Home solar powers training; households earn a share of the model
LIVE from NEARCON Day 1 In SF (The Rollup)
“If training can go inside the house, it can tap into my solar, where my marginal cost of energy is like minimal, like almost zero… I have a lot of excess solar — using the excess solar to do some part of the training… in exchange I get back some token, some representation of the model, and when the model goes into inference and starts making money, I get some money out.” — 03:19:24
Context: Home GPUs earning income from distributed training runs; concedes it’s less efficient than centralized data centers (“nothing that can beat physics”) but argues it’s excess compute at near-zero marginal energy cost. Names little-known distributed-training companies (name garbled in captions).