Mira Murati, the former chief technology officer (CTO) at OpenAI, launched a new artificial intelligence startup on Tuesday, with about 30 top researchers and engineers from competitors, including Meta, Mistral and her last company. Her venture, Thinking Machines Lab, aims to build AI systems that incorporate human values and expand their application beyond current models.

Ms Murati joins a growing list of ex-OpenAI executives who have launched their own AI ventures, including those behind Anthropic and Safe Superintelligence. She has been raising funds from venture capitalists to support the new venture, Reuters reported in October.

Around two-thirds of Thinking Machines Lab’s team consists of former OpenAI employees, Reuters reported. Among them is Barret Zoph, a prominent researcher who left OpenAI on the same day as Ms Murati in late September and will now serve as the startup’s Chief Technology Officer.

Another big name is OpenAI co-founder John Schulman, who joins as Chief Scientist. Mr Schulman left OpenAI for rival Anthropic in August, citing his focus on “AI alignment” – a core priority for Ms Murati’s startup, which seeks to make AI systems safer and more reliable.

Thinking Machines Lab positions itself as an AI research and product company dedicated to making AI more understandable, customisable, and widely applicable. It aims to bridge gaps in AI knowledge and usability by developing systems that collaborate with humans and adapt to various industries. The company prioritises open research, sharing its work through papers, blogs, and open-source projects, the website said.

“While current systems excel at programming and mathematics, we’re building AI that can adapt to the full spectrum of human expertise and enable a broader range of applications,” the company said. Thinking Machines Lab also plans to share code, datasets, and model specifications to encourage external research on AI alignment.

Mira Murati, a key public face of OpenAI alongside CEO Sam Altman, resigned abruptly amid governance changes at the company. Before OpenAI, she worked at Tesla and augmented reality startup Leap Motion. She played a major role in developing ChatGPT, DALL-E, and Codex.

Mira Murati quit as the Chief Technology Officer of OpenAI on Wednesday. She’s the latest to join a growing list of high-profile exits from the company. Ms Murati’s decision comes after six years with the artificial intelligence pioneer behind ChatGPT.

In a post on X, Ms Murati described her time at OpenAI as “an extraordinary privilege” and her decision to leave as “difficult.” “There’s never an ideal time to walk away from a place one cherishes, but the moment feels right. I want to create the time and space to do my exploration,” she said.

I shared the following note with the OpenAI team today. pic.twitter.com/nsZ4khI06P

— Mira Murati (@miramurati) September 25, 2024

OpenAI CEO Sam Altman responded with a touching tribute to X, thanking her for her contributions and conveying his support. “It’s hard to overstate how much Mira has meant to OpenAI, our mission, and to us all personally,” the CEO wrote. “I feel tremendous gratitude towards her for what she has helped us build and accomplish.” 

He said he was “excited for what she’ll do next.”

I replied with this. Mira, thank you for everything.

It’s hard to overstate how much Mira has meant to OpenAI, our mission, and to us all personally.

I feel tremendous gratitude towards her for what she has helped us build and accomplish, but I most of all feel personal…

— Sam Altman (@sama) September 25, 2024

Mira Murati, 35, played a major role in developing ChatGPT and overseeing releases of image generator Dall-E and AI code generator Codex. Her exit comes after her brief tenure as interim CEO following Mr Altman’s ousting last November. 

Sam Altman announced Mark Chen as senior VP of research and Josh Achiam as head of mission alignment to ensure a smooth transition.

Mira Murati’s departure follows those of co-founders Greg Brockman (on extended leave) and John Schulman (joined rival Anthropic), as well as a product team leader and former Meta employee. Earlier this year, co-founder Ilya Sutskever exited OpenAI after a boardroom dispute. Ms Murati’s exit now leaves only two of its 11 founders remaining. 

The shakeup comes as OpenAI releases its new Strawberry AI model, designed to improve “thinking” in generative AI chatbots. The company aims to address “hallucinations” – persuasive but incorrect content – and provide more accurate responses.

OpenAI chief technology officer Mira Murati on Wednesday said she is leaving the company behind ChatGPT, adding to a growing list of high-level departures. Murati called her more than six years at the San Francisco-based company “an extraordinary privilege” and described her decision to leave as difficult in a post on X, formerly Twitter.

“There’s never an ideal time to walk away from a place one cherishes, but the moment feels right,” Murati said in the post.

“I want to create the time and space to do my own exploration.”

OpenAI chief Sam Altman replied to Murati’s post with his thanks for what Murati has helped the company build, and promised details on a transition plan soon.

Murati is the latest influential member of the OpenAI team to step away from their position.

Co-founder Greg Brockman is on an extended leave of absence and fellow OpenAI founder John Schulman left for AI sector rival Anthropic, according to a report in The Information.

A product team leader who OpenAI had lured over from Meta also left, the report said.

OpenAI co-founder Ilya Sutskever exited OpenAI early this year in the wake of a boardroom battle that saw Altman temporarily ousted from the company.

OpenAI earlier this month released a new series of artificial intelligence models designed to spend more time thinking — in hopes that generative AI chatbots provide more accurate and beneficial responses.

The new model, nicknamed Strawberry, is designed to tackle complex tasks and solve more challenging problems in science, coding and mathematics — something that earlier models have been criticized for failing to provide consistently.

Unlike their predecessors, these models have been trained to refine their thinking processes, try different methods and recognize mistakes before they deploy a final answer.

OpenAI’s push to improve “thinking” in its model is a response to the persistent problem of “hallucinations” in AI chatbots.

This refers to their tendency to generate persuasive but incorrect content that has somewhat cooled the excitement over ChatGPT-style AI features among business customers.

The new release came as OpenAI was raising funds that could see it valued around $150 billion, which would make it one of the world’s most valuable private companies, according to US media.
 

(Except for the headline, this story has not been edited by NDTV staff and is published from a syndicated feed.)