The next step up the value chain would be to train the AI engine in a specific niche area. Open AI, etc. used huge data bases to create their basic broad functionality and that process is very expensive and creates awe inspiring but less than perfect results.
It is possible to create much more accurate results by training the general AI with relatively small amount of niche specific data. This is definitely an area to build a business or career around. ~Gary Gomes 12/22
ChatGPT is a large language model chatbot developed by OpenAI based on GPT-3.5. It has a remarkable ability to interact in conversational dialogue form and provide responses that can appear surprisingly human.
Large language models perform the task of predicting the next word in a series of words. It’s a revolutionary technology because it’s trained to learn what humans mean when they ask a question.
Reinforcement Learning with Human Feedback (RLHF) is an additional layer of training that uses human feedback to help ChatGPT learn the ability to follow directions and generate responses that are satisfactory to humans. ~Roger Montti in SEJ
Our thought experiment seeks to find a way of determining what, of all that great teachers do, can be scaled up and evolved as online tools that work for students world wide, or at least in the US. Say as part of our experiment that we identify the above as a Niche Area.
What would the needed data set for that niche training consist of? Would we need to drill down into specific “teacher’s of the year” practices to mine relevant data? Could ChatGPT find some of that for us? What sort of API might we need to evolve? This seems like a huge question for which expert insight might be required even to analyze the possibilities, much less realize them.
ChatGPT: What Is It & How Can You Use It?
There currently exists, parallel to Large Language Chat Bots such as OpenAI Chat GPT, an evolving “Knowledge Management” discipline, or niche.
This capability is also called “Second Brain”, and it attempts to both accumulate information content, and to structure that content in such a way that it is readily available when needed. Which is in part the point or usefulness of ChatGPT.. Might we might also call that Knowledge Management, or even a Second Brain.?
But while similar they also work in different ways and produce different results. See following post on Knowledge Management and Second Brain tools.