Claude Fable 5 has impressive spatial reasoning capabilities that are immediately relevant for engineering and design: In this example I gave Claude Fable 5 a single photo of a hierarchical mesh torus (image generated using a text-to-image model); and in one shot, no iteration, it built a full interactive... 3D simulation: ~1,400 nodes & 4,400 fibers, realistic mass–spring physics you can grab, compress, twist… plus strain-based sonification so you can 'hear' the structure vibrate. It inferred the complete 3D topology from a partial 2D view. ⤵️show more

Here is another example, this time modeling fracture (irreversible deformation) of a foam material based off an image ⤵️

Here is the image if others want to use it:

I made this and many other images using our series of fine-tuned bio-inspired imagegen models:

learned physics incoming

Simulated virtual worlds that we can use to experiment in, explore, optimize and hillclimb with.

you can use it to decompose your data into @mathildepapillo's more robust topological data representations for higher-order message passing! @TAGinDS would be fun to come to for a day [in boston]!

might be wrong but a torus is one of the more geometrically regular shapes you could pick for this. curious if it holds up with something less symmetric, like a branching scaffold or irregular lattice?

Might be useful for controlling plasma in TOMOHAWK.

Inferring a full 3D topology from a partial 2D image and making it physically interactive with sound in one shot 😭 this is way beyond game demos

Finally getting to a point of real world hardware design with AI

Yes a threshold has been crossed

Mmmm, donut 🍩

Yes🙌

demo link?

Finally found someone else thinking about the sonic spectrum as well. Want to hear a black hole? :P

import numpy as np import matplotlib.pyplot as plt from mpl_toolkits.mplot3d import Axes3D from matplotlib.animation import FuncAnimation fig = plt.figure(figsize=(12, 10), facecolor='black') ax = fig.add_subplot(111, projection='3d') ax.set_facecolor('black') ax.axis('off') def draw_torus(frame): ax.clear() ax.set_facecolor('black') ax.axis('off') t = np.linspace(0, 2*np.pi, 600)