Yeah but it would be fun to get two minute summaries of papers that are not about visually interesting simulations and/or AI. Because that's pretty much the only thing this channel does. AI and "aesthetic" simulations.
@@narzissus Even in those subjects, listening to most people speak about simulations; it is really hard to make a laymen interested in the nitty-gritty of even the aesthetic simulations. Do you have an example of simulations outside of the spectrum you mentioned? The "non-aesthetic" simulations.
They said the same thing about video games. In this paper, one frame takes from 24.5s to 216.3s to render. Not exactly real time. Still amazing results nevertheless.
so true when we take the goggles off in 2034 we will so be amazed by Real life The place is just awesome. we can only make what we already are inside limitless but within the rules of the Creator ~
They are, they're using the same computers that you do. It's just that it takes hours or days to render. What you see in this video is pre-rendered. No computer can run this in realtime yet.
@@lpharmer3496 They probably are a fair bit more powerful than your computer, but at the end of the day any computer can render something like this. It's just a matter of how long it takes for the computer to do all the calculations, thus better computers will render faster than a weaker one.
I just wanted to say: Thank you Karoly, for all of your efforts. I studied at your institute, and even though i moved on to something new years later, its so nice to stay in touch with this discipline thanks to your short summaries. They’re very entertaining and insightful. I’m very grateful for that! What a time to be alive :)
Love the videos! I appreciate your work and always learn a little something. Also, you talk very friendly, it's a nice change from most of youtube today.
Oh, that was so much fun to watch! Thank you. The moment it made *click* and I got what they were doing. You're always doing great work on these videos. They're short but must still take some while to make. Awesome work! Keep it up. :)
There's so many ways you could apply this type of thinking too, as an example, imagine some kind of MMO where "chunks" of the world get "demoted" in the server when they're not being used, going from being in RAM to being in an SSD to being in an HDD, for example, and then getting promoted again when necessary. This might even already be a thing that I don't know about!
@@nagualdesign If the simulation is modeled after quantum mechanics, perhaps the tree is in a state of being fallen and not fallen at the same time, and once a player observes it, it'll collapse into-oh, now the whole forest is gone!
@@OrangeC7 this is pretty much what every single video game has been doing for years now. If there is no player around to see it, nothing is happening. What makes this paper interesting is that they also apply this principle to parts of the "world" that *are* being observed; it's just that some need less refinement than others. And even this has been a thing in video games for years; if in a typical game you look at a mountain in the distance, it's not gonna load all of its polygons and then zoom out so much that you won't see most of them. Instead, it will just load a much coarser representation that won't nearly be as detailed, but since it's so far in the distance there is no observable difference to the player.
definitely something to look into. the challenge would be that these adaptive algorithms have much less predictable costs, their worst-case performance isn't much better then no adaptiveness at all. so some kind of rough fallback would need to be hacked in to handle "all details" moments in sims which could otherwise blow their budgets and cause stuttering.
This sound early similar to space-time dilation in relativity, maybe the universe has space-time dilation implemented to save computational resources too, suggesting that reality might be some kind of simulation.
The same temporal optimizations can be used for high efficiency video compression. Slow scenes with little movement can have fewer key frames, with new key frames added whenever there’s a camera cut
I don't know why I continue to watch literally all your videos, I've never modeled or even animated. Hell making a video with a black screen and white text would be a struggle for me. I love them all though, and it may inspire me to make an animation or simulation at some point
Stay tuned for an absolute killer paper on bubble and foam formation in the next few weeks. It made it to this plugin (and presumably other places too in the industry): blendermarket.com/products/flipfluids
Studying business, I have yet no idea what I will need the information I'm gathering from this channel for, but I'm entertained and have set my mind to understand this.
You might want to look into econophysics -- using models from physics (e.g. from fluid dynamics) to simulate the movement of market stocks and other aspects of the economy.
Man, I don't know what any of this means, but it's so cool to watch these rendered fluid animations. Makes me want to play around with Blender when I get home.
