Post

Within the next 5 years most people on the planet will run inference locally on their device using mostly open source and distilled models. The massive scale OpenAI and nVidia megaprojects being developed likely only used for training and maybe government scale science, infra, and war related tasks