find yourself in a pixelated world with almost no complexity you know your near the bottom of the nested simulation chain. (Or playing Minecraft) 😂

[@erikziak1249]

@@gooblepls3985 True. I am part of it. I am conscious now. Was I just 30 seconds ago? Are my memories real? Do I recognize the comment above, to which you reply as being "mine"? Yes. Do I answer that really factually correctly? No way to know or prove it. Right now I do not even remember the first sentence of this reply when I began to write it. I am not looking back. If I read it, i might remember. Or I will make up a story just to fit my narrative and put it into context, stored in my long-term memory (from which it will fade pretty fast). Are all our conscious lives not just mere illusions and we keep making up stories that fit the narrative, as to make any sense of this vast chaotic world, that we have to navigate with incomplete information? I mean, consciousness might very well, or it actually is, only a byproduct of our oversized brains. It was not selected for, it came into being and is being dragged on as an evolutionary spandrel, while what was selected for is our ability to analyze, conceive plans and, stay alive and reproduce. Or maybe we need that spandrel in order to survive? What if it is forming the backbone of our "lazy system" thinking? The fast one, giving quick and usually correct answers without even consciously thinking about them. How big is our conscious mind? 5% of its overall energy consumption? More? Less? Why? Am I overthinking it? Probably. Do you ask yourself whether I am sober? I am. I wish I had intoxicated myself a little with drinking alcohol. It is relief sometimes. Why does my head spin now? It feel heavy. I should go to sleep. Good night all.

[@das_it_mane]

Wish I could simulate a girlfriend

[@gandalf8216]

If possible, it becomes nigh impossible for our Universe to not be a simulation already. The probability for our Universe to be a simulation would increase towards infinity.

[@kronosol5779]

The second half of the Video , where the Simulations show starts with the nice music is just Superb :)

[@riccfire]

One of my favorites videos from you guys to date. Really awesome!

[@mihaiandreioprea]

you guys do such a great job

[@discreet_boson]

Loved these past 2 videos on simulations! Would love to see more such videos of the meeting point of physics and computation

[@gmailcal]

The Voice to Background music is just perfect on this one. Thanks

[@MrMrbloobloo]

It’s really nice having these videos showing how the simulations usually shown to us are actually made. I’d personally like to see a similar video about how the CMB is measured.

[@esraeloh8681]

I have literally been thinking about this question since just after the new year, &, here you guys are, discussing exactly this subject, talk about coincidense.

[@antonbatura8385]

I've been watching SpaceTime for a few years, and I won't lie, it's always nice when there's an episode which I can kinda understand.

[@ericvosselmans5657]

another amazing episode. they keep on getting better and better

[@stephencashen1199]

These computational physics videos are awesome.

[@stephanieparker1250]

Thanks for another great video, Matt! 🙌

[@douglasharley2440]

*computational nitty-gritty!* love it, much thanks.

[@btc54723]

Never thought I'd see algorithm time complexities discussed in a spacetime video. I love it

[@neogoo123]

Lol, loved the Aperture Science reference! Gotta love that fuzzy and friendly supercomputer GLaDOS

[@TheNathanSproul]

Loved this one!

[@infonityarchive]

excellent video as always

[@Neo-br3uc]

The Light Bulb Idea is a one well-thought and creative Experiment for sure

[@4RafaCaetano]

this video is simply stunning

[@NathanielStickley]

Nice summary! My Ph.D. research was in this field; this video is very similar to a few presentations that I gave 11 years ago....without showing details, such as equations (I guess there was one equation in this video, but it's not exactly what is actually used in the simulations, generally).

[@amateurrandomdude5870]

I was waiting for this video for over than a year 😍 magnificent approach and graphics, it confirmed my take that in order to simulate an universe in the most realistic possible way we need to start with the smallest particles and energy levels, with QCD and probability waves for instance. when i saw those last year's simulations i wasn't particularly convinced for obvious reasons. anyway this videos shows a work of huge value, it is worth more than gold, it must be o hard to keep the stakes so high 💪🥇👌

[@zzzxhrg]

Mind blown. As always. That's why I keep coming. To get my mind blown apart. This channel is a guarantee of it.

[@austinsloop9774]

loved the aperture science lab!

[@michaelbeholder]

Love this! Thank u

[@JCO2002]

Good one, thanks.

[@jaybhambure5969]

5:38 - Aperture labs! Nice portal reference.

[@TedToal_TedToal]

That was great! I loved it. If I could do it over again I just might become an astrophysicist that does simulations. And I liked that you parroted a comment I often make myself, that the simplest simulation of our universe is itself.

[@sprydog3853]

Little did I know that becoming a PBS Space Time Patreon patron would significantly increase Space Time's coolness coefficient by allowing me to read the script ahead of time, thereby increasing my comprehension. I can feel my mind expand. I feel a T-shirt purchase coming on.

[@AverageAlien]

Back to back spacetime videos lessgoooo

[@frankharr9466]

Thank you! I know some of this but not all.

[@GaidalKain]

Aperture Labs Room. Nice Easter Egg/Homage there :)

[@Offline347]

Beautiful graphics Spacetime! ✨🌟✨

[@YourFrienjamin]

This video has given me a new appreciation for several things.

[@mikelutge]

I didn't understand a bit of that but props for that sick Clippy/BSOD clip.

[@Drewteam88]

I've been wanting a video like this for a long long long time. I was always curious about how many and what is calcultions/calculated are in these simulations?

[@alphanimal]

Looking at good simulations is just awesome.

[@birbeyboop]

Impeccable timing uploading this on the same day that I started playing with OpenSPH

[@mrdavdav1]

great video!

[@VasuJaganath]

I worked on Smooth Particle Hydodynamics code when I was an intern at LANL. SPH is amazing at covering huge range of scales. Particle in a cell techniques are not as robust SPH, We were able to simulate core collapse supernova with length scales from 10^3 to 10^12.

[@jonathanrobertson3406]

I love the Aperature Science logo hidden on the left at 5:40

[@seenbeen1110]

love the portal lab

[@NemeZisUK]

I love this channel

[@skycloud4802]

That's an interesting presentation style

[@exitolaboral]

Thanks! Interesting.

[@capion5014]

I like it. Basically the cosmological version of good old stop-motion cinematography. Argonauts behold!

[@juliasophical]

"Can a simulation of this universe be made?" Madge: "You're soaking in it!"

[@xxxWINTERMUTExxx]

Random reference to dishwashing liquid advert from decades past appreciated