Generative thermodynamic computing produces images with zero neural network calls
mikeflache · x · 2026-09-24
Researchers demonstrated "Generative Thermodynamic Computing": a physical thermodynamic system that generates images without a single neural network call. Current image generators rely on billions of parameters and millions of dollars in GPU compute to mathematically denoise static. Instead of coding the denoising, this approach lets physical matter evolve over time according to the laws of physics—information is stored in the physical system rather than software weights. If engineered at scale, it could dramatically cut the cost of generation.
More from Multimodal
- Ling 3.0-flash-VL lands in llama.cpp: 124B MoE with only 5.5B active params — jacek2023 · 2026-09-24
- Full Suno prompt shared: dusty country trap with a 7-year-old Southern girl vocal — techhalla · 2026-09-24
- NoSpoon, a $20-for-5-Minute AI Microdrama Tool, Shuts to the Public This Month — Kyrannio · 2026-09-24
- Fractal Bioluminescent Bloom: a Reusable Prompt Template for Deep-Sea Glow Aesthetics — LudovicCreator · 2026-09-24
- Pika launches API Club aggregator claiming up to 88% cheaper gen-media pricing than Fal and Runway — Kyrannio · 2026-09-24
- Midjourney Preps Major Create Overhaul, Thinking Mode, New Edit and Niji Models — LudovicCreator · 2026-09-24