Here’s another SmoothLife article referenced by this article, with a really captivating long youtube video of an example SmoothLife evolution
Great video. They take on an organic look reminiscent of cells combining and dividing.
Wow. I wonder if there are still interesting structures in this version like a glider. Hmmm.
Yes, you can see the equivalent of a glider in the gif. It’s a lopsided donut.
The fact that they’re already using continuous operations gives me an project idea for whoever wants to one-up it. They could build it as a dedicated, analog computer. The game doesn’t have a lot of functions or value range either. Should make it easier. It will then run this in real time on real values using hardly any circuitry or watts. Could even be turned into a toy appliance if combined with a screen. Something to look at like the old, lava lamps.
An analog electronic computer can model small (whole!) numbers of continuous variables and relations between them, but I’ve never seen one that can actually model a continuous space of variable points. For that we’d use real chemistry: the most famous are the B-Z reactions, but the generic term is reaction-diffusion system. Of course, the chemicals are made up of discrete molecules, but their motion is continuous at least down to the Planck length!
This Python thing isn’t properly continuous either, of course: it’s using floating point numbers for the values, and a numpy mrange for a 2D cell grid. The neighborhoods are circular discs which cover a relatively high number of cells, thereby approximating a continuous space. It’s a little ironic, because Conway explicitly described Life as a discretization of a differential equation system.
If you think this stuff is cool, check out some of the work on “artificial chemistry” systems composed of discrete particles which move and interact in a continuous space. My personal favorite examples are from Sayama’s “swarm chemistry” experiments, but I haven’t really kept up with the field.