Four billion years in four minutes – Simulating worlds on the GPU • Hacker News Board

Re: Four billion years in four minutes – Simulating worlds on the GPU

#91

Post by mensetmanusman » Tue, Jul 30, 2024, 3:12 PM UTC

xwolfi wrote: post #39

shanxS wrote: post #2

Sometimes I wonder what it’d be like to live in simulated universe.

Well you are in a way, what is a simulation ? It's just a set of rules you follow that are simpler than the more complex environment that it runs in: I suppose if we could "see" "outside" the "universe", we'd understand that maybe our reality is very simple and limited compared to the "reality" outside. Maybe this would be true infinitely, or maybe the outside reality would be much more logical than ours, and we'd ac…

That concept of God is stupid, some are tautological, and some are experiential. Eg the concept of God is love (if you have faith in the concept and reality of love being real).

Re: Four billion years in four minutes – Simulating worlds on the GPU

#92

Post by exe34 » Tue, Jul 30, 2024, 3:40 PM UTC

netcraft wrote: post #83

exe34 wrote: post #65

Earlier quoted context omitted.

there's a short story in his anthology that is a sequel to diaspora. oceanic.

I had no idea it was related to diaspora. After reading only the first few paragraphs I can't wait to read it all. Thank you!

tbh a number of the other short stories could be considered prequels to diaspora.

Re: Four billion years in four minutes – Simulating worlds on the GPU

#93

Post by Retric » Tue, Jul 30, 2024, 3:47 PM UTC

LoganDark wrote: post #80

Retric wrote: post #66

Earlier quoted context omitted.

Infinite processing power, infinite storage, zero latency, etc is about as soft of a setting as it gets.

Is it? A new fundamental particle doesn't necessarily make it soft sci-fi. The story seems perfectly technical to me. Can't black holes already encode an infinite amount of information past the event horizon?

> Can't black holes already encode an infinite amount of information past the event horizon?

No, current physics has a bunch of hard limits on information.

> A new fundamental particle doesn’t necessarily make it soft sci-fi

Calling something a particle doesn’t matter here. It’s a fully formed computational device that happens to magically solve all problems.

Consider, how do you encode and read information from such a particle. Qubit’s are physical properties of something, such as spin. Is this particle supposed to have infinite properties which you can access with infinite precision? No, it’s just a magic macguffin that does whatever the author wants.

Re: Four billion years in four minutes – Simulating worlds on the GPU

#94

Post by MaxikCZ » Tue, Jul 30, 2024, 7:22 PM UTC

candiddevmike wrote: post #11

shanxS wrote: post #2

Sometimes I wonder what it’d be like to live in simulated universe.

Watch the show Pantheon sometime! No spoilers, just highly relevant.

That series blew me away, both seasons, every episode is just totally awesome. Hard to reccomend without spoiling.

Re: Four billion years in four minutes – Simulating worlds on the GPU

#95

Post by Andrews54757 » Tue, Jul 30, 2024, 8:26 PM UTC

Waterluvian wrote: post #22

disillusioned wrote: post #3

Only tangentially related, but the lovely (quite) short story "I don't know, Timmy, being God is a big responsibility" is fantastic and kind of hits on simulating worlds, in a sense: https://qntm.org/responsibility

I’m not sure I really know how to handle this. Do I just read the latest version linked at the top of the page? Or is the one offered by this like somehow more canonical or something? I’m not sure I’ve ever been faced with different versions of a fiction before. Just textbooks. I really love thought pieces like this if you refrain from poking holes in the logic or physics. I love the idea of a multiverse where they’r…

I read both. They are both great but I like the original better. The new edition is accessible but I think the original's prose is beautiful.

Re: Four billion years in four minutes – Simulating worlds on the GPU

#96

Post by kmoser » Tue, Jul 30, 2024, 9:13 PM UTC

fersaysrelax wrote: post #63

kmoser wrote: post #40

Back in 1996/1997 I worked on a CD-ROM game that simulated movement of the tectonic plates, as well as temperature, elevation, and precipitation, over millions of years. Amazing to see how evolution (heh!) of computing hardware and software have come so far in 28 years: https://www.kmoser.com/evolution/

I think you are the first person I come across that knows that game and you were an actual dev on it! So thank you! I played it when I was around 13(I think) and I loved it. The whole concept of simulating a planet and the animals on it was pretty much mind blowing for me at the time(it still is). I should have kept the poster that came with it...the different sapien evolutionary branches were really interesting.

Cool! It's admittedly an obscure game. My main contribution included the random generation and movement of the cratons. This involved mapping Cartesian to spherical coordinates but harder because each row of tiles was offset by half a tile's width from the rows above and below, like a brick wall.

We wrote an in-house editor that let us create and animate the different cratons, starting with the Pangaea configuration (basically all cratons stuck together as a giant continent) all the way through to how they are arranged now, to simulate how they moved historically. That was also difficult!

I also worked on the initial placement of some visual elements that dotted the landscape, and the serialization of the different data structures for loading/saving games.

After it shipped they gave each of us devs a couple of shrink-wrapped copies of the boxed final product but weirdly they were not the US version. I still have an (unopened) Portuguese version that probably has the poster.

