I'm trying out a different editing style compared to my previous videos; all suggestions & criticisms appreciated! Would you prefer: - More/less editing? - Longer/shorter? - Less process of making it / more pretty animations of the end result? Let me know your thoughts! Edit: Thanks to everyone who's left feedback so far! As many have pointed out, there are a million ways in which this sim falls short as a true model of reality. I'll be returning to this project in the future to attempt to incorporate more accurate physics. Aside from that, this video has already gotten way more attention than I ever anticipated, but I'm happy to see people engaging with it. The best is yet to come!
i like the 3 leinght types, 10 min or less simpler topic 15/20min most likely to pick in a feed to watch 40+ lil risky but from time to time can be true gems by my understaning of the alg.. glad if it helps!
Oh boy i can't wait to watch him slowly realize you can't just "ignore quantum physics" when it comes to recreating fundamental particles and watch him slowly lose his mind over weird definitions, wave functions, and literal unsolved phenomena
well actually you totally can ignore quantum physics, because the whole concept of it is trying to fill in what happens to particles while they arent interacting with something. unless he wanted to observe the wave properties of light at a macroscopic scale - which would require him to simulate billions of wave-particles - then it doesnt matter at all.
@lightningbmw2309 You Actually cant, you need the quantum mechanics at least to form stable atoms ( unless you ignore Maxwell's equations). To actually model interactions you need qft
@@pacocaapps6775 I mean he could totally just make the electrons follow a rigid set of rules when orbiting, and just implement all of the results instead of the actual mechanics that allow it to happen. He seems to have made the whole code particle and force based, so it is quite impossible for him to make anything quantum without starting back from scratch.
Tbh I think this is more of entertainment than anything. More accurate simulations already exist, this is just a dude having some fun making a rough approximation and nothing's wrong with that lol
Yeah like the electrons not orbiting around two atoms when theirs more than 3 electrons and then them orbiting further away has to do with the electron shell... I think bro just made an accurate quantum simulation 💀
As someone with a physics degree too, I actually thought that the sheer number of times he neglected some force or other makes him fit in quite well with the physics community.
This video reminds of the joke about programming with AI Coding takes me about five hours and then debugging takes another four. But now, thanks to Chat GPT, I can code in just 3 minutes and debug in just 3 days!
The reason a clump of neutrons isn't stable has to do with quantum energy levels in the nucleus, and the fact that the weak nuclear force allows neutrons to turn into protons and vice versa. In real life, about half the neutrons would quickly turn into protons via beta decay until the energy of the nucleus was minimised.
If there cannot be more than one god, then this god cannot be a MonoBehaviour unless you instantiate it. In which case it's an existing being and not a force. We learn everyday.
My brother , there is no such thing as proton radius. Infact, there is no such thing as radius. All particles are probability waves in a space time soup.
It's a good enough approximation for protons and neutrons, especially for the purposes of building a sandbox like in this video. I agree that it's a much worse approximation for electrons, though.
@@jackbrand4507 We have literally already coded and made models of the "probability clouds"of particles at this scale, so it wouldn't have been too hard to literally just take the average of the cloud chart and convert them to vector points for a model, then just take it's radius
@@damprat141 Fair enough. An issue with that is you will have to store the position distribution (cloud chart) of each energy eigenstate of the free electron and proton (the neutron will be practically the same as the proton), and assign each particle a cloud based on how much kinetic energy it has. You will then have to update this each time the particle's energy changes. There are infinite excited states, so you will have to create an upper energy limit. Additionally, the interactions between free particles will create new energy eigenstates that have their own unique clouds, so you will have to store these as well, and clamp particles to certain eigenstates based on how close they are to other particles and which "particle group" they are in (eg. two protons and two neutrons forming an alpha particle). Now that I type this, it does actually sound do-able.
if god is real. imagining god doing this for like 1 bllion years and being like “FUCK WHY ARE THEY HALFWAY ACROSS THE UNIVERSE!” and “HOW ARE ALL THE ELECTRONS IN THE FOURTH DIMENSION”
Suggestion: Replace the electrostatic force with the Lorentz force, which combines the electrostatic and magnetic forces. You may have to program some numerical integration in order to solve Maxwell's equations at each time step, since the electric and magnetic fields are coupled by differential equations. But, I think it will be worth it, since you may currently be losing a lot of phenomena by not including the magnetic force.
