Good ideas and conversation. No ads, no tracking. Login or Take a Tour!
A Roomba traverses the world gathering IRL data. The difference is the Roomba can incorporate that data, the LLM can't - the basic make up of "large language models" is "giant dataset" + "careful, empirical training" = "pantomime of intelligence." The response of the model must be carefully, iteratively tuned by technicians to get the desired results. It's that iterative tuning that makes it work, and it is exquisitely sensitive to context.