{"id":39603,"date":"2025-07-16T20:43:56","date_gmt":"2025-07-16T12:43:56","guid":{"rendered":"https:\/\/www.1ai.net\/?p=39603"},"modified":"2025-07-16T20:43:56","modified_gmt":"2025-07-16T12:43:56","slug":"%e6%a0%b9%e6%9c%ac%e5%81%9c%e4%b8%8d%e4%b8%8b%e6%9d%a5%ef%bc%8cmeta-%e8%a2%ab%e6%9b%9d%e5%8f%88%e4%bb%8e-openai-%e6%8c%96%e8%b5%b0%e4%b8%a4%e5%90%8d%e9%a1%b6%e7%ba%a7%e7%a0%94%e7%a9%b6%e4%ba%ba","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/39603.html","title":{"rendered":"Meta has been revealed to have poached two more top researchers from OpenAI."},"content":{"rendered":"

July 16 (Bloomberg) -- According to Wired magazine today, citing multiple sources familiar with the matter, theOpenAI<\/a> Researcher Jason Wei is coming soonbecome a member of an alliance or union Meta<\/a> Newly established Superintelligence Laboratory<\/strong>.<\/p>\n

\"8ac9d495j00szhsnu0030d000v900k9p\"<\/p>\n

According to his personal website, Wei was involved in the development of the\u00a0OpenAI's o3 model and the Deep Research project<\/strong>He joined OpenAI in 2023. He joined OpenAI in 2023 after working at Google, where he focused on \"thought chaining\"-an approach designed to train AI models to solve complex problems step-by-step.<\/p>\n

During his time at OpenAI, Wei became very passionate about reinforcement learning. This way of training AI through positive and negative feedback has become a high-profile area of research, and one that Meta is focusing on as it recruits members of its superintelligence team.<\/p>\n

Another OpenAI researcher, Hyung Won Chung, will also be joining Meta, and multiple sources have confirmed that Wei and Chung's internal Slack accounts have been deactivated. Chung's personal website indicates that he has worked with Wei on a number of projects.Includes o1 models from Deep Research and OpenAI<\/strong>. His research focuses on reasoning and intelligences. According to Collingwood, the two worked together at Google and joined OpenAI at the same time.<\/p>\n

Several sources noted that Wei and Chung\u00a0Long-term cooperation and close relationship<\/strong>Meta has in the past.Solicitation of Research Teams by Entire Groups<\/strong>The precedent is set, for example, by the addition of three former Google researchers to the superintelligence team from OpenAI's Swiss office.<\/p>\n

In the past month, Meta has opened a new round of AI talent competition, offering up to $300 million over four years (note: the current exchange rate is about 2.154 billion yuan) in salary terms.Meta CEO Zuckerberg has issued an internal memo to clarify the company's AI strategic plan, and pointed out that the newly joinedresearch worker<\/a>Mostly from OpenAI<\/strong>.<\/p>\n

The war for talent is still heating up, and OpenAI is fighting back. Last week, OpenAI recently managed to poach four senior engineers from Tesla, xAI, and Meta.<\/p>\n

On Tuesday, Wei posted on social media about the \"life lessons\" he learned from reinforcement learning. He wrote, \"While living and training AI models, I've learned a lot.Imitation is a necessary first step<\/strong>But to surpass the teacher, one must dare toGo your own way.<\/strong>that take risks and reap rewards from the environment.\"<\/p>","protected":false},"excerpt":{"rendered":"

July 16 (Bloomberg) -- OpenAI researcher Jason Wei is joining Meta's new superintelligence lab, according to Wired magazine today, citing multiple sources familiar with the matter. According to his personal website, Wei worked on OpenAI's o3 model and the Deep Research program. He joins OpenAI in 2023 from Google, where he focused on \"thought chaining\" -- an approach designed to train AI models to solve complex problems step-by-step. During his time at OpenAI, Wei developed a passion for reinforcement learning. This approach to training AI through positive and negative feedback has become a high-profile area of research, and it's one of the areas in which Me<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[146],"tags":[297,190,7210],"collection":[],"class_list":["post-39603","post","type-post","status-publish","format-standard","hentry","category-news","tag-meta","tag-openai","tag-7210"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/39603","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=39603"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/39603\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=39603"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=39603"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=39603"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=39603"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}