@@unicorns537 I'm looking for Zbrush Character Sculptors and Riggers plus animation but I will need them once the game development process begins which can take up to 3-5 years, I'm still learning and planning how the game would be. Research and learn is my current aim.
Just like in video compression, we are able to use less data to achieve a similar result. We don't notice if the time between calculations is 1000 a second, or 1 a second if nothing changes in that section.
Amazing idea! So glad to hear about this. I can see some application for gaming simulations as well! It would be quite easy to put this in a level-editor. Literally a game-changer :)
"It's astounding ..Time is fleeting..Madness takes its toll But listen closely ..Not for very much longer I've got to keep control I remembeeeer doing the time-warp"... :P
This reminds me of an N-body gravity simulation I made that used regionally specific timestepping. It was so hard to also make sure that quantities like energy and momentum were being conserved.
I lost my shit when he said, "What a time to be alive". I've been ranting about this to a group of friends for the last half hour. Truly, if ever there was an argument to say we're in the future. Amazing.
Great article, but I was distracted from 0:37 on: "Dragon Massage" *gets vaporized* On a more serious note, I was kind of reminded of ray marching, where you take longer steps if you know there will be no possible collisions. Not to promo other channels, but both Sebastian Lague and The Coding Train did great videos demonstrating it in, for me personally, understandable terms. I wonder if this regional time stepping could have similar applications, i.e. in real time physics simulations, kind of like the RTX made ray tracing viable for real time.
@Mugabe Try Blender's built-in fluid/smoke simulator, it is completely free and super fun, and there are also example scenes available on BlendSwap and other corners of the web.
Now we just have to wait 10 years before it's actually implemented anywhere. Although SideFX seems to be pretty quick about it, it just takes to loong to get all this amazing tech implemented in widely used, non proprietary software!
the next step is conceptionally trivial, spatial-time-adaptation. I wonder how difficult the logic to assume the spatial and time parameters will be though.
rendering time has always been a pain in the ass but not as bad as before and not as good as what is to come. Once rendering time is almost unseen we can create our own simulated Planet just like the one we simulate life on now.
@Isaac Dweck Yeah haha I'm glad you're stoked about it haha. I get it. __the Theory I'm working on basically says that Time is a few different things: - _an object in itself -- this object is pure Time and it is not singularly any tense, as in past present future, or otherwise. What it takes to describe this Time is *comprehensive* time descriptors that are unique to this reference frame. New formulas and concepts and their respective terms must be designated here. This Time operates, in reference to other things, in a conjuntive-transformative syntax. This means that this Time relates individual things to other things in concrete, static manners, and also in transforming manners known as "flux." - _the blended solution of the pure Time and things that it *is able* to contextualize -- this Temporal Mixture (def. wanna rename that haha) is where we, as People which are things that use language, desire to refer to things as "... *that time* when we..." and so on. This still *is* real time, although it is specifically where pure Time's mechanics become considered in tandem with what they engage. This denomination is important because it is where pure Time *relates* to People, who are the only form of subject that is able to *rationalize* Time at all. - _the objects with which pure Time engages -- these are not Time, per se, however they are important to discuss because they are integral elements in temporal phenomena. With these ingredients, Time is able to express structure beyond itself, much like a thigmotropic plant using the fences around it to grow. These are the components I have considered so far as ~keystones for this Theory. I believe what needs to be developed now is the literal, absolute syntactical elements (similar to numbers, addition, subtraction, etc. in mathematics). It needs to be considered, also, what *other* fundemental elements might exist. Once those passive and active elements are defined, I believe a well-developed dedicated Temporal Theory will be available.
