# Turning Predictability into a Revolution By:: [[Brian Heath]] 2024-04-25 If one has used an AI Large Language Model (LLM) like ChatGPT, one will likely be very impressed initially and then find areas where it struggles. This is normal with most new technologies. There is the hype, and then reality sets in. Once one has seen the reality of the technology, one can put it to work in the right ways. Many are finding that feeding the proper context into an AI LLM produces remarkable results. This is called Retrieval Augmented Generation (RAG). Essentially, one is leveraging the language probability model of the LLM and feeding it relevant context to customize the response (the RAG component). As a young professional, one learns the "correct" way to compose a business email and then inserts the appropriate topic and context. The same is true for the combination of RAG and LLMs. LLMs know the "correct" way to construct a business email, and RAG provides the relevant content. This is how most knowledge work is done today. Knowledge workers spend years learning the "correct" ways to send emails, build presentations, write code, and write reports. Then, they apply these communication and work models to their current work context. Thus, it should be little wonder that we've figured out how to mimic this process within computational frameworks effectively. Yet, its degree of effectiveness and efficiency still baffles us. While the hype will likely die down, there are deeper insights to be gained beyond the shininess of the new technology. In particular, LLMs combined with RAG shows us how predictable and uncreative work has become. At their core, LLMs are just probability engines and only useful in situations where words and phrases are predictable. Almost all of the knowledge work today is entirely predictable. We've ventured so far down the rabbit hole that we don't even realize how repetitive we are. We've gotten so wrapped up in our imaginary world of meaning, boxed in by the "right" way of doing things, that we've failed to realize how robotic we've become. LLMs are showing us just how true this is. In an ironic twist, our robots are showing us a way back to a better way of being human. We must embrace this opportunity to change how we work, organize, and make meaning for the betterment of ourselves and society. #### Related Items [[Artificial Intelligence]] [[Work]] [[Creative]] [[Knowledge Work]] [[Models]] [[Society]] [[Future]] [[Technology]]