"Research Assistant": Exploring UXs Besides Chat

LangChain has introduced a new "Research Assistant" template, heavily inspired by the GPT Researcher team, for helping with research tasks. The main focus of this template is to provide long-running, general, non-chat applications that can perform complex reasoning tasks and generate high-quality outputs. Unlike chat-based LLM applications, which often struggle with latency expectations and require humans in the loop, this new template allows users to ask the LLM to do a large amount of work while still resulting in something that is not expected to be perfect. The "Research Assistant" template can generate code files, respond to GitHub issues, and write full research reports, making it one of the most useful applications seen so far.