Discover and read the best of Twitter Threads about #Action2023
Most recents (1)
🧵 1/
Seems that training for all AI systems needs to start at a set starting point of a specific topic in order to allow that topic to act as a tuner that can add to the overall collective system intelligence of some topic that is in the same realm or parallel to another topic
Seems that training for all AI systems needs to start at a set starting point of a specific topic in order to allow that topic to act as a tuner that can add to the overall collective system intelligence of some topic that is in the same realm or parallel to another topic
🧵 2/
Example: Person w/a plumbing issue, people would start at a LLM chat interface specific for plumbing, and it would be trained and have input coming from that specific topic. It also seems like there could also be some type of expert 'contributor' input vs 'asker' output.
Example: Person w/a plumbing issue, people would start at a LLM chat interface specific for plumbing, and it would be trained and have input coming from that specific topic. It also seems like there could also be some type of expert 'contributor' input vs 'asker' output.
🧵 3/
The contributor would be an expert plumber who could be tested by the LLM chat interface and upon passing a specific level of competency (would need a verification process the contributor was not just copy/pasting from Google or using the LLM to copy/paste answers)
The contributor would be an expert plumber who could be tested by the LLM chat interface and upon passing a specific level of competency (would need a verification process the contributor was not just copy/pasting from Google or using the LLM to copy/paste answers)
