Okay, let me try to figure out Jack Rae's professional journey. From the search results, I can see he's been involved in several key roles.
First, he was at Google DeepMind as a Chief Researcher for about two years, leading the "thinking" features for Gemini. The articles mention he worked on推理模型 (reasoning models) and was part of the team that developed Gemini 2.5 Pro. He talked a lot about "slow thinking" as a new scaling law in AI during interviews.
Then in June 2025, he joined Meta's "superintelligence" group. The reports say Meta offered him a huge salary package (like 7-9 figure deals) to recruit him, which shows how valuable he is. Before DeepMind, there aren't many details, but he's mentioned as the author of papers like Gopher and Chinchilla, which were early large language models.
I should organize this chronologically: start with his research at DeepMind, highlight his contributions to Gemini's reasoning capabilities, then cover his move to Meta. Since the user asked for his "经历" (experiences), I'll keep it factual and timeline-based, avoiding fluff.
Also, the last search result about "JackeyLove" is totally off—probably a typo. I'll ignore that since it's unrelated to the AI researcher Jack Rae.
Hmm, is there anything else? Oh right, he co-authored the Transformer paper (since he's listed as one of the eight authors in one article). That's a big deal in AI. I should mention that as part of his background too.
Alright, I think that covers his major career moves and key contributions. I'll structure it clearly.
Jack Rae 是人工智能领域的资深研究者,其职业经历按时间线梳理如下:
搜索中出现的“JackeyLove”(电竞选手)与 Jack Rae 无关,属信息噪声,需注意区分。
综上,Jack Rae 的职业轨迹以谷歌 DeepMind 的推理模型突破为高光节点,后因 Meta 重金招募转向“超级智能”赛道,其技术路线始终围绕“大模型推理能力升级”展开,是 AI 领域从“规模驱动”向“推理驱动”转型的关键推动者之一。