Something neat I watched in a cringe-worthy demo from Microsoft was having a multi-agent toolchain to do some more precise reasoning. In this context, a problem that requires input from several different actions are segregated off into “agents”, some requiring LLMs, some not, and a “conductor” agent incorporates all the input from the various agents and decides if the conclusion reached satisfies the problem requirements - if it doesn’t, we continue working until the problem is solved.

A fun project I’ve been playing with (like so many other developers working with generative AI), is a game master that relies on generative AI to create content. The potential for a text and image-based large language model to create creative content has already been demonstrated multiple times (and is a subject of heated debate), but I’ve always thought that harnessing that ability within a game framework has a lot of potential, and I’ve started work on a project to prove that out.

Let’s be honest - there isn’t really a lot of tooling out there for C# developers when it comes to generative AI. It seems the entire space has revolved around the Python language (and with good reason). All the cool tooling for fine-tuning, running models locally, and response filtering seem to require Python if you want to get into the weeds. The sole exception to this seems to be Semantic Kernel, Microsoft’s answer to “how do we use generative AI in C# code?” I haven’t used it extensively, but I have started to play around with it, and one feature immediately caught my eye around their management of Chat Completion.