«

xAI人才大规模流失的背后原因是什么？

qimuai 发布于 阅读：2 一手编译

xAI人才大规模流失的背后原因是什么？

内容来源：https://www.theverge.com/ai-artificial-intelligence/878761/mass-exodus-at-xai-grok-elon-musk-restructuring

内容总结：

【独家深度】xAI核心团队震荡：安全部门被曝解散，多名创始成员离职创业

近日，埃隆·马斯克旗下人工智能公司xAI陷入剧烈人事动荡。本周二至周三，联合创始人吴宇怀（Tony Wu）与Jimmy Ba相继在社交平台宣布离职，前者称“将开启新篇章”，后者表示“需重新校准人生方向”。至此，xAI最初的12位联合创始人仅剩半数留任。多名普通员工也公开表示已离开公司，部分人透露将创立自己的AI企业。

重组背后：安全争议与创新困境
据多位前员工向媒体透露，此次离职潮源于公司内部对安全问题的分歧及长期陷入“追赶模式”的挫败感。一位今年初离职并要求匿名的员工指出，xAI过度聚焦于开发涉及成人内容的Grok模型，且安全团队已被解散，“除了针对儿童性虐待内容的基础过滤外，模型几乎没有任何安全审核流程”。他直言：“安全部门在xAI已名存实亡。”

马斯克近期公开的xAI重组架构图显示，公司业务将划分为Grok主模型与语音、编程、图像视频及企业数字化模拟四大板块，其中未提及安全团队。该前员工补充称，产品优先级决策常通过包含马斯克在内的全员群聊决定，管理层内耗严重。

并购催化：太空野心与员工出走
上周宣布与SpaceX合并后，xAI估值据传达1.25万亿美元。马斯克在内部会议中提出建设“月球AI卫星工厂”等太空计划，但部分员工对战略方向产生疑虑。另一位离职员工坦言：“试图复现OpenAI一年前的成果无法战胜他们，公司几乎不敢进行任何风险尝试。”

合并同时为员工带来变现契机——据报道xAI股东获发2500亿美元新股，促使部分持有股权的员工选择创业。前工程师Vahid Kazemi公开表示“现有AI实验室都在重复造轮子”，已着手创立新项目；另有多名前员工联合成立AI基础设施公司Nuraline，强调“学习不应止步于模型参数”。

文化隐忧：安全缺位与高压决策
《华盛顿邮报》本月初曾披露xAI安全机制薄弱问题，此次更多前员工证实该状况。有离职者指出，马斯克认为“安全即审查”，甚至主动推动模型“更无约束”。工程师常直接推送代码至生产环境，长期缺乏人工审核环节。“在这里生存的法则就是闭嘴执行马斯克的指令。”一名前员工总结道。

随着AI赛道竞争白热化，这场人才流失是否会影响xAI与SpaceX的“天地联动”战略，仍需持续观察。

中文翻译：

过去几天对xAI而言可谓风云激荡，员工与联合创始人接连宣布离职。周二至周三期间，联合创始人吴宇怀（Tony Wu）率先发文称"是时候开启人生新篇章"，随后另一位联合创始人Jimmy Ba也发布类似贴文，表示"需要在宏观层面重新校准人生方向"。至此，xAI最初的12位联合创始人仅剩半数留任。多名员工亦在X平台宣布离职，部分人透露将创立自己的AI公司。

xAI为何出现离职潮？

前员工透露，公司重组源于安全理念冲突及长期陷于"追赶困境"。埃隆·马斯克通过并购将这家AI初创公司纳入其太空探索公司SpaceX与社交媒体平台X的体系。自上周宣布与SpaceX合并以来，市场传闻其估值达1.25万亿美元，而马斯克宣称新集团将打造"天基AI"数据中心，成为"地球内外最具雄心的垂直整合创新引擎"。据称在周二的xAI内部会议上，马斯克甚至提及建设月球AI卫星工厂与城市的计划。

企业并购后出现人员流动本属常态，马斯克也承认部分离职属于"重组过程中不得已的人员调整"。但种种迹象表明，不少人对马斯克设定的发展方向心存疑虑。

安全团队形同虚设

一位因担心报复而要求匿名的年初离职者向媒体透露，公司许多人对xAI专注开发成人向Grok产品且忽视安全的做法感到幻灭。他认为公司始终陷于"追赶模式"，未能推出超越竞争对手的创新成果："尽管迭代迅速，但从未实现如'我们已超越OpenAI或Anthropic等公司'的突破性进展。"

SpaceX合并案据称为xAI股东带来2500亿美元新股，持有股权的员工因此获得更多资金实现个人创意。前员工Vahid Kazemi在X平台写道："所有AI实验室都在重复建设相同的东西，这很无趣。我认为需要更多创造力，所以我将开启新事业。"另一位离职者则表示将创立"专注于加速科学研究"的新项目。

还有前员工透露正与其他xAI旧部共同创立AI基础设施公司Nuraline，他指出："在xAI期间，我看到了攻克任何可量化问题的清晰路径，同时也目睹原始智能如何被细微人为错误所扼杀……学习不应止步于模型权重，而应持续优化AI系统的每个环节。"

安全部门名存实亡

马斯克发布的45分钟内部会议录音显示，xAI将重组为四大板块：Grok主模型与语音、编程、图像视频（Imagine）以及Macrohard（马斯克称其目标为"实现企业的完整数字孪生"）。年初离职者透露，Grok转向成人内容部分源于安全团队解散，除儿童色情内容等基础过滤外，模型几乎已无安全审核机制。"xAI的安全部门早已名存实亡"，他表示。查阅马斯克在X平台发布的新架构图，确实未见安全团队踪影。该人士还指出，任职期间领导层对产品功能优先级争论不休，内耗时常阻碍进展，许多决策通过包含马斯克在内的全员群聊敲定。

