Meta hires key OpenAI researcher to work on AI reasoning models

Meta has appointed a highly influential OPENCE researcher to work on her AI rational models, trapit bansal The company’s new AI superintelization unitA person familiar with the subject tells TechCrunch.

OpenAI spokesman Kayla Wood confirmed to TechCrunch that Bansal had left the Opina. Bansal LinkedIn Page Saying that he left the open in June.

Bonsal has worked in the OpenAI since 2022 and co-founder Elias Sutskver was the main player of the company’s work on learning reinforcement. OpenAI’s first AI argument model, and in 1 he is listed as a Foundational contributor.

Bansal Meta’s AI Superintelligence Lab may give an important lift, as well as leaders as well The scales have your CEO Alexandra Wang And is Looking to add Former Github CEO is Nat Fredman and Daniel Gross, co-founder of Safe Superintelligence. Bansal Meta can help develop a Frontier AI argument model, which is competitive with industrial-leading technologies like Openai and 3 or DEPSK R1. Currently, Meta does not publicly provide an AI argument model.

In recent months, Mark Zuckerberg Meta has been in the appointment to create the new AI team, $ 100 million to offer compensation package Top researchers who joined his organization. It is unclear what Bansal proposed to join the agreement.

Nevertheless, it seems that Zuckerberg has been successful in achieving the Top AI research talent.

Three former OpenAI Researchers – Lucas Bayer, Alexander Colosnikov and Ziaoua Jhai – have joined the Mater AI Superintelism team in recent weeks, Wall Street Journal Report on Wednesday. Bansal will join them, ex -Google Dipmind Researcher Jack Roy and Startup Sesame, Johan Shalkwick’s ex -machine Learning Leader, join them BlombergThe

In order to further fill its new AI unit, Zuckerberg tried to report Achieve Startups with Heavy Hit AI Research LabsSuch Superintelies of Sutskver, Mira Murarati Thoughts of machine labes, and confusionThe However, these discussions never moved to the final stage.

Sam Altman, CEO of OpenAI at a recent podcast, said that Meta was trying to give her top talent in her startup, but claimed that ”None of our best people He decided to take him. ”

Meta did not agree to comment.

AI rational models present a key zone to be correct for the Mater AI Superintelism team. Last year, OpenAEA, Google and DEPSEC have sent high performance AI rational models that have pushed the limit to what software can do. AI models have been successful in improving AI labs in benchmark and real-world functions, using extra time and computing resources, using extra time and computing resources.

Mater AI Superintelise Labs can become a key internal group that gives strength to products throughout the company, much like Google’s Depthmind Unit. Meta has an ambitious attempt Create AI agents for business Clara Shih under the former AI Salesforce CEO. In order to create a competitive agent, Border AI rational models need to be developed to give Meta their strength.

Bansal and other key AI researchers are hoping to move forward, with the addition of AI researchers. It may be difficult to give OpenAI has planned to publish an open AI argument model In the upcoming weeks – an offer that can add more stress to Mater open AI offers.

This story was updated after publishing with more detailsThe