@@garrettbates2639 idk what det means out of specific contexts, i think ive heard of ikke before but i cant remember what it means either because of skill issue or cuz i woke up 2 hours ago, but good afternoon. day 20 of duolingo lol
i love how in the end he gave up trying to have electrons orbit so he just forced them. he did a bohr. its an interesting video... when you strip away the physics from the equations, you get weird stuff like that. i believe you can get better behaving systems if you respect the physics a little more. the first thing i notice, that it also made laugh, was you trying to grap the "electron radius"... because this doesn't sense in our current understanding of the electron, lol. You see, the size of something is determined by the volume in space that encloses what that things can interect with or not. But in the case of the electron, its volume of interactivity is not a fixed thing. it changes based on whats around it, whats interacting with it. in one case it can be a ball, but in other cases it can be a ballon-shaped, a flower, a ring with ballon - shaped. it can even be a sine wave-shape. then there is another problem. this shape cannot be precisely determined in fixed values without you loosing track of that electron, and, thus, changing its shape. for if you try and measure its radius, you have to know at which point interaction starts to happen, but as soon as it happens (its instant), the shape changes. you change the property of the mf by trying to measure it. i also found it funny you using Newton's laws to calculate 2 very quantum mechanical systems lmao. yes, the strong force does have this quality, which can be modeled by forces (you calculate the potential and take the negative gradient), but then you loose track of all the other Quantum stuff, such as tunneling, fluctuations, superposition, etc. like, nuclar decay happens ONLY because of tunneling. you can't get nuclear decays with newton's laws. if you think you have it, its because the particles had enough energy to decay. tunneling happens even if you don't have enough energy. it simply happens, you know? the strong force is just too strong. finnaly i also found it funny how you kept running into the problem of things going faster than light. in real life, when things approach light speed, their rate of passage of time changes, it increases, it takes literally longer for time to move, when compared to other. this happens at the time that their apparent lengths contracts, so the objects look smaller along the direction of travel. i dont quite know just how much that affects the interactions, but the most interesting thing about this is this phenomenon literally explains why we have magnets (look it up, go deep). nice vid, interested to see if you'll follow it up.
@@gcewing i wonder if or how he will incorporate relativistic effects... the real momentum equation , for example, that actually govern collisions... in classical theory atoms can't even form yknow?
@@zeynaviegas Yeah, that's the problem with these kinds of semi-classical simulations. Atoms and nuclei are highly quantum-mechanical objects, so you're never going to get the simulation to behave realistically without tackling the quantum effects head-on.
@3RR0RNULL If you are doing something like making accurate physics in a game, rounding up can cause a lot of problems, nuclear physics require very accurate numbers because of how small atoms are, (irrational numbers are usually rounded up to the 7th-15th digit).
Love how you start talking about how you love AI and every issue here is the exact reason I don't use it. If my program is bugged, I want it to at least be my own fault. Also those trails do look like photo rays! There's fog boxes to see radiation trails. Very fun.
Yeah, it was sort of upsetting watching him look at a nice stable simulation and go "hm this isnt exciting enough" and then crank the forces until everything was breaking 😭
this is literally the type of project that inspired me as a youngin to got me started on my phd in computational chemistry!! I thought it would be more efficient to encode all of physics into a game engine than to arbitrary tell a barrel to explode and needed to find out why no one was doing it yet lol
Oh that sounds so cool and difficult. It's probably useful for poorly understood steps in reactions or whatever. I don't know much about chemistry but I'm sure you must have to do some weird programming optimizations to get good performance for your simulations. How is it going?
I think an accurate simulation of this to play around with would be so god damn cool, but the fact that you got this cartoonish version working like this is also really cool, and i love it
Absolutely love this, it’s so sick! I have some suggestions, if you don’t mind: 1. Manually continuously variable universe size 2. Manually continuously variable scale (aka by increasing the scale, it makes everything smaller while making the universe bigger and your own movement faster, and vice versa) Also, gravity isn’t actually that strong, so if you made it weaker, it might result in the atoms individually acting with less regard for it, while collective bodies experience it more and behave more like we see celestial bodies behave irl
I've started learning programming recently and good god dude, this unfathomable to me, I mean I can fathom it because I watched, but GOOD GOD this is insane, subbed
I don't know how to code, I don't know much about physics, it is 2am and this is video just hits all the right spots. I don't know why but I could watch you do this for hours. (and the music is great)
If you have any experience with gpt coding you would know that gpt is not very good at all at original concepts, but is great at adding simply modifications to your code. So for example it could have in a million years write this code alone, but if you give it the full skeleton, muscle skin and some other stuff it might be able to add the stomach, or make the skin function properly. It’s like a research assistant, doing the menial tasks that are minuscule next to the larger theoretical picture.
@kadephillips576 I get it, yeah. This argument is exactly why people may tend to use it. But it doesn't quite fit my original comment and the problem I saw with the video. I do have a lot of experience with coding and know very well how and in which situations ChatGPT can assist the programmer. And this usage (in the video) is not one of them. You said ChatGPT can be really helpful if you give it enough structure / additional context to the specific problem. But this only holds (and also just kinda) true, as long as the surrounding code and context isn't also generated by ChatGPT. And this is the case in the video and what I wanted to criticize. He didn't use it to assist him every once in a while for boilerplate or that kinda stuff, he used it almost exclusively instead of quickly coding the things himself and rather spent the time fixing the (because of the also ai generated context-code) shitty output produced by ChatGPT. Not to mention that ai today, although it may seem impressive at first, is not at all good at generating code in the first place. You can really see that in the video too. In the same way that image generation ai may hallucinate some extra fingers or wierd random shapes, code generation ai often does the exact same thing with countless nonsensical and not needed statements in disguise. So: Using ChatGPT isn't actually good no matter how you put it, does not give off the sign of being a good programmer and yet, this precisely is what the video does/shows. This is what I criticized.
