Definitely not a videogame. I wouldn't want the responsibility of allocating that quantity of energy and resources. I think ChatGPT and its brethren are fascinating, amazing, and useful, but your question makes me think probably nobody should have that compute power. Maybe it's hubris to think one could responsibly use it. (Now I feel uncool for failing to have fun with your question.)
>There are several tightly coupled parts of a simulation game that must be designed closely together: the simulation model, the game play, the user interface, and the user’s model.
>In order for a game to be realizable, all of those different parts must be tractable. There are games that might have a great user interface, be fun to play, easy to understand, but involve processes that are currently impossible to simulate on a computer.
>There are also games that are possible to simulate, fun to play, easy to understand, but that don’t afford a useable interface: Will has designed a great game called “Sim Thunder Storm”, but he hasn’t been able to think of a user interface that would make any sense.
He also wanted to do SimTapeWorm, but just could not talk the marketing department into that. And I imagine the user interface would be quite slippery.
Aren't those simulations run on clusters of similar size (or atleast within an order of magnitude) and an agreeably responsible use of computational resources?
Feels like there should be an xkcd about this. I asked Claude to write one, but it wasn't very funny. Actually, Claude agrees with you that the energy used for training AIs is mere pocket change compared to climate simulations. (Can I trust Claude? Seems far from disinterested.)
Definitely not a videogame. I wouldn't want the responsibility of allocating that quantity of energy and resources. I think ChatGPT and its brethren are fascinating, amazing, and useful, but your question makes me think probably nobody should have that compute power. Maybe it's hubris to think one could responsibly use it. (Now I feel uncool for failing to have fun with your question.)
From "Will Wright on Designing User Interfaces to Simulation Games (1996) (2023 Video Update)":
https://donhopkins.medium.com/designing-user-interfaces-to-s...
>There are several tightly coupled parts of a simulation game that must be designed closely together: the simulation model, the game play, the user interface, and the user’s model.
>In order for a game to be realizable, all of those different parts must be tractable. There are games that might have a great user interface, be fun to play, easy to understand, but involve processes that are currently impossible to simulate on a computer.
>There are also games that are possible to simulate, fun to play, easy to understand, but that don’t afford a useable interface: Will has designed a great game called “Sim Thunder Storm”, but he hasn’t been able to think of a user interface that would make any sense.
He also wanted to do SimTapeWorm, but just could not talk the marketing department into that. And I imagine the user interface would be quite slippery.
What about climate/weather models?
Aren't those simulations run on clusters of similar size (or atleast within an order of magnitude) and an agreeably responsible use of computational resources?
Would the climate model take its own climate impact into account?
1 reply →
Feels like there should be an xkcd about this. I asked Claude to write one, but it wasn't very funny. Actually, Claude agrees with you that the energy used for training AIs is mere pocket change compared to climate simulations. (Can I trust Claude? Seems far from disinterested.)