One of the creative minds behind OpenAI’s next big leap in AI is Yash Kumar, a Member of Technical Staff at the company and the product lead on ChatGPT Agent.
During a demo and briefing with The Verge, Yash and research lead Isa Fulford discussed the capabilities of the newly launched tool. He explained ChatGPT Agent has access to “an entire computer,” not just the browser.
ALSO READ| Gemini for Google Workspace is what ChatGPT wishes it could be
Who is Yash Kumar?
Yash, who is of Indian origin, studied Computer Science at IIIT Hyderabad, one of India’s top engineering institutes. He joined OpenAI in November 2023 and now works out of the company’s San Francisco headquarters, per LinkedIn.
ChatGPT Agent, which the company describes as a tool that can complete tasks using its own “virtual computer,” was built on a model OpenAI specifically developed for this purpose. The tool can do everything from checking a user’s calendar and summarizing meetings to helping plan a family breakfast.
Yash told The Verge that they are now working on “optimising for hard tasks” so users can have a smooth and seamless experience.
{{/usCountry}}Yash told The Verge that they are now working on “optimising for hard tasks” so users can have a smooth and seamless experience.
{{/usCountry}}“Even if it takes 15 minutes, half an hour, it’s quite a big speed-up compared to how long it would take you to do it,” Fulford noted. “It’s one of those things where you can kick something off in the background and then come back to it.”
Before ChatGPT Agent takes any action, like sending off an email or booking a reservation, it makes sure to check in with you first, Fulford added. Because the model powering this new tool is more capable than previous versions.
ALSO READ| ChatGPT isn’t disrupting Google Search just yet, report finds
The company explained that it’s turned on special safeguards originally designed for scenarios involving “high biological and chemical capabilities.” OpenAI also clarified there's currently no “direct evidence that the model could meaningfully help a novice create severe biological or chemical harm.”
Notably, back in May, Anthropic activated similar protections when it rolled out Opus 4, one of the more advanced versions of its Claude model.