@ very fair. On that front I agree. It’s important to understand your code enough to debug it yourself. The generate code and plug it back in and regenerate and plug back in is not always an effective strategy.
This is like genuinely one of the best videos I’ve been recommended lately. Probably because I both enjoy coding and physics but the editing and flow is also phenomenal
4:24 : gravity? I don’t think gravity plays a non-negligible role in an atomic nucleus, with the exception of neutron stars 5:21 : ah, so your universe is RP^3 shaped, cool. 7:03 : I really would prefer if people reserved “exponentially” for c^x , and didn’t describe things that increase quadratically (x^2) as “increasing exponentially”.
I hate it when people say "It's not exponential because it isn't exactly m^n" when in general, exponential just means it increases more and more, not that specific function (which I wish wasn't just called 'exponential' because again, outside of a specifically mathematical context it just means that it's more than linear)
@@stellarx20 I think the word “exponent”, and so “exponential”, originated in mathematics? Etymology online says that the earliest use of “exponent” is the mathematical one.
Well the alternative is asking mediocre programmers like you on stack overflow. But on another note it is annoying how instead of seeing the writing process we just see him copy pasting chatgpt lmao
@@crimmerz2000 I dont think you get it. chatgpt gets its data from the internet. so a lot of its responses to questions are going to be _from stack overflow_. chatgpt is only as good as its input data; mediocre in, mediocre out. plus, you wont even need stack overflow a lot of the time if you're actually a _good programmer that reads the goddamn documentaion_. god.
@@crimmerz2000 i don't think you get it. chatgpt gets its data from the internet, so most of its responses to questions are going to be *from stack overflow.* so the quality is going to be the same, if not worse; garbage in, garbage out. plus, you won’t actually _need_ stack overflow or chatgpt that often if you’re actually *a GOOD programmer* that *reads the goddamn documentation.* god.
With your circle wrapping around, it's topologically not a torus. For a topological torus you want a square wrapping around. With your circle wrapping around, i think it is topologically a projective plane (something that is halfway between a klein bottle and a torus). It might be a kline bottle tho. I'm not completely sure.
@nuke_clear no. If you take something from the center, cross the boundary, then return to the center, it will have become mirrored. That isn't possible on a 2-sphere or a torus.
You need to include quantum effects my guy. A lot of phenomena are more rare IRL than in your sim because they require the particles to tunnel through potentials rather than have enough energy to overcome them.
ok so, neutrons do not interact with electrons as they dont have any charge, neutrons and protons only interact and bind because of strong nucleus force, and electrons are not effected by this anyway the strong force gets stronger the further away a proton gets from a neutron, and at a certain point it just repels them and thus they bind neutrons arent attracted to each other, neither are protons, protons repel eachother just like electrons repel eachother
THIS HAS TO BE THE BEST VIDEO EVER If drawing something in MS Paint and immediately turning it into a 1 day unity project with feature creep was a contest, you'd be the winner. I've tried many times but failed, I bow to you
When you made the strong force length-dependent I chuckled to myself “yeah, I remember when I thought it made sense”. I hesitate to mention a fundamental flaw in how you’re simulating electrons: we cannot know both the position and momentum of any electron each with unlimited certainty. This becomes important for electron-electron interactions.
i like the strategy of "figure out ways to make it kind of like how it would be ish through unconventional methods" instead of "look up exactly how it works and code that"
this would be sick as a vr chat avatar or something, also this is the first video I have seen of yours and its really awesome! Love the editing style and how you break down the process.
7:00 if the force is (inversely) proportional to distance *squared*, then as distance gets small the force increases *quadratically*, not exponentially.
Technically a more accurate way to program movement is to check how close a non-light particle is to the speed cap, and have applied velocities exponentially decrease in how much a particle's velocity is actually increased by, being multiplied by zero once an object reaches light speed. Not that they should ever get there, as the formula should decrease the speed by so much that non-photon objects can never reach light speed. Unfortunately this would require calculus to use without serious bugs. Also unfortunately I do not know the formula for this off the top of my head, but it's part of relativistic physics near the speed of light. (In the real world increases in velocity at near-light-speeds also decreases the amount of time experienced by an object, causing near-light objects to age slower. Oddly enough this means time stops for you at the speed of light. Quantum- and astro-physics are weird and awesome.)
Cool video, look into the Barnes-Hut algorithm to to approximate some of the calculations so you can have wayyyyyy more particles. Not sure how it would be implemented in Unity though...
This editing style feels nearly perfectly tuned to my (admittedly odd) brain. I aspire to this level of competence in pretty much anything. You have an impressive mind.
