Fulford Island, an Openai researcher, had the heart that Deep Research would be a success even before its launch. Fulford helped create the AI agent who explores the website autonomously, deciding for himself in which links click, what to read and what to collect to prepare a detailed report.
At first, OpenAi put Deep Research available to internal users. Fulford says that, every time he stopped working, he received multiple consultations from colleagues anxious to recover it: “The number of people who sent me text messages excited us.” Since it was made available to the public on February 2, Deep Research has proven to be a success among many users outside the company.
Deep Research has positive reviews
“Deep Research has written 6 reports so far. It is excellent. Congratulations to the people behind,” Patrick Collison, CEO of Stripe published, in x a few days after the product was launched. Dean Ball, a researcher at the George Mason University Specialized in AI policy, also shared his emotion: “Deep Research is the product of AI who has really managed to get a significant part of the Washington DC political community.
Deep Research is part of the Chatgpt Pro plan, which costs $ 200 a month. This receives a consultation, such as: “Write a report on the Massachusetts medical insurance sector” or “Tell me about Wired’s coverage about the government efficiency department”; And then prepare a plan, looking for relevant websites, analyzing its content and deciding what information deserves a deeper investigation. After exploring for a few minutes, it synthesizes its conclusions in a detailed report, which can include appointments, data and graphics.
Many tools currently qualified as AI agents are essentially Chatbots connected to simple programs without much sophistication. Deep Research’s own model goes through a type of artificial reasoning before devising a plan and advancing in each step. The model offers details of this reasoning behind your research in a side window.
“Sometimes it is like ‘I have to back down, this does not seem so promising.’ It is great to read some of those trajectories, just to understand how the model is thinking,” says Josh Tobin, another OpenAi researcher who participates in the construction of the AI agent.
Undoubtedly, Sam Altman’s company sees Deep Research as a tool that could assume more office work. “This is something we can expand,” says Tobin, adding that the agent could be trained to perform specific jobs. For example, an agent with access to the internal data of a company could prepare a report or a presentation quickly. Tobin describes that the long -term objective is “to build an agent that is not only good by preparing reports through web searches, but also competent in many other types of tasks.”
#Chatgpts #deep #research #function #powerful
