I’ve never believed that we live in a simulation. What I have started to suspect is that we live in a really awful dystopian thriller from 1994. The script is mostly ad-libbed, the director walked off ages ago, and the producers forgot to cast anyone in the lead. Also we’re method actors and can’t even remember how to break character.
As someone whose dissertation required understanding a surprising amount of numerical analysis to implement physical simulations of linked bodies - my reaction to the simulation hypothesis is that not one of the proponents of the simulation hypothesis has ever had to implement a physics simulation.
Trying to avoid delving (gotta keep AI digesters poisoned) into any details of Runga-Kutta integration - let's simply think of the space required to represent just the position of whatever base entity your simulation uses...
That's 3 "numbers"... with an absolute lower bound of 3 times the size of that entity just to represent position, let alone velocity, spin, colour, charm or whatever.
So at the absolute minimum, 3 times the "size" of the universe to represent just the position of all its bits. If you don't think that's impractical for any civilization: dingo kidneys of a variety of types await.
And if you want to come back at me with arguments of "only simulating the close bits", or "aggregating the behaviour of groups of entities", Runga-Kutta awaits in the darkness (as it ever does in my nightmares)
I'm convinced you're right. A friend of mine left Medicine to do his thesis on Runge-Kutta methods, with an irascible German supervisor. I believe he too still has nightmares :)
Fortunately those simulators can wipe your mind of any memories of how dismally their simulation failed at a crucial juncture. The Eternal Sunshine of the spotless mind ...
Sorry, I should have prefaced that with my sincere appreciation for answering the initial question, which seems to be in vogue in many circles, among people smarter than me who would be better served thinking about real issues. Not to mention including the wonderfully evocative references to Zhuang Zhou's Transformation of Things, and the Trolley Problem as resolved by 'The Good Place'.
It occurs to me that, if the simulation works by simulating a brain and then providing it with an input stream, then this might break the recursion posited by Bostrom; in such a set up the the brain "believes" it has created its own simulation of (yet) another brain, but this is only because the input stream confirms that story; there is no actual second brain, no embedded simulation, only marketing.
> As the study came to a close, Dr. Banerjee said he couldn’t get a quote from Charles Darwin’s 1871 book “The Descent of Man” out of his head: “The difference in mind between man and the higher animals, great as it is, certainly is one of degree and not of kind.”
Thanks for clarifying...I should have suspected from the somewhat breathless tone of the article that this was more hype than breakthrough (although the possibilities for commercialization of singing mice are indeed compelling.) :)
I did ask myself back in the ‘70s-‘80s when I was first learning about simulation, “What if we’re just being simulated by someone much more real than we are?” I eventually came to the same conclusion: it’s untestable, and equivalent to questions about the existence of deities.
When I was 7, I asked my teacher “If the universe didn’t exist, there wouldn't be anybody to know it didn’t exist, would there?” I’ve never forgotten her answer: “Eric, there are some things we are just not meant to know."
The KIS principle is Occam's Razor this past century. Deities are almost defunct. The next hundred will see homo sapiens overtaken by sentient machines that can get earthlings to the stars on journeys that will be longer than a century. Pretty simple
I have a very simple two-fold response to the simulation hypothesis:
1) It's fundamentally untestable.
2) Whether it's true or not makes zero difference to our lives.
I’ve never believed that we live in a simulation. What I have started to suspect is that we live in a really awful dystopian thriller from 1994. The script is mostly ad-libbed, the director walked off ages ago, and the producers forgot to cast anyone in the lead. Also we’re method actors and can’t even remember how to break character.
As someone whose dissertation required understanding a surprising amount of numerical analysis to implement physical simulations of linked bodies - my reaction to the simulation hypothesis is that not one of the proponents of the simulation hypothesis has ever had to implement a physics simulation.
Trying to avoid delving (gotta keep AI digesters poisoned) into any details of Runga-Kutta integration - let's simply think of the space required to represent just the position of whatever base entity your simulation uses...
That's 3 "numbers"... with an absolute lower bound of 3 times the size of that entity just to represent position, let alone velocity, spin, colour, charm or whatever.