The high-energy to low-energy photon is very realistic! The upper and lower bounds for photon emissions are much lower, and also atom dependent more than proton/neutron dependent, but you will see that kind of behavior all the time. It's why light can warm things up, for one.
this video is incredible! unironically a great teaching example for what programming is like and how much of it is just constantly learning/digging through new things
I like to think that this is how the universe was made except god didn’t have any outer sources so it was just trial and error for quintillions of years
![BLACK BAG - Official Trailer [HD] - Only in Theaters March 14](http://i.ytimg.com/vi/Du0Xp8WX_7I/mqdefault.jpg)
I'm trying out a different editing style compared to my previous videos; all suggestions & criticisms appreciated!
Would you prefer:
- More/less editing?
- Longer/shorter?
- Less process of making it / more pretty animations of the end result?
Let me know your thoughts!
Edit: Thanks to everyone who's left feedback so far! As many have pointed out, there are a million ways in which this sim falls short as a true model of reality. I'll be returning to this project in the future to attempt to incorporate more accurate physics.
Aside from that, this video has already gotten way more attention than I ever anticipated, but I'm happy to see people engaging with it. The best is yet to come!
i like the 3 leinght types,
10 min or less simpler topic
15/20min most likely to pick in a feed to watch
40+ lil risky but from time to time can be true gems
by my understaning of the alg..
glad if it helps!
so 15/20min 60% of the time i think is cool, editing and jokes are on point, the classical music is the cherry on top, peace!
It's fine how it is. No changes; you found the best middle point.
whatever style this is, keep it up.
I think this style of editing is perfect, I love underrated channels like yours
Oh boy i can't wait to watch him slowly realize you can't just "ignore quantum physics" when it comes to recreating fundamental particles and watch him slowly lose his mind over weird definitions, wave functions, and literal unsolved phenomena
@@whomidity3953 and when he tries to put relativity in the mix? the quantum stuff + the relativity stuff lol
well actually you totally can ignore quantum physics, because the whole concept of it is trying to fill in what happens to particles while they arent interacting with something. unless he wanted to observe the wave properties of light at a macroscopic scale - which would require him to simulate billions of wave-particles - then it doesnt matter at all.
@lightningbmw2309 actual iterator discussion threads right here
@lightningbmw2309 You Actually cant, you need the quantum mechanics at least to form stable atoms ( unless you ignore Maxwell's equations). To actually model interactions you need qft
@@pacocaapps6775 I mean he could totally just make the electrons follow a rigid set of rules when orbiting, and just implement all of the results instead of the actual mechanics that allow it to happen. He seems to have made the whole code particle and force based, so it is quite impossible for him to make anything quantum without starting back from scratch.
The amount of wrong physics in this video is so fascinating, it actually is fun
@@thehexagon_yt i think this can used in classes for a good discussion
yeah right
Tbh I think this is more of entertainment than anything. More accurate simulations already exist, this is just a dude having some fun making a rough approximation and nothing's wrong with that lol
Erm you are not using the correct formula for collapsing the wave function 🤓🤓🤓🤓
"Because here we embrace the future"
_Promptly wastes 30 hours thanks to the "future"_
The future seems incredibly bleak
I'm all for using chatGPT to solve a problem you can't fix yourself or implement known algorithms, but my god.
Man slowly discovers why quantum is very important in nuclear phyisics
funny thing is some of the bugs he was finding remind me of actual weird phenomenon that were found in quantum physics
Yeah like the electrons not orbiting around two atoms when theirs more than 3 electrons and then them orbiting further away has to do with the electron shell... I think bro just made an accurate quantum simulation 💀
Makes you wonder if some of this stuff, like quantum entanglement, are just bugs in the code of the universe.
Quantum Theory is just unintuitive, not inexplicable.
@@b0r0din988 quantum properties are features, the fact that you can't intuitive them is a user error
so he liteally made quantum mechanics by breaking classical mechanics
“okay, let’s implement this fix” *tells chatgpt to implement the fix*
bro didn't update his atomic model to Schrodinger's and is still using Bohr's deprecated libraries
"This formation seems pretty stable
*shows a hydrogen atom with three neutrons*
that's tritium, oh wait, it's not tritium
Physicist here. I have to say, I wasn't too optimistic about this for a while lol. But I'm glad I stuck around til the end! that looks really cool.
What degree have you done?
he basically stripped the physics out of physics, and only worked with the formulas lmao 🤣 no wonder things broke so fast
Thank you for saying that, made me stick around too and it was worth it. Made me wonder why electrons don't in fact collide with protons irl
@@racoons3645 because they aren't just small spheres like in his simulation
As someone with a physics degree too, I actually thought that the sheer number of times he neglected some force or other makes him fit in quite well with the physics community.
This video reminds of the joke about programming with AI
Coding takes me about five hours and then debugging takes another four. But now, thanks to Chat GPT, I can code in just 3 minutes and debug in just 3 days!
The reason a clump of neutrons isn't stable has to do with quantum energy levels in the nucleus, and the fact that the weak nuclear force allows neutrons to turn into protons and vice versa. In real life, about half the neutrons would quickly turn into protons via beta decay until the energy of the nucleus was minimised.
