SpeakingAlthough I haven't put much thought into the language out problem (transforming ideas into written statements), something occurred to me today: Any time a person, or an AI, goes to say something, they have the opportunity to try and parse what they're about to say using their language in layer. If the resulting meaning is the same or similar enough to the idea they were trying to express, they have the green light to go ahead and speak. Otherwise, the speaker needs to go back to the drawing board and figure out how else to express their idea. |