@Isaac Dweck oh yeah! (if you've already read the other comment) I forgot to mention, this video relates to that theory because I believe that one of the somewhat fundemental mechanics of Time in relation to things is a type of Temporal Compensation Function! It has to do with *necessary* or intended events in the universe that supercede other events and their effects. So, like... if Person A and Person B are *Supposed* to be in the same place at Time Zero, but they each take paths of differing lengths to get there, due to a requisite aspect of their meeting, Time sort-of *alters* itself so that the two objects and the event coincide. ^^like in the video! haha! .. funny stuff
@Isaac Dweck Yeah liberal systems developments like the number system in that video, I think, are part of the next step of civilization. Yeah, I may write a paper about that and some other things once I get some other stuff done and think of some ways to validate/support my ideas physically or mathematically. I'll let you know! haha take care
So, the simulation processing costs less by changing its time steps. In relativity, time passes differently depending on the object's speed in relation to the light speed. If our reality is a simulation, it would make sense to use relativity as a time adaptivity technique to reduce processing costs.
Go study physics before making poor analogies. The universe has no need to "optimize processing", in actuality a lot of information is spurious and non-optimized from a computing perspective. Even if this is supposed to be some argument for the simulation hypothesis, there is no need to optimize what doesn't require extra computing power in a quantum machine, which is actually the basis of physics as we know it. Simulating multiple probabilities of paths is "free" in a quantum computer. Not that this proves or disproves a simulation hypothesis, as it is an untestable conjecture and useless model of reality by any means.
the idea behind it seems simple to the point that im amazed it hasnt been done before, but just thinking about how you choose the timestep seems like a headache
The sharp borders dont really make sense to me. Alongside the simulation you would need a prognose of the movement to detect if there is a collision in the next timestep. Lets say a ball crashes into a wall, then you chose a small timestep for the moving ball and the ball and and a big timestep for the wall since its not moving. If the collision occurs at the timestep of the ball it could land in between a timestep of the wall and the collision would be asynchrous and delayed. To be in the same detail the collision region of the wall would need to be adapted in its timestep before the collision which adds a whole layer of prognose to the calculation making it slower instead of faster, wouldnt it? If you select the regions manually its for sure doing good, but thats kind tedious.
I haven't seen a channel where talking about nuance of simulations is actually entertaining. You're awesome.
Heartwarming message of the week. You are very kind, thank you so much! 🙏 - twitter.com/karoly_zsolnai/status/1217123538085261314
Yeah but it would be fun to get two minute summaries of papers that are not about visually interesting simulations and/or AI. Because that's pretty much the only thing this channel does. AI and "aesthetic" simulations.
@@narzissus Even in those subjects, listening to most people speak about simulations; it is really hard to make a laymen interested in the nitty-gritty of even the aesthetic simulations. Do you have an example of simulations outside of the spectrum you mentioned? The "non-aesthetic" simulations.
@@devilsadvocate2876 I think they mean papers that are beyond the scope of computer science, maybe? Just white papers in general?
Its all in the enthusiasm. The most terse topics can be interesting if the presenter is enjoying himself, and clearly TMP is enjoying this stuff.
"You bet your papers it does" is now my favorite sentence
Quite fond of "Hold on to your papers"
"It can be 10 times faster, so I can do all my simulations in 1 night
What a time to be alive"
@@thehiddenninja3428
Pedersen
@@ОлексійГрекул Olexei Grekul
_W H A T A T I M E T O B E A L I V E_
_W A T T B A_
Now imagine what it will be like after next decades.
@@zufex2029 W H A T A T I M E I T W I L L B E T O B E T O B E A L I V E T H E N !
@@luminousdragon L O L
computer virtual things and "be alive" don't fit in my mind
VR is going to be so amazing in 10 years. Everything is going to be so realistic.
What a time to be alive!
If they incorporate generative neural networks in video cards I would be willing to say 2 years (if not 1)
They said the same thing about video games. In this paper, one frame takes from 24.5s to 216.3s to render. Not exactly real time.
Still amazing results nevertheless.
so true when we take the goggles off in 2034 we will so be amazed by Real life The place is just awesome. we can only make what we already are inside limitless but within the rules of the Creator ~
10 years? You're so optimistic. At best it'll be 20-30 years.
0:37 *that does not look like a massage.*
That's one hell of a massage! LOL
dragon OBLITERATION
The massage obliterated the discomfort!
