Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

ALiEn – a GPU-accelerated artificial life simulation program

alien-project.org

Post reply on HN
Print view View original

121–130 of 217 posts

Re: ALiEn – a GPU-accelerated artificial life simulation program

#121

Post by StandardFuture »

tgv wrote: post #87
p_j_w wrote: post #85

Earlier quoted context omitted.

Everyone saying it would be computationally expensive to simulate our universe is failing to put their mind outside of the box of our universe. Imagine for a moment that there's a universe which compares to our universe similarly to how our world compares to one inside of Conway's Game of life. Granted, this scenario doesn't provide us with anything we can take action on, but the idea that we're in a simulation at al…

So you only have to assume physics totally different from ours, and we can’t observe it. Isn’t that a bit of a weak point? And what would be the point of this simulation that has been running for billions of years?

> And what would be the point of this simulation that has been running for billions of years?

There is zero evidence for or against the simulation hypothesis, so why would some random person on HN be able to have the answer to this question even if we are in a simulation or even if we simply assume that we are?

Re: ALiEn – a GPU-accelerated artificial life simulation program

#122

Post by cwkoss »

StandardFuture wrote: post #118

Could someone explain the obsessive devotion to doing all Artificial "Life" research in terms of Cellular Automata? If you could supply the mathematical reasoning for this, I would be very interested in hearing the answer. And preferably an answer that goes beyond just saying that Von Neumann used Cellular Automata in his ALife research. Also, if anyone knows of alternative methods to CA in studying the properties of…

I think many artificial life simulations end up underwhelming because life is incredibly complex, and so it's very hard to simulate at scale. This ALiEn is perhaps the most advanced one I've seen, and it looks like even still they take some shortcuts (like copypasting interesting organisms from previous simulations together to create interesting interactions).

What you see as a criticism of this line of research I think is actually its reason: Life is arguably the most interesting thing in the universe, and if we can create it digitally it will surprise us. Evolution yields insights and solutions that you cannot predict. If we can synthesize what the minimal set of key properties are necessary for artificial lifeforms to create interesting unexpected outcomes, it helps us clarify the definition of what a non-biological organism could be.

I'm personally fascinated by the idea of autonomous digital agents that exist and self replicate while trying to earn cryptocurrency, which is used to pay for the hosting costs of themselves and their progeny. I think we are about two decades away from this being realized, but in the future, software services could self assemble, replicate imperfectly and evolve to please humans without any humans writing additional code: we'd just have to code a profitable LUCA, create suitable 'nests' and pay the organisms that please us. "What is life" is debatable, but IMO this would be a valid digital lifeform.

Re: ALiEn – a GPU-accelerated artificial life simulation program

#123

Post by StandardFuture »

dragonwriter wrote: post #120
StandardFuture wrote: post #118

Could someone explain the obsessive devotion to doing all Artificial "Life" research in terms of Cellular Automata? If you could supply the mathematical reasoning for this, I would be very interested in hearing the answer. And preferably an answer that goes beyond just saying that Von Neumann used Cellular Automata in his ALife research. Also, if anyone knows of alternative methods to CA in studying the properties of…

> Could someone explain the obsessive devotion to doing all Artificial "Life" research in terms of Cellular Automata? Uh, it doesn't exist. Plenty of A-life research doesn't use cellular automata as a model. > Also, if anyone knows of alternative methods to CA in studying the properties of life then I would also be interested in learning of these. https://avida.devosoft.org/

> https://avida.devosoft.org/

Thanks for the link!

> Plenty of A-life research doesn't use cellular automata as a model.

While I would like to believe you on that, one link does not seem sufficient to support the word "plenty" when the ratio of ALife projects built around CAs to not is extremely high.

Re: ALiEn – a GPU-accelerated artificial life simulation program

#124

Post by chimprich »

wyager wrote: post #52
zild3d wrote: post #42

Earlier quoted context omitted.

Is there any evidence suggesting we're not in a simulation?

A lot of physics seems unnecessarily expensive to compute. Quantum mechanics suggest we either have nonlocality or exponential blowup, both of which cause simulation challenges. With just classical physics you don’t need to deal with that. On the other hand, there are a lot of things that make physics tractable to compute, such as the +++- metric tensor and other factors forbidding causality violations. A universe wi…

> A lot of physics seems unnecessarily expensive to compute.

How do you know it's all computed? To make a convincing simulation, you just need to simulate in detail the bits that are actually being observed.

Everything else that happens could just be approximated at larger and larger granularity the further it is away from an observer.