If there cannot be more than one god, then this god cannot be a MonoBehaviour unless you instantiate it. In which case it's an existing being and not a force. We learn everyday.
My brother , there is no such thing as proton radius. Infact, there is no such thing as radius. All particles are probability waves in a space time soup.
It's a good enough approximation for protons and neutrons, especially for the purposes of building a sandbox like in this video. I agree that it's a much worse approximation for electrons, though.
the last sentence is not true but everything else is
@@jackbrand4507 We have literally already coded and made models of the "probability clouds"of particles at this scale, so it wouldn't have been too hard to literally just take the average of the cloud chart and convert them to vector points for a model, then just take it's radius
@@damprat141 Fair enough. An issue with that is you will have to store the position distribution (cloud chart) of each energy eigenstate of the free electron and proton (the neutron will be practically the same as the proton), and assign each particle a cloud based on how much kinetic energy it has. You will then have to update this each time the particle's energy changes. There are infinite excited states, so you will have to create an upper energy limit. Additionally, the interactions between free particles will create new energy eigenstates that have their own unique clouds, so you will have to store these as well, and clamp particles to certain eigenstates based on how close they are to other particles and which "particle group" they are in (eg. two protons and two neutrons forming an alpha particle). Now that I type this, it does actually sound do-able.
You can give your particles a non-defined radius if you ever decide to be the god of a universe. Don't backseat divine his universe.
if god is real. imagining god doing this for like 1 bllion years and being like “FUCK WHY ARE THEY HALFWAY ACROSS THE UNIVERSE!” and “HOW ARE ALL THE ELECTRONS IN THE FOURTH DIMENSION”
Well.. God is supposedly All powerful and All knowing. So no bugs occur for God.
@@AnasKG2063just because he knows it’ll happen does make it any less irritating for him
Also dont interpret god saying the F-word, god would never say the f-word, this is the word from a fellow christian, have a good day
@ i am not christian, as you can tell by me saying “if god is real”, but of course, why wouldnt god swear?
@@fonsworththethirdcause he’s amazing and *totally* never does anything bad? idk Im not Christian
24:44
“You would think just a bunch of neutrons would be stable wouldn’t you”
my brother in christ do you know what a radioisotope is?
Suggestion: Replace the electrostatic force with the Lorentz force, which combines the electrostatic and magnetic forces. You may have to program some numerical integration in order to solve Maxwell's equations at each time step, since the electric and magnetic fields are coupled by differential equations. But, I think it will be worth it, since you may currently be losing a lot of phenomena by not including the magnetic force.
chatgpt really did a good job making this!
@Hi-pk5mf Guten Tag mr.lila
@@_PapstFranziskus_ The boy tag mr lila
idk if tag means something in norwegion ive only been practicing for like 20 days
you use chat gpt and do the same thing then
@@Radiance_alphaHann skrev på tysk, ikke norsk. Det betyr "god dag."
@@garrettbates2639 idk what det means out of specific contexts, i think ive heard of ikke before but i cant remember what it means either because of skill issue or cuz i woke up 2 hours ago, but good afternoon.
day 20 of duolingo lol
i love how in the end he gave up trying to have electrons orbit so he just forced them. he did a bohr.
its an interesting video... when you strip away the physics from the equations, you get weird stuff like that. i believe you can get better behaving systems if you respect the physics a little more.
the first thing i notice, that it also made laugh, was you trying to grap the "electron radius"... because this doesn't sense in our current understanding of the electron, lol. You see, the size of something is determined by the volume in space that encloses what that things can interect with or not. But in the case of the electron, its volume of interactivity is not a fixed thing. it changes based on whats around it, whats interacting with it. in one case it can be a ball, but in other cases it can be a ballon-shaped, a flower, a ring with ballon - shaped. it can even be a sine wave-shape.
then there is another problem.
this shape cannot be precisely determined in fixed values without you loosing track of that electron, and, thus, changing its shape. for if you try and measure its radius, you have to know at which point interaction starts to happen, but as soon as it happens (its instant), the shape changes. you change the property of the mf by trying to measure it.
i also found it funny you using Newton's laws to calculate 2 very quantum mechanical systems lmao. yes, the strong force does have this quality, which can be modeled by forces (you calculate the potential and take the negative gradient), but then you loose track of all the other Quantum stuff, such as tunneling, fluctuations, superposition, etc. like, nuclar decay happens ONLY because of tunneling. you can't get nuclear decays with newton's laws. if you think you have it, its because the particles had enough energy to decay. tunneling happens even if you don't have enough energy. it simply happens, you know? the strong force is just too strong.
finnaly i also found it funny how you kept running into the problem of things going faster than light. in real life, when things approach light speed, their rate of passage of time changes, it increases, it takes literally longer for time to move, when compared to other. this happens at the time that their apparent lengths contracts, so the objects look smaller along the direction of travel. i dont quite know just how much that affects the interactions, but the most interesting thing about this is this phenomenon literally explains why we have magnets (look it up, go deep).
nice vid, interested to see if you'll follow it up.
