Closing remarks
We have looked at various aspects of emerging LLM frameworks and how we can build better solutions using these frameworks. While this technology is exciting and opening new doors for how we look at and solve problems, it is still in its nascent stages. It requires a lot of processing power and may not have the service-level agreement (SLA) response times we might be looking for. While we might get excited about solving all the problems using this approach, it’s important to evaluate whether trying to solve a given problem really needs to leverage these technologies or whether we can solve them more reasonably and effectively. For example, many people are excited about LLM capabilities and trying to generate data models (be it graphs, SQL, or other) by describing a problem. This can be a double-edged sword. If we are not familiar with the data we are working with, we may not be able to validate the model generated. If we are very familiar with the data, we might...