Re: ALiEn – a GPU-accelerated artificial life simulation program

#125

Post by StandardFuture »

cwkoss wrote: post #122
StandardFuture wrote: post #118

Could someone explain the obsessive devotion to doing all Artificial "Life" research in terms of Cellular Automata? If you could supply the mathematical reasoning for this, I would be very interested in hearing the answer. And preferably an answer that goes beyond just saying that Von Neumann used Cellular Automata in his ALife research. Also, if anyone knows of alternative methods to CA in studying the properties of…

I think many artificial life simulations end up underwhelming because life is incredibly complex, and so it's very hard to simulate at scale. This ALiEn is perhaps the most advanced one I've seen, and it looks like even still they take some shortcuts (like copypasting interesting organisms from previous simulations together to create interesting interactions). What you see as a criticism of this line of research I th…

> so it's very hard to simulate at scale.

But, this is a very unaddressed point. Why focus on "simulation" when mathematical formalisms & theories could be potentially even more useful? Especially when most "simulations" are running on some arbitrary set of hard-coded assumptions?

> What you see as a criticism of this line of research

To clarify, I was in no way criticizing ALife research. Quite the opposite. I am actually trying to help ensure it does not get stuck in a rut.

Re: ALiEn – a GPU-accelerated artificial life simulation program

#126

Post by westoncb »

sooheon wrote: post #112
IkmoIkmo wrote: post #109

Earlier quoted context omitted.

Could you recommend one or two of these short stories? Thank you!

Pick up Axiomatic, try the first story (The Infinite Assassin) and see if it grabs you. It's got the best parts of something like Snow Crash, which drops you in without much exposition, and lets the visuals/action lead. Learning to Be Me from the same collection really stood out to me, it's a unique take on the "uploading your consciousness" trope.

My recollection from Infinite Assassin is that for someone not pretty literate in mathematics it would nearly be nonsense.

Re: ALiEn – a GPU-accelerated artificial life simulation program

#127

Post by kadoban »

411111111111111 wrote: post #49
dijit wrote: post #23

Earlier quoted context omitted.

Are we celebrating monopolies now?

the difference between current AMD and Nvidia GPUs isn't even that large if viewed from price/performance ratio... Comparing cards at similar price has AMD having slightly less performance while having significantly more GDDR memory. i still use an RTX3080 though, thankfully got one before the current craze started

The difference between AMD and Nvidia is _huge_ when you look at software support and drivers and etc. Part of this is network effects and part of it is just AMD itself. But the hard reality is I'd never buy AMD for compute, even if in specs it were better.

Just as a random anecdote, I grabbed an AMD 5700xt around when those came out (for gaming). Since I had it sitting around between gaming sessions, I figured I'd try to use it for some compute, for Go AI training. For _1.5 years_ there existed a showstopping bug with this, it just could not function for all of that time. They _still_ do not support this card in their ROCm library platform last I checked. The focus and support from AMD is just not there.

Re: ALiEn – a GPU-accelerated artificial life simulation program

#128

Post by mistermann »

tgv wrote: post #30
Cthulhu_ wrote: post #27

Earlier quoted context omitted.

It's more likely than you think: https://en.wikipedia.org/wiki/Simulation_hypothesis

There’s no evidence we live in a simulation, and a whole lot to suggest we aren’t.

That so many people believe they have knowledge of all the evidence suggests to me that something very strange is going on here.

This may not literally be a simulation, but it seems to behave like one in many ways.

Re: ALiEn – a GPU-accelerated artificial life simulation program

#129

Post by PaulDavisThe1st »

tgv wrote: post #111
p_j_w wrote: post #97

Earlier quoted context omitted.

>you only have to assume physics totally different from ours You don't have to assume anything is true if you don't want to, but if you want to consider whether we're living in a simulation, it's probably worth considering. >And what would be the point of this simulation that has been running for billions of years? First, it's billions of years in our time. Second, what's the point of Conway's Game of Life?

> You don't have to assume anything is true if you don't want to, but if you want to consider whether we're living in a simulation, it's probably worth considering. Which, for me, is a dead end. It's the same as assuming there's a God, except the moral implications are worse. > First, it's billions of years in our time. So, if billions of years of our time fly by like your average simulation run in "their" universe,…

Your ability to comprehend something is your shortcut to assessing how likely it is?

Re: ALiEn – a GPU-accelerated artificial life simulation program

#130

Post by ColonelPhantom »

raphlinus wrote: post #48
hyperluz wrote: post #44

Earlier quoted context omitted.

What would be the best alternative? OpenCL? Vulkan?

OpenCL is sadly stagnant. Vulkan is a good choice but not itself portable. There are frameworks such as wgpu that run compute shaders (among other things) portably across a range of GPU hardware.

In what way is Vulkan not portable? It runs on all operating systems (Windows 7+, Linux, Android, and Apple via MoltenVK) and all GPUs (AMD GCN, nVidia Kepler, Intel), and shaders (compute and rendering) are to my knowledge standardized in the portable SPIR-V bytecode.

WGPU is more portable, since it can use not only Vulkan but also other APIs like OpenGL and Direct3D 11, but Vulkan is already very highly portable for almost everyone with a computer modern enough to run anything related to GPU compute.

Post reply on HN