Thanks for the tips! I will have to follow this project up with a more accurate version in the future for sure.
@ThomasWaldYT oh please do! ill definitely watch it
Also using the Newtonian formula for kinetic energy when particles are moving near the speed of light...
@@gcewing i wonder if or how he will incorporate relativistic effects... the real momentum equation , for example, that actually govern collisions...
in classical theory atoms can't even form yknow?
@@zeynaviegas Yeah, that's the problem with these kinds of semi-classical simulations. Atoms and nuclei are highly quantum-mechanical objects, so you're never going to get the simulation to behave realistically without tackling the quantum effects head-on.
at 4 am I watch these videos
2:07 "We'll just round this off" 💀
???
@@3RR0RNULLif you round constants... well... its not good
@@im_kas Pi:
Also, why? It’s not like it matters in a simulation game.
Bro doesn't know @@3RR0RNULL
@3RR0RNULL If you are doing something like making accurate physics in a game, rounding up can cause a lot of problems, nuclear physics require very accurate numbers because of how small atoms are, (irrational numbers are usually rounded up to the 7th-15th digit).
Love how you start talking about how you love AI and every issue here is the exact reason I don't use it. If my program is bugged, I want it to at least be my own fault.
Also those trails do look like photo rays! There's fog boxes to see radiation trails. Very fun.
you gotta drop your dt, so many non-physical bandaid fixes when the real problem is just bad numerical integration
Yeah, it was sort of upsetting watching him look at a nice stable simulation and go "hm this isnt exciting enough" and then crank the forces until everything was breaking 😭
this is literally the type of project that inspired me as a youngin to got me started on my phd in computational chemistry!! I thought it would be more efficient to encode all of physics into a game engine than to arbitrary tell a barrel to explode and needed to find out why no one was doing it yet lol
Oh that sounds so cool and difficult. It's probably useful for poorly understood steps in reactions or whatever. I don't know much about chemistry but I'm sure you must have to do some weird programming optimizations to get good performance for your simulations. How is it going?
The laws of physics were absolutely VIOLATED in this simulation
I got so hyped when I saw the two hydrogen atoms form a covalent bond.
Apparently ChatGPT is a good teacher for compute shaders!
well, a whole section of machine learning science focuses specifically on them so
i mean... that's what nvidia uses for their "fake" frames on heavy hungry RTX games
Not really... Did they actually learn anything, or were they just given a load of code they didn't understand?
I think an accurate simulation of this to play around with would be so god damn cool, but the fact that you got this cartoonish version working like this is also really cool, and i love it
Absolutely love this, it’s so sick! I have some suggestions, if you don’t mind:
1. Manually continuously variable universe size
2. Manually continuously variable scale (aka by increasing the scale, it makes everything smaller while making the universe bigger and your own movement faster, and vice versa)
Also, gravity isn’t actually that strong, so if you made it weaker, it might result in the atoms individually acting with less regard for it, while collective bodies experience it more and behave more like we see celestial bodies behave irl
would love to see these changes, hopefully he sees it!
variable scale would just make the atomic theory lol
Watching this knowing even like a surface level amount of physics is so hilarious because it just doesn’t make any sense
I've started learning programming recently and good god dude, this unfathomable to me, I mean I can fathom it because I watched, but GOOD GOD this is insane, subbed
Man, I am just VIBING to that vivaldi in the background. Nice music choice :)
I don't know how to code, I don't know much about physics, it is 2am and this is video just hits all the right spots. I don't know why but I could watch you do this for hours. (and the music is great)
Hearing the "We will be skipping all the quantum mumbo jumbo" while actively preparing for exam on quantum chemistry broke me on fundamental level.
Cool but the title should be: "ChatGPT Coded a Nuclear Physics Simulator to Play God in VR"
cry harder
If you have any experience with gpt coding you would know that gpt is not very good at all at original concepts, but is great at adding simply modifications to your code. So for example it could have in a million years write this code alone, but if you give it the full skeleton, muscle skin and some other stuff it might be able to add the stomach, or make the skin function properly. It’s like a research assistant, doing the menial tasks that are minuscule next to the larger theoretical picture.
@kadephillips576 I get it, yeah. This argument is exactly why people may tend to use it. But it doesn't quite fit my original comment and the problem I saw with the video. I do have a lot of experience with coding and know very well how and in which situations ChatGPT can assist the programmer. And this usage (in the video) is not one of them. You said ChatGPT can be really helpful if you give it enough structure / additional context to the specific problem. But this only holds (and also just kinda) true, as long as the surrounding code and context isn't also generated by ChatGPT. And this is the case in the video and what I wanted to criticize. He didn't use it to assist him every once in a while for boilerplate or that kinda stuff, he used it almost exclusively instead of quickly coding the things himself and rather spent the time fixing the (because of the also ai generated context-code) shitty output produced by ChatGPT. Not to mention that ai today, although it may seem impressive at first, is not at all good at generating code in the first place. You can really see that in the video too. In the same way that image generation ai may hallucinate some extra fingers or wierd random shapes, code generation ai often does the exact same thing with countless nonsensical and not needed statements in disguise.
