the next frontier for AI-native products is for them to autonomously make decisions and take actions.
but it's hard! traditional RL methods are too brittle to handle new environments. luckily, LLMs have been shown to easily adapt and generalize to all sorts of tasks.
in this project, i built a simple prototype to show what a personal assistant might look like in the age of LLMs. language models serve as the glue between your intent and the different digital tools/services you use everyday -- translating messy, unstructured natural language commands into the right interface calls and vice versa.
increase agency by letting humans do what they do best, and let everything else be automated.
Log in or sign up for Devpost to join the conversation.