It turns out that the 18th-century English had the same BVLOS problems we do now. But with boats. Specifically, big boats.

A vision encoder is a necessary component for allowing many leading LLMs to be able to work with images uploaded by users.