So: Using ChatGPT isn't actually good no matter how you put it, does not give off the sign of being a good programmer and yet, this precisely is what the video does/shows. This is what I criticized.
@ very fair. On that front I agree. It’s important to understand your code enough to debug it yourself. The generate code and plug it back in and regenerate and plug back in is not always an effective strategy.
@@kadephillips576 🤝
This is like genuinely one of the best videos I’ve been recommended lately. Probably because I both enjoy coding and physics but the editing and flow is also phenomenal
4:24 : gravity? I don’t think gravity plays a non-negligible role in an atomic nucleus, with the exception of neutron stars
5:21 : ah, so your universe is RP^3 shaped, cool.
7:03 : I really would prefer if people reserved “exponentially” for c^x , and didn’t describe things that increase quadratically (x^2) as “increasing exponentially”.
but its not quadratic as well lol its 1/r²
@ quadratic in (1/r) though
I hate it when people say "It's not exponential because it isn't exactly m^n" when in general, exponential just means it increases more and more, not that specific function (which I wish wasn't just called 'exponential' because again, outside of a specifically mathematical context it just means that it's more than linear)
@@stellarx20 I think the word “exponent”, and so “exponential”, originated in mathematics? Etymology online says that the earliest use of “exponent” is the mathematical one.
@@drdca8263 it doesn't matter though. semantics are dumb.
ChatGPT coded a nuclear physics simulator to play god in vr.
> "I coded"
> look inside
> chatgpt
well the alternate is stack overflow
@snugpig and knowing how to code instead of relying on a horrible excuse for a machine to do it for you
Well the alternative is asking mediocre programmers like you on stack overflow. But on another note it is annoying how instead of seeing the writing process we just see him copy pasting chatgpt lmao
@@crimmerz2000 I dont think you get it. chatgpt gets its data from the internet. so a lot of its responses to questions are going to be _from stack overflow_. chatgpt is only as good as its input data; mediocre in, mediocre out. plus, you wont even need stack overflow a lot of the time if you're actually a _good programmer that reads the goddamn documentaion_. god.
@@crimmerz2000 i don't think you get it. chatgpt gets its data from the internet, so most of its responses to questions are going to be *from stack overflow.* so the quality is going to be the same, if not worse; garbage in, garbage out. plus, you won’t actually _need_ stack overflow or chatgpt that often if you’re actually *a GOOD programmer* that *reads the goddamn documentation.* god.
With your circle wrapping around, it's topologically not a torus. For a topological torus you want a square wrapping around.
With your circle wrapping around, i think it is topologically a projective plane (something that is halfway between a klein bottle and a torus). It might be a kline bottle tho. I'm not completely sure.
Isn't it just a 2-sphere?
@nuke_clear no. If you take something from the center, cross the boundary, then return to the center, it will have become mirrored. That isn't possible on a 2-sphere or a torus.
5:03 flat earthers trying to explain how we can’t fall off the disk
You need to include quantum effects my guy. A lot of phenomena are more rare IRL than in your sim because they require the particles to tunnel through potentials rather than have enough energy to overcome them.
Electrons being in random ass spots works so perfectly to explain an electron field
ok so, neutrons do not interact with electrons as they dont have any charge, neutrons and protons only interact and bind because of strong nucleus force, and electrons are not effected by this
anyway the strong force gets stronger the further away a proton gets from a neutron, and at a certain point it just repels them and thus they bind
neutrons arent attracted to each other, neither are protons, protons repel eachother just like electrons repel eachother
Actually goated video wtf. Loved the format and hope you make more like this!
watching this video at 4AM really made it more immersive. i regret not sleeping though, i cant lie
31 views is less than I expected. Just keep up the good videos and your channel will blow up .
THIS HAS TO BE THE BEST VIDEO EVER
If drawing something in MS Paint and immediately turning it into a 1 day unity project with feature creep was a contest, you'd be the winner. I've tried many times but failed, I bow to you
Imagine how easy this would have been if you'd just wrote the program yourself
I'm feeling like i just found a gemstone! I can see the subs count doing numbers when the algorithm catches you, as sinister it's sounds. Keep up!
When you made the strong force length-dependent I chuckled to myself “yeah, I remember when I thought it made sense”.
I hesitate to mention a fundamental flaw in how you’re simulating electrons: we cannot know both the position and momentum of any electron each with unlimited certainty. This becomes important for electron-electron interactions.
