
上周二,前OpenAI首席技術(shù)官米拉·穆拉蒂正式分享了其新成立的人工智能初創(chuàng)公司的細節(jié),宣布了一個由眾多前同事組成的管理團隊,并公布了解決最先進的“前沿”人工智能系統(tǒng)存在的“關(guān)鍵差距”的計劃。上周二發(fā)布的一篇博文中披露了有關(guān)Thinking Machine Lab的這些細節(jié),距離穆拉蒂離開OpenAI已大約六個月之久,同時也揭開了人工智能行業(yè)最受矚目的項目之一的面紗。
這家從事人工智能研究與產(chǎn)品開發(fā)的初創(chuàng)公司將其使命定義為讓“人工智能系統(tǒng)得到更廣泛的理解、更具可定制性和通用性”。
該博文稱:"關(guān)于這些系統(tǒng)的訓(xùn)練方法的知識集中在頂級研究實驗室中,這不僅限制了人們關(guān)于人工智能的公開討論,還限制了人們有效使用人工智能的能力。”
OpenAI聯(lián)合創(chuàng)始人約翰·舒爾曼(John Schulman)將擔任這家初創(chuàng)公司的首席科學(xué)家。據(jù)《財富》雜志此前率先報道,舒爾曼在加盟Anthropic僅5個月后便離職,隨后加入了穆拉蒂的新公司。曾任OpenAI研究副總裁的巴雷特·佐夫(Barret Zoph)將擔任Thinking Machines Lab的首席技術(shù)官。
在上周二公布的29名Thinking Machines Lab員工中,至少有7人曾就職于OpenAI。喬納森·拉赫曼(Jonathan Lachman)曾任特別項目負責人;翁麗蓮(Lilian Weng)曾任副總裁;盧克·梅茨(Luke Metz)、山姆·施萊弗(Sam Shleifer)和斯蒂芬·羅勒(Stephen Roller)曾擔任研究科學(xué)家。其余員工包括來自Meta、谷歌DeepMind、CharacterAI和Mistral的研究人員。
穆拉蒂于2024年9月離開了OpenAI,稱她希望“為自己創(chuàng)造時間和空間,以便進行探索”。據(jù)媒體報道,2024年10月,她曾就為這家神秘的初創(chuàng)公司籌集超過1億美元資金進行洽談。上個月,這家初創(chuàng)公司招聘了幾位備受矚目的員工,其中包括OpenAI前特別項目負責人拉赫曼。(財富中文網(wǎng))
譯者:中慧言-王芳
上周二,前OpenAI首席技術(shù)官米拉·穆拉蒂正式分享了其新成立的人工智能初創(chuàng)公司的細節(jié),宣布了一個由眾多前同事組成的管理團隊,并公布了解決最先進的“前沿”人工智能系統(tǒng)存在的“關(guān)鍵差距”的計劃。上周二發(fā)布的一篇博文中披露了有關(guān)Thinking Machine Lab的這些細節(jié),距離穆拉蒂離開OpenAI已大約六個月之久,同時也揭開了人工智能行業(yè)最受矚目的項目之一的面紗。
這家從事人工智能研究與產(chǎn)品開發(fā)的初創(chuàng)公司將其使命定義為讓“人工智能系統(tǒng)得到更廣泛的理解、更具可定制性和通用性”。
該博文稱:"關(guān)于這些系統(tǒng)的訓(xùn)練方法的知識集中在頂級研究實驗室中,這不僅限制了人們關(guān)于人工智能的公開討論,還限制了人們有效使用人工智能的能力。”
OpenAI聯(lián)合創(chuàng)始人約翰·舒爾曼(John Schulman)將擔任這家初創(chuàng)公司的首席科學(xué)家。據(jù)《財富》雜志此前率先報道,舒爾曼在加盟Anthropic僅5個月后便離職,隨后加入了穆拉蒂的新公司。曾任OpenAI研究副總裁的巴雷特·佐夫(Barret Zoph)將擔任Thinking Machines Lab的首席技術(shù)官。
在上周二公布的29名Thinking Machines Lab員工中,至少有7人曾就職于OpenAI。喬納森·拉赫曼(Jonathan Lachman)曾任特別項目負責人;翁麗蓮(Lilian Weng)曾任副總裁;盧克·梅茨(Luke Metz)、山姆·施萊弗(Sam Shleifer)和斯蒂芬·羅勒(Stephen Roller)曾擔任研究科學(xué)家。其余員工包括來自Meta、谷歌DeepMind、CharacterAI和Mistral的研究人員。
穆拉蒂于2024年9月離開了OpenAI,稱她希望“為自己創(chuàng)造時間和空間,以便進行探索”。據(jù)媒體報道,2024年10月,她曾就為這家神秘的初創(chuàng)公司籌集超過1億美元資金進行洽談。上個月,這家初創(chuàng)公司招聘了幾位備受矚目的員工,其中包括OpenAI前特別項目負責人拉赫曼。(財富中文網(wǎng))
譯者:中慧言-王芳
Mira Murati, former chief technology officer at OpenAI.
PATRICK T. FALLON—AFP/GETTY IMAGES
Former OpenAI chief technology officer Mira Murati, formally shared details of her new AI startup on Tuesday, announcing a leadership team stacked with former colleagues and a plan to tackle “key gaps” around the most advanced, “frontier” AI systems. The details about Thinking Machine Lab, announced in a blog post on Tuesday, come roughly six months after Murati left OpenAI and shed light on one of the AI industry’s most highly anticipated projects.
The AI research and product startup described its mission as making “AI systems more widely understood, customizable, and generally capable.”
“Knowledge of how these systems are trained is concentrated within the top research labs, limiting both the public discourse on AI and people’s abilities to use AI effectively,” the blog post said.
OpenAI cofounder John Schulman will serve as the startup’s chief scientist, joining Murati after departing Anthropic just five months after signing on, as Fortune was first to report last week. Barret Zoph, who was OpenAI’s vice president of research, will be the chief technology officer of Thinking Machines Lab.
At least seven of the 29 Thinking Machines Lab employees announced on Tuesday were formerly with OpenAI. Jonathan Lachman was former head of special projects; Lilian Weng was former vice president; Luke Metz, Sam Shleifer, and Stephen Roller were former research scientists. The remaining employees include researchers from Meta, Google DeepMind, CharacterAI, and Mistral.
Murati left OpenAI in September 2024, saying she wanted to “create the time and space to do my own exploration.” In October 2024, she was in talks to raise over $100 million in funding for the stealth startup, according to media reports. Last month, the startup made several high-profile hires, including Lachman, former head of special projects at OpenAI.