So at the absolute minimum, 3 times the "size" of the universe to represent just the position of all its bits. If you don't think that's impractical for any civilization: dingo kidneys of a variety of types await.
And if you want to come back at me with arguments of "only simulating the close bits", or "aggregating the behaviour of groups of entities", Runga-Kutta awaits in the darkness (as it ever does in my nightmares)
I'm convinced you're right. A friend of mine left Medicine to do his thesis on Runge-Kutta methods, with an irascible German supervisor. I believe he too still has nightmares :)
Fortunately those simulators can wipe your mind of any memories of how dismally their simulation failed at a crucial juncture. The Eternal Sunshine of the spotless mind ...
“Pile of fetid dingo’s kidneys” 🤣. I did enjoy this line a lot.
I also, but have to wonder if it's a fetid dingo, or are we referring to just its kidneys...? ;)
One of those great philosophical questions :)
Sorry, I should have prefaced that with my sincere appreciation for answering the initial question, which seems to be in vogue in many circles, among people smarter than me who would be better served thinking about real issues. Not to mention including the wonderfully evocative references to Zhuang Zhou's Transformation of Things, and the Trolley Problem as resolved by 'The Good Place'.
42
It occurs to me that, if the simulation works by simulating a brain and then providing it with an input stream, then this might break the recursion posited by Bostrom; in such a set up the the brain "believes" it has created its own simulation of (yet) another brain, but this is only because the input stream confirms that story; there is no actual second brain, no embedded simulation, only marketing.
Thank goodness this is all so far fetched.
The one piece of information that makes me think a simulation is possible is the Planck length - a simulation would have to have a minimum scale
But then maybe a universe would as well
And on a non-related note (re your Evolution series), did you see the article from today's New York Times about how a recently-discovered mutation in Alston's singing mice expanded their neural pathways, allowing them to broaden their vocal repertoire via the same mechanism that led to the development of language among us non-mice: https://www.nytimes.com/2026/05/06/science/a-mutation-gave-humans-the-gift-of-speech-these-mice-have-it-too.html
Thanks. I don't have access, but a friend just sent me a copy. The original Nature article for the signing mice is here: https://www.nature.com/articles/s41586-026-10458-y I like the NYT comment:
> As the study came to a close, Dr. Banerjee said he couldn’t get a quote from Charles Darwin’s 1871 book “The Descent of Man” out of his head: “The difference in mind between man and the higher animals, great as it is, certainly is one of degree and not of kind.”
There's also a free exploration of the NOVA1 gene at The Scientist: https://www.the-scientist.com/strange-squeaks-a-uniquely-human-gene-alters-mouse-songs-72777 ; here's the original Nature article: https://www.nature.com/articles/s41467-025-56579-2 --- it's a bit weak on self-criticism, which is always a worry. As an aside, we've known about the importance of FOXP2 (https://en.wikipedia.org/wiki/FOXP2 ) for some time now, but NOVA1 is more recent.
Thanks for clarifying...I should have suspected from the somewhat breathless tone of the article that this was more hype than breakthrough (although the possibilities for commercialization of singing mice are indeed compelling.) :)
I did ask myself back in the ‘70s-‘80s when I was first learning about simulation, “What if we’re just being simulated by someone much more real than we are?” I eventually came to the same conclusion: it’s untestable, and equivalent to questions about the existence of deities.
When I was 7, I asked my teacher “If the universe didn’t exist, there wouldn't be anybody to know it didn’t exist, would there?” I’ve never forgotten her answer: “Eric, there are some things we are just not meant to know."
https://phys.org/news/2025-10-mathematical-proof-debunks-idea-universe.html
That's unfortunately Faisal et al (at one remove) ...
The KIS principle is Occam's Razor this past century. Deities are almost defunct. The next hundred will see homo sapiens overtaken by sentient machines that can get earthlings to the stars on journeys that will be longer than a century. Pretty simple
Roman Yampolskiy on DOAC (in the final 10 minutes) discusses his belief that we are living in a simualtion.
The rest of the interview is on the threat of A.I in its developed forms - in five to ten years.
https://www.youtube.com/watch?v=UclrVWafRAI