-and everything else too-
the walls were massaged with the entrails of the dragon
It made him so relaxed he melted
0:39 "Dragon massage"
More like dragon massacre! Poor thing!!!
as a dumbass who loves dragons
that "dragon massage" mentally hurt.
Absolutely stunning. I look forward to the day when everyday home computers are fast enough to render these kinds of fluid motions without issue!
They are, they're using the same computers that you do.
It's just that it takes hours or days to render. What you see in this video is pre-rendered. No computer can run this in realtime yet.
@@AtomicBlastPony I am fairly sure their computers are a _bit_ more powerful graphically than mine
@@lpharmer3496 They probably are a fair bit more powerful than your computer, but at the end of the day any computer can render something like this. It's just a matter of how long it takes for the computer to do all the calculations, thus better computers will render faster than a weaker one.
Well , maybe Quantum computers will be able to perform such tasks in real-time.
I implemented the material point method for snow simulation as my final project this last semester! Definitely going to read this paper as well.
I just wanted to say: Thank you Karoly, for all of your efforts. I studied at your institute, and even though i moved on to something new years later, its so nice to stay in touch with this discipline thanks to your short summaries. They’re very entertaining and insightful. I’m very grateful for that! What a time to be alive :)
Amazing, happy to hear from you and thank you very much for the kind message!
This is literally implementing a spacetime compression. Awesome!
This is brilliant! I love your videos!
WHAT A TIME TO BE ALIVE!!
Your enthusiasm is so uplifting.
Love the videos! I appreciate your work and always learn a little something. Also, you talk very friendly, it's a nice change from most of youtube today.
The idea isn't 'crazy' or even novel; but I am sure the implementation of _how_ they did it, is!!
I was thinking the same, dynamic-meshing and adaptive time steps have been used in CFD and FEA for years
0:30
Observe the hand movements after Dino is shot.
It's like he is begging to leave the simulation
*tired from being shot repeatedly*
You forget to see him spewing the water back..!?
@@Princedemars my bad (๑•́ ₃ •̀๑)
This channel is such a hidden gem, thank you for these videos.
Oh, that was so much fun to watch! Thank you. The moment it made *click* and I got what they were doing. You're always doing great work on these videos. They're short but must still take some while to make. Awesome work! Keep it up. :)
I love these videos
Thank you for all you do
Thank you for making these papers accessible and interesting
Damn that's actually quite brilliant. Wonder how well temporal adaptiveness would work alongside spatial, specially if you want real time rendering.
They were using both, that bird had timesteps changing with time and location, same with the snowy hill
There's so many ways you could apply this type of thinking too, as an example, imagine some kind of MMO where "chunks" of the world get "demoted" in the server when they're not being used, going from being in RAM to being in an SSD to being in an HDD, for example, and then getting promoted again when necessary. This might even already be a thing that I don't know about!
@@nagualdesign If the simulation is modeled after quantum mechanics, perhaps the tree is in a state of being fallen and not fallen at the same time, and once a player observes it, it'll collapse into-oh, now the whole forest is gone!
@@OrangeC7 this is pretty much what every single video game has been doing for years now. If there is no player around to see it, nothing is happening. What makes this paper interesting is that they also apply this principle to parts of the "world" that *are* being observed; it's just that some need less refinement than others. And even this has been a thing in video games for years; if in a typical game you look at a mountain in the distance, it's not gonna load all of its polygons and then zoom out so much that you won't see most of them. Instead, it will just load a much coarser representation that won't nearly be as detailed, but since it's so far in the distance there is no observable difference to the player.
definitely something to look into. the challenge would be that these adaptive algorithms have much less predictable costs, their worst-case performance isn't much better then no adaptiveness at all. so some kind of rough fallback would need to be hacked in to handle "all details" moments in sims which could otherwise blow their budgets and cause stuttering.
That clip of the dragon waving furiously with both arms is delightful
I could watch these simulations for hours and hours. Awesome stuff!
