three body problem solved?

[xevious2]

https://arxiv.org/abs/1910.07291

"fixed computational cost and up to 100 million times faster than a state-of-the-art solver"

So does this mean we can get more accurate physics now?

[Funkenhausen]

i loaded the earth and 5000 moons stress test and it took 90% of RAM after a few minutes, then crashed. i am still waiting for the US2 to go back to using CUDA cores for calculations.

[Jar]

It's really amazing and interesting work, but it unfortunately doesn't mean much for Universe Sandbox. The scope of what it can be used for is fairly limited and doesn't make sense for interactive sandbox simulations like Universe Sandbox. Here's a relevant quote from an article about the paper[www.livescience.com] that explains this:
"Neural networks must be trained by being fed data before they can make predictions. So the researchers had to generate 9,900 simplified three-body scenarios using Brutus, the current leader when it comes to solving three-body problems."