Re: Four billion years in four minutes – Simulating worlds on the GPU

#97

Post by jonah » Wed, Jul 31, 2024, 12:04 AM UTC

kmoser wrote: post #40

Back in 1996/1997 I worked on a CD-ROM game that simulated movement of the tectonic plates, as well as temperature, elevation, and precipitation, over millions of years. Amazing to see how evolution (heh!) of computing hardware and software have come so far in 28 years: https://www.kmoser.com/evolution/

This is rad! I recently read Assembling California by John McPhee[1][2] - highly recommended BTW.

Anyway, I was musing that I'd love to see a 3D globe with a time slider that I could play around with and visualize the (projected of course) movement of the plates to get us to where we are now.

Do you know of anything like that?

[1] https://www.goodreads.com/book/show/19898.Assembling_Califor...

[2] excerpt: https://www.newyorker.com/magazine/1992/09/07/assembling-cal...

Re: Four billion years in four minutes – Simulating worlds on the GPU

#98

Post by kmoser » Wed, Jul 31, 2024, 3:19 AM UTC

jonah wrote: post #97

kmoser wrote: post #40

Back in 1996/1997 I worked on a CD-ROM game that simulated movement of the tectonic plates, as well as temperature, elevation, and precipitation, over millions of years. Amazing to see how evolution (heh!) of computing hardware and software have come so far in 28 years: https://www.kmoser.com/evolution/

This is rad! I recently read Assembling California by John McPhee[1][2] - highly recommended BTW. Anyway, I was musing that I'd love to see a 3D globe with a time slider that I could play around with and visualize the (projected of course) movement of the plates to get us to where we are now. Do you know of anything like that? [1] https://www.goodreads.com/book/show/19898.Assembling_Califor... [2] excerpt: https://ww…

Googling "pangaea animation" gives you many video demos of just that.

Re: Four billion years in four minutes – Simulating worlds on the GPU

#99

Post by LoganDark » Wed, Jul 31, 2024, 6:43 AM UTC

Retric wrote: post #93

LoganDark wrote: post #80

Earlier quoted context omitted.

Is it? A new fundamental particle doesn't necessarily make it soft sci-fi. The story seems perfectly technical to me. Can't black holes already encode an infinite amount of information past the event horizon?

> Can't black holes already encode an infinite amount of information past the event horizon? No, current physics has a bunch of hard limits on information. > A new fundamental particle doesn’t necessarily make it soft sci-fi Calling something a particle doesn’t matter here. It’s a fully formed computational device that happens to magically solve all problems. Consider, how do you encode and read information from such…

> No, it’s just a magic macguffin that does whatever the author wants.

As a general rule I consider this unfair. It's okay for science fiction to be set in universes that have different physical laws than our own. Hard or soft sci-fi has nothing to do with whether the rules match up to reality, only whether they're internally consistent and whether they are delved into in technical detail.

Sure, there is no known particle in our universe that could do this, nor is there any known way for such a particle to exist. Does that mean its existence in the story makes the sci-fi soft? Take a look at how it's described:

> They - by which we now refer to Tim, Diane, their eight colleagues, their two supervisors, four chemical engineers, six electrical engineers, the janitor, a countable infinity of TEEO 9.9.1 ultra-medium-density selectably-foaming non-elasticised quantum waveform frequency rate range collapse selectors and the single tormented tau neutrino caught in the middle of it all - represented the sum total of the human race's achievements in the field of quantum computing. Specifically, they had, earlier that week, successfully built a quantum computer. Putting into practice principles it had taken a trio of appallingly intelligent mathematical statisticians some 10 years to mastermind, and which only about fifty-five other people in the world had yet got a grip on, they had constructed an engine capable of passing information to and processing the responses from what could, without hyperbole, be described as a single fundamental particle with infinite processing power and infinite storage capacity.

Sure you could call it a MacGuffin, but it's not magic, it's clearly science (just within their universe). It's clearly explained to have been science. Maybe it's not explained in infinite detail, but it does not need to be. Nothing about this looks like soft sci-fi to me.

Nothing about qntm's other stories - Ra, Fine Structure, and a bunch of the short ones - seem like soft sci-fi to me either. I mean, seriously, Ra contains academic lectures[0] where the whole speech, parts of the equations, and the principles behind them are all actually explained to the reader in detail, not just glossed over. There is even an appendix where the actual in-universe mechanics behind the "magic" are explained.[1] And another one where a specific case of it is thoroughly considered and explained.[2] This is science. Fictional science, but that is what science fiction is. There's not really anything soft about it, imho.

[0]: https://qntm.org/know (warning: few chapters into Ra)

[1]: https://qntm.org/what, https://qntm.org/spells (many spoilers)

[2]: https://qntm.org/invisibility

Re: Four billion years in four minutes – Simulating worlds on the GPU

#100

Post by Bluestein » Wed, Jul 31, 2024, 7:15 AM UTC

"Four seasons in one day / Laying in the depths of your imagination / Worlds above and worlds below / Sunshine on the black clouds hanging over the domain ..."