This sound early similar to space-time dilation in relativity, maybe the universe has space-time dilation implemented to save computational resources too, suggesting that reality might be some kind of simulation.
"All experiments were done on a six core Intel i7 8700k " Now there go all my excuses...🤔😂
*I don't have a computer*
The science and results are awesome but mostly I love how passionate and excited you are about all of this! "What a time to be alive" :)
it reminds me the quantum physics problems of small particles collision/interference and influence of an observer on it
The same temporal optimizations can be used for high efficiency video compression. Slow scenes with little movement can have fewer key frames, with new key frames added whenever there’s a camera cut
You make everything very clear while keeping the pace fast enough
I don't know why I continue to watch literally all your videos, I've never modeled or even animated. Hell making a video with a black screen and white text would be a struggle for me. I love them all though, and it may inspire me to make an animation or simulation at some point
Heartwarming message of the day. Thank you so much, you are very kind! 🙏
What a time to be alive, never thought we could change time step on different regions, master piece 👏
I still have no idea what this channel is even talking about really but I'm watching all the videos because of how entertaining it is
This is the same hack our alien overlords discovered when creating our simulation universe. They call it relativity
0:28 This is like one of the traps from the Cube movies, poor dragon squirrel!
Papers and Software are living in two parallel worlds. I remember fantastic papers from 2003,2004...2014.... and it is still on a paper....
Stay tuned for an absolute killer paper on bubble and foam formation in the next few weeks. It made it to this plugin (and presumably other places too in the industry): blendermarket.com/products/flipfluids
Studying business, I have yet no idea what I will need the information I'm gathering from this channel for, but I'm entertained and have set my mind to understand this.
You might want to look into econophysics -- using models from physics (e.g. from fluid dynamics) to simulate the movement of market stocks and other aspects of the economy.
Man, I don't know what any of this means, but it's so cool to watch these rendered fluid animations. Makes me want to play around with Blender when I get home.
I never thought I'd be on a channel where they explain papers about simulation and AI in >10 minutes, but here I am. What a time to be alive, indeed
Cannot wait to see simulations which implement temporal and spatial adaptivity in a single framework.
What a time to be alive!
Wattba!
I'm a developer and I aim for super realism, hopefully studying all of these papers will help me to achieve making a super realistic game!
Unless you are writing your own game engine...
@@deep.space.12 You can also modify engines likr unity or unreal thats what many people have done for triple A games like ark:survival for example
@Jackie Tearie thanks but it will take some time cause ita hard to create
@@unicorns537 I'm looking for Zbrush Character Sculptors and Riggers plus animation but I will need them once the game development process begins which can take up to 3-5 years, I'm still learning and planning how the game would be. Research and learn is my current aim.
@@unicorns537 thx
One day, someone may be writing a paper about the voice we hear in these videos... ya never know! :)
Just like in video compression, we are able to use less data to achieve a similar result. We don't notice if the time between calculations is 1000 a second, or 1 a second if nothing changes in that section.
Amazing idea! So glad to hear about this. I can see some application for gaming simulations as well! It would be quite easy to put this in a level-editor. Literally a game-changer :)
This adaptive time step seems like an absolute HEADACHE to benchmark
Nah.
Just compute some areas at t =0, 1, 2, 3, 4, 5....
And compute others at t=0, 2, 4, 6, ...
etc.
This is the best video you've ever made.
Help papers resisted my hold-down attempt, am now covered in paper cuts
I forgot to hold on, now it's papers everywhere 😁
"dragon massage"
*dragon is straight up liquefied*
wow
God, I love optimization! Working smarter not harder is the name of the game here!
Watching fluid simulations is so satisfying.
I really like the two minute papers logo animation
Yes! Now map and model a physarium (slime mold) and have an ai predict its movements until we learn what its decision making patterns are.
