Deep research agents, like OpenAI, Gemini, or Claude's deep researchers, are extremely powerful to gather information, surf the web, and compile reports.

It has become a daily habit of mine to spin up multiple deep researchers to help gather knowledge and insights, but how do they work under the hood?

In this talk, we will go through how DR agents evolved, from chain-of-thoughts, STORM, and finally explore the more modern approaches to let AI explore A LOT of information accurately. We will even try to build some of them together!