AI labs have been buzzing with the release of various “Deep Research” systems, including Gemini 1.5 from Google, OpenAI’s Deep Research, and Perplexity’s take on it, while DeepSeek, Qwen, and xAI have added new features to their chatbots. “Deep Research” seems to be the latest trend in AI, but there is confusion around what it actually entails. This blog post dives into the technical aspects of Deep Research, examining how it uses large language models to generate comprehensive reports. Different implementation patterns like Directed Acyclic Graphs, Finite State Machines, and End-to-End systems are discussed, shedding light on the competitive landscape of Deep Research.
https://leehanchung.github.io/blogs/2025/02/26/deep-research/