Nagyon király hogy mire képes a technológia márcsak optimalizálni kell :D
I've been planning on learning hungarian, I need to get to it
@@the_biblioklept2533 I just heard his name and I wanted to let him know that hungarians watch his videos too
"It's astounding
..Time is fleeting..Madness takes its toll
But listen closely
..Not for very much longer
I've got to keep control I remembeeeer doing the time-warp"... :P
Can you add this time step method together with the fine/coarse method?
What multiple times to be simultaneously alive!
This reminds me of an N-body gravity simulation I made that used regionally specific timestepping. It was so hard to also make sure that quantities like energy and momentum were being conserved.
I lost my shit when he said, "What a time to be alive". I've been ranting about this to a group of friends for the last half hour. Truly, if ever there was an argument to say we're in the future. Amazing.
Great article, but I was distracted from 0:37 on: "Dragon Massage" *gets vaporized*
On a more serious note, I was kind of reminded of ray marching, where you take longer steps if you know there will be no possible collisions. Not to promo other channels, but both Sebastian Lague and The Coding Train did great videos demonstrating it in, for me personally, understandable terms. I wonder if this regional time stepping could have similar applications, i.e. in real time physics simulations, kind of like the RTX made ray tracing viable for real time.
Any time I see a new two minute papers video, I want to start playing around with some simulations.
Do you have software/ code suggestions on how to start?
@Mugabe Try Blender's built-in fluid/smoke simulator, it is completely free and super fun, and there are also example scenes available on BlendSwap and other corners of the web.
@@TwoMinutePapers Thank you:)
I always watch these videos hugging an A4 ream
this is like the brain frame rates O wait O!!!!
Thanks to Lambda to make this video possible!
Fckng amazing!! These scientists rock
eye gasms for 4 minutes and 47 seconds? I'm so in
This is why I love engineering. Problems then solutions then more problema then more solutions. Faster then any government/ideology stuffs.
holy heck now we're using King Crimson in simulations
"Dragon Massage"
*dragon fucking explodes*
finally! spatial delta t adaption in cfd!
Fabulous words. Love it.
He really explained this concept in an easily comprehendable method. I mean just try to understand this video with no sound😁
Now we just have to wait 10 years before it's actually implemented anywhere. Although SideFX seems to be pretty quick about it, it just takes to loong to get all this amazing tech implemented in widely used, non proprietary software!
I’m subscribed to this channel just to watch new simulation videos 😆
It's not the same. Look at the snow attached to the ball at the end of the simulation. SyncMPM leaves a larger trail behind it
I can't wait to use it in next Houdini FX
the next step is conceptionally trivial, spatial-time-adaptation.
I wonder how difficult the logic to assume the spatial and time parameters will be though.
Kind of reminds me of frustum culling. Very cool.
you sounded like GreatScott at the end! HAHA
I can't wait until we can have thin films of water that doesn't look like gel, without destroying a computer in the process.
pub.ist.ac.at/group_wojtan/projects/thin_fluid_features/thin_fluid_features.html
wonder if anyone has tried both methods in tandem. adaptive spatial and temporal detail.
HOLY SHIT *TEN TIMES??!?* FUCKING AMAZING! God, this truly is the best time to be alive in computer science.
Games will be so cool once they figure out how to make graphics like this play in real time
rendering time has always been a pain in the ass but not as bad as before and not as good as what is to come. Once rendering time is almost unseen we can create our own simulated Planet just like the one we simulate life on now.
this is actually similar to my own theory-in-development of how real time works
@Isaac Dweck Yeah haha I'm glad you're stoked about it haha. I get it.
__the Theory I'm working on basically says that Time is a few different things:
- _an object in itself -- this object is pure Time and it is not singularly any tense, as in past present future, or otherwise. What it takes to describe this Time is *comprehensive* time descriptors that are unique to this reference frame. New formulas and concepts and their respective terms must be designated here. This Time operates, in reference to other things, in a conjuntive-transformative syntax. This means that this Time relates individual things to other things in concrete, static manners, and also in transforming manners known as "flux."
