WHISPERS OF THE FUTURE

Summary from the TLDR Founders newsletter

LLMs are amazing, but they require immense amounts of hardware. For this reason, the software will evolve to have a hardware architecture where some of the model is processed on the computer and some in the cloud.