RUNNING LOCAL MODELS IS GOOD NOW

Summary from the TLDR newsletter

Local models may still be slow and resource-intensive, but they can run many tasks competently and can help save on costs.