- _the blended solution of the pure Time and things that it *is able* to contextualize -- this Temporal Mixture (def. wanna rename that haha) is where we, as People which are things that use language, desire to refer to things as "... *that time* when we..." and so on. This still *is* real time, although it is specifically where pure Time's mechanics become considered in tandem with what they engage. This denomination is important because it is where pure Time *relates* to People, who are the only form of subject that is able to *rationalize* Time at all.
- _the objects with which pure Time engages -- these are not Time, per se, however they are important to discuss because they are integral elements in temporal phenomena. With these ingredients, Time is able to express structure beyond itself, much like a thigmotropic plant using the fences around it to grow.
These are the components I have considered so far as ~keystones for this Theory. I believe what needs to be developed now is the literal, absolute syntactical elements (similar to numbers, addition, subtraction, etc. in mathematics). It needs to be considered, also, what *other* fundemental elements might exist.
Once those passive and active elements are defined, I believe a well-developed dedicated Temporal Theory will be available.
@Isaac Dweck oh yeah! (if you've already read the other comment) I forgot to mention, this video relates to that theory because I believe that one of the somewhat fundemental mechanics of Time in relation to things is a type of Temporal Compensation Function! It has to do with *necessary* or intended events in the universe that supercede other events and their effects. So, like... if Person A and Person B are *Supposed* to be in the same place at Time Zero, but they each take paths of differing lengths to get there, due to a requisite aspect of their meeting, Time sort-of *alters* itself so that the two objects and the event coincide. ^^like in the video! haha! .. funny stuff
@Isaac Dweck Yeah liberal systems developments like the number system in that video, I think, are part of the next step of civilization.
Yeah, I may write a paper about that and some other things once I get some other stuff done and think of some ways to validate/support my ideas physically or mathematically.
I'll let you know! haha take care
King Crimson skips over tiny collisons
we all love you Károly ❤
Now create a gravity function and link it to this time step technique, you have linked spacetime to a simulation
So, the simulation processing costs less by changing its time steps. In relativity, time passes differently depending on the object's speed in relation to the light speed. If our reality is a simulation, it would make sense to use relativity as a time adaptivity technique to reduce processing costs.
Yep, also spacial adaptions like octrees. Which I want to say is what was shown off in the video.
Go study physics before making poor analogies. The universe has no need to "optimize processing", in actuality a lot of information is spurious and non-optimized from a computing perspective. Even if this is supposed to be some argument for the simulation hypothesis, there is no need to optimize what doesn't require extra computing power in a quantum machine, which is actually the basis of physics as we know it. Simulating multiple probabilities of paths is "free" in a quantum computer. Not that this proves or disproves a simulation hypothesis, as it is an untestable conjecture and useless model of reality by any means.
Awesome simulation
These papers are crazy.
Wow this reminds me of how world simulation works in Terraria.
I wish these 5 minute, two minute papers were 20 minutes.
This kind of explains the quantum double slit experiment and how everything might be simulations in real life.
Funny how that asynchronous simulation method functions much like time does in our universe.
Can we simulate sand for beach eg crabs or any reptiles crawling in sand
Is this how black holes work?
the idea behind it seems simple to the point that im amazed it hasnt been done before, but just thinking about how you choose the timestep seems like a headache
The sharp borders dont really make sense to me. Alongside the simulation you would need a prognose of the movement to detect if there is a collision in the next timestep. Lets say a ball crashes into a wall, then you chose a small timestep for the moving ball and the ball and and a big timestep for the wall since its not moving. If the collision occurs at the timestep of the ball it could land in between a timestep of the wall and the collision would be asynchrous and delayed. To be in the same detail the collision region of the wall would need to be adapted in its timestep before the collision which adds a whole layer of prognose to the calculation making it slower instead of faster, wouldnt it? If you select the regions manually its for sure doing good, but thats kind tedious.
" A simulation where time advances in different amounts at different spatial regions"... so, like reality? :)
Noooo, time in reality is like truth or love, an object of the spirit.
Would it be possible to combine the two approaches? Having a type of spacio-temporal adaptive simulation?
finally a vid I understand