“We’re releasing it initially with our lightweight model version of LaMDA. This much smaller model requires significantly less computing power, enabling us to scale to more users, allowing for more feedback.”
Google announced Bard to “trusted testers” with a public launch promised in “the coming weeks.” The model shipping is the cut-down version of LaMDA, picked because it’s cheap to run. Pichai reminds everyone that Google wrote the 2017 Transformer paper and “re-oriented the company around AI six years ago.” Six years of head start, and the public gets a promise while testers get the budget model.