Oh you sweet summer child. You cannot just ignore quantum physics…
"I've done a bit of coding off camera" is the programmers equivalent of "I did a little mining off camera", if using ChatGPT was frowned upon
I love how you program and I love your honesty. We have the exact same style
I love imagining the guy as god himself, and this is how the universe was created. Just some dude messing around
float4, float3, and float2 are just the shader versions of Vector4, Vector3, and Vector2
0:10 I'm more of a bottom, but I can be on top of you if you want to put me on UWU XD
this joke is unbelievably dumb and i love it
@qu3sti0nuble With tgis username, my bi ass just had to do it XD
@@NerdGlasses256 RESPECT. my ass is bisexual as well
This got me confused at first, but then I read your username right as he said "I gotta put on some nerd glasses." and you got a chuckle out of me.
LOL
this is amazing!! KEEP GOING I WANT TO SEE MORE THIS COULD BE VERY USEFUL.
bro literally made a low-poly universe sandbox on the atomic level i love it
31:33 optimal circle packing jumpscare
i like the strategy of "figure out ways to make it kind of like how it would be ish through unconventional methods" instead of "look up exactly how it works and code that"
this would be sick as a vr chat avatar or something, also this is the first video I have seen of yours and its really awesome! Love the editing style and how you break down the process.
7:00 if the force is (inversely) proportional to distance *squared*, then as distance gets small the force increases *quadratically*, not exponentially.
Please continue this, I want to see you go theoretically insane.
Technically a more accurate way to program movement is to check how close a non-light particle is to the speed cap, and have applied velocities exponentially decrease in how much a particle's velocity is actually increased by, being multiplied by zero once an object reaches light speed. Not that they should ever get there, as the formula should decrease the speed by so much that non-photon objects can never reach light speed. Unfortunately this would require calculus to use without serious bugs. Also unfortunately I do not know the formula for this off the top of my head, but it's part of relativistic physics near the speed of light. (In the real world increases in velocity at near-light-speeds also decreases the amount of time experienced by an object, causing near-light objects to age slower. Oddly enough this means time stops for you at the speed of light. Quantum- and astro-physics are weird and awesome.)
Cool video, look into the Barnes-Hut algorithm to to approximate some of the calculations so you can have wayyyyyy more particles.
Not sure how it would be implemented in Unity though...
The name is ringing a bell; I'll take a look.
didn't realize you only had 1k subs. production value was great and the video was a fun watch
independently discovering that you need a concept of momentum to make physics work at 33:10 was not on my bingo card for this video
I was thinking about if somebody made a simulator game and I'm glad somebody went out and did it.
Love the just copy and paste of large sections of code, from an obviously questionable source. Let me just add some more tape to my ball of tape.
I can't wait until this eventually evolves into a life simulator of some kind, being able to literally make life via buildings the atoms over time
I've always wanted to do this, haven't finished the video but I'm incredibly excited to see how this plays out
You know, a great way to avoid having AI fuck your shit up is to not use it
This editing style feels nearly perfectly tuned to my (admittedly odd) brain.
I aspire to this level of competence in pretty much anything. You have an impressive mind.
16:02 it's quantum tunneling guys in it's true form, but in 2d. they go to the third dimension to cross the wall.
no idea why youtube recommended this to me but it was super fun so new subscriber gained
oh man two minutes in wait till this guy learns how important QCD is
adding gravity to a nuclear physics simulator seems like a recipe for floating point error nonsense
I also hear that Claude Ai is realy good at working with code.
claude 3.5 sonnet is literally better than gpt4o at everything
fun fact, this is the exact set of circumstances that happened on day zero when God was still figuring out how to build the universe.
I feel your pain, on the rest of the code on changed, or placeholder, or just a small snippet. Drives me up the wall
Particle physics simulation made by someone who doesn't understand particle physics
to be fair he did a decent job
bro coded 15% of this, and how is it a simulator if you ignore half of the physics that makes it work?
"i coded all physics and a simulation of the big bang and billions of years in the future"
did he just say... "we're not gonna worry about any sort of quantum physics uwu :3"
Now you need to coda the probability clouds that fundamental particles actually consist of
-Is not gonna be too complex...
*said him, 2 years ago.....
that one unemployed friend on a sunday
I too at 4AM just decide to play god
This looks incredible for a dude with ~90 subscribers
I mean, did ChatGPT do most of the work.
@pl-AEtheRR I'm mostly talking about the video editing
@pl-AEtheRRMost is a exaggeration
Watching that while the new year begins and old ends
The high-energy to low-energy photon is very realistic! The upper and lower bounds for photon emissions are much lower, and also atom dependent more than proton/neutron dependent, but you will see that kind of behavior all the time. It's why light can warm things up, for one.
this video is incredible! unironically a great teaching example for what programming is like and how much of it is just constantly learning/digging through new things
having a master's degree in telling the truth, is of a different level of degrees bro 😭😭
this video is a testament to the fine tuning argument
It directly disproves it.
Splendid video, good sir! Subscription added, underrated channel. Good day!
Love your video. I've also wanted to work with atomic simulation for some time, and this made it easier!
you did exactly what I want to do lol
now I kinda want to make the code myself lol
Found my video to watch tomorrow with a snack and drink
I like to think that this is how the universe was made except god didn’t have any outer sources so it was just trial and error for quintillions of years