Using the physical traits of cats to initialize AI models works better than using standard math.
April 1, 2026
Original Paper
Schrödinger's Seed: Purr-fect Initialization for an Impurr-fect Universe
arXiv · 2603.29115
The Takeaway
Researchers mapped the mass, coat patterns, and 'name complexity' of 21 domestic cats to create random numbers for deep learning, outperforming industry-standard methods by 2.5%. It suggests that the random biological messiness of biology might be better for training AI than arbitrary computer-generated numbers.
From the abstract
Context. Random seed selection in deep learning is often arbitrary -- conventionally fixed to values such as 42, a number with no known feline endorsement. Aims. We propose that cats, as liminal beings with a historically ambiguous relationship to quantum mechanics, are better suited to this task than random integers. Methods. We construct a cat-driven seed generator inspired by the first Friedmann equation, and test it by mapping 21 domestic cats' physical properties -- mass, coat pattern, eye