Just like working with any other large language model, prompt engineering is an important part of shaping the input and output of a model to achieve your specific task. When working with generative AI tools, however, the explicit control over the query sent into the model is often abstracted away. But, that doesn’t mean you can’t still shape the input - let’s discuss how to do this with Github Copilot specifically.
Within a few weeks of Github Co-Pilot being released, software development companies were touting its success as a software development tool, claiming savings of up to 56% on productivity gains. Google has been recently quoted saying “over 25% of our code is now written by AI”. What are these companies doing (or doing differently) to see such huge returns on their generative AI tool investment? Let’s dive in!
Extending some of our findings in conversational AI prompting, I’ve made a simple app that uses an OpenAPI spec file to walk an end-user through making an API call based on the available routes and request body defined within the specification. With this, we could theoretically support any well-defined API request!
As promised, this post will break down how I implemented a role playing game where several independent AI agents take the role of players (rather than the game master) using SemanticKernel and managing the game state through the ChatHistory object. I’m using the popular Alice is Missing role playing game published by Hunters Entertainment, which is a text-based system that focuses on interactions between players and doesn’t require a game master.
Last week I worked on a side project inspired by some whitepapers I read on using Generative AI to play social games. I thought “Hey, why don’t I do something similar with the social RPG ‘Alice is Missing’?” I soon ran into some limitations with the AgentGroupChat module that represents the multi-agent chat workflow that is currently in beta within Semantic Kernel that I’d like to discuss.