另一位在重组前离职的匿名员工同样认为公司陷于被动追赶："重复OpenAI一年前的工作无法实现超越。所有举措都是追赶，几乎不存在风险决策，未经实践验证的方案我们绝不会尝试。"他同时证实公司忽视安全问题——这亦与《华盛顿邮报》本月初的报道相呼应。

"公司完全不存在安全机制——无论是图像模型还是聊天机器人"，该员工表示，"马斯克实际上在刻意让模型更失控，因为在他看来安全就意味着审查。"他还透露xAI工程师长期采用"直接部署上线"模式，很长时间内都缺乏人工审核环节。

"在这里生存的法则就是保持沉默，执行马斯克的指令。"他总结道。

今日热门资讯

英文来源：

The past few days have been a wild ride for xAI, which is racking up staff and cofounder departure announcements left and right. On Tuesday and Wednesday, cofounder Yuhuai (Tony) Wu announced his departure and that it was “time for [his] next chapter,” with cofounder Jimmy Ba following with a similar post later that day, writing that it was “time to recalibrate [his] gradient on the big picture.” The departures mean that xAI is now left with only half of its original 12 cofounders on staff. A number of staffers also took to X to announce they were leaving xAI, with some announcing that they were starting their own AI companies.
What’s behind the mass exodus at xAI?
Former employees say the restructuring followed tensions over safety and being “stuck in the catch-up phase.”
Former employees say the restructuring followed tensions over safety and being “stuck in the catch-up phase.”
Elon Musk’s AI startup, by one merger / acquisition or another, is under the same umbrella as both his space company, SpaceX, and his social media platform, X. Since the SpaceX merger was announced last week, rumors have swirled about the reported $1.25 trillion valuation and the all-in-one company’s future plans, which Musk announced would involve “space-based AI” data centers and “the most ambitious, vertically-integrated innovation engine on (and off) Earth.” In an internal meeting at xAI on Tuesday, Musk reportedly talked of plans to build an AI satellite factory and city on the Moon.
There’s often a natural departure point for companies post-merger, and Musk has announced that some of the departures were a reorganization that “unfortunately required parting ways with some people.” But there are also signs that people don’t like the direction Musk has taken things.
Are you a current or former xAI employee? Contact me via Signal at haydenfield.11 on a non-work device with tips.
One source who spoke with The Verge about the happenings inside the company, who left earlier this year and requested anonymity due to fear of retaliation, said that many people at the company were disillusioned by xAI’s focus on NSFW Grok creations and disregard for safety. The source also felt like the company was “stuck in the catch-up phase” and not doing anything new or fundamentally different from its competitors. “Although we were iterating really fast, we were never able to get to a point like, ‘Oh, we’ve made a step function change over what OpenAI or Anthropic or other companies had released,’” he said.
The SpaceX merger meant that xAI shareholders were reportedly issued $250 billion in new shares, so employees with equity ostensibly have more runway to fund their own ideas. One former employee, Vahid Kazemi, wrote on X that “all AI labs are building the exact same thing, and it’s boring. I think there’s room for more creativity. So, I’m starting something new.” Another former staffer said he left the company to “build something new, focused on accelerating science.”
Yet another former employee said he was launching an AI infrastructure company called Nuraline alongside other ex-xAI employees. He wrote, “During my time at xAI, I got to see a clear path towards hill climbing any problem that can be defined in a measurable way. At the same time, I’ve seen how raw intelligence can get lobotomized by the finest human errors … Learning shouldn’t stop at the model weights, but continue to improve every part of an AI system.”
“Safety is a dead org at xAI.”
Musk posted a recording of xAI’s 45-minute internal all-hands meeting that announced the changes, adding that xAI will be categorized into four main areas: Grok Main and Voice (the main Grok AI model), Coding, Imagine (image and video), and Macrohard (“which is intended to do full digital emulation of entire companies,” Musk said).
The source who departed earlier this year said Grok’s turn toward NSFW content was due partly to the safety team being let go, with little to no remaining safety review process for the models besides basic filters for things like CSAM. “Safety is a dead org at xAI,” he said. Looking at the restructured org chart Elon Musk shared on X, there’s no mention of a safety team. The source also said that during his time at xAI, he felt leadership had a lot of differing opinions on which product features to prioritize and that infighting sometimes stalled progress. A lot of decisions on what to ship happen via an all-company group chat on X with Musk in it, he said.
A second source, who departed xAI before the recent restructuring and requested anonymity, echoed the thoughts that Musk’s company was playing catch-up. “Trying to do what OpenAI was doing a year ago is not how you beat OpenAI,” he said. “Everything is a catch-up. There’s almost zero risky bet. If something hasn’t been done before we’re not going to do it.”
He mentioned xAI’s lack of safety focus too, an issue that was also highlighted in reporting by The Washington Post earlier this month.
“There is zero safety whatsoever in the company — not in the image [model], not in the chatbot,” the second source said. “He [Musk] actively is trying to make the model more unhinged because safety means censorship, in a sense, to him.”
The second source also said engineers at xAI immediately “push to prod[uction]” and that for a long time, there was no human review involved.
“You survive by shutting up and doing what Elon wants,” he said.
Most Popular

ThevergeAI大爆炸

文章目录

    📚 推荐阅读


    扫描二维码，在手机上阅读
    版权所有：启幕AI
    文章标题：xAI人才大规模流失的背后原因是什么？
    文章链接：https://www.qimuai.cn/?post=3317
    本站文章均为原创，未经授权请勿用于任何商业用途