科技行业广泛采纳的招聘和管理方法正受到严格质疑。作者Tejas Kumar指出，包括Bar Raiser面试、OKR目标管理、Spotify组织模型等在内的这些流程大多缺乏科学依据支持，反而倾向于优化那些追求秩序而非创新的人才1。
谷歌的研究数据表明现有面试制度存在根本性问题。该公司对数万份面试评分与实际工作表现的对比中发现"零相关性"，将面试描述为"完全的随机混乱"1。然而谷歌的"四次面试规则"(Rule of Four)能以86%的置信度预测招聘决定，并将招聘周期从6个月缩短至45天1。
Bar Raiser面试制度源自亚马逊1999年的创新，该公司目前拥有超过10,000名Bar Raiser1。与此同时，Spotify从未完全执行其2012年白皮书中的模型；该模型共同作者曾坦言告诉产品经理"我们写它的时候都没在做"1。OKR倡导者Marty Cagan在2020年表示已"停止推荐这个实践"，称其"浪费时间和精力，收效甚微"1。敏捷开发先驱Dave Thomas在2014年发表"Agile已死"，称"agile"一词"已被颠覆到实际上毫无意义"1。
面试流程的问题在实践中已有明证。北卡罗来纳州立大学2020年的研究表明，被观察的候选人在白板面试中的表现是单独房间的一半，其中"所有接受公开面试的女性都失败了"1。Columbia学生Roy Lee在2025年用隐藏的AI工具通过Amazon面试获得实习offer，该事件引发的关注导致Amazon删除相关视频，但Lee因此融资1500万美元1。
Stanford基于ADP数据的研究发现了AI时代的新风险。到2026年6月，22-25岁在AI高暴露职位中的就业率比低暴露职位低19%1。当前招聘机制会在AI时代加速淘汰只能执行命令的工作1。
作为改革方向，Automattic等企业已尝试替代方案，为试用项目支付标准费率$25/小时，试用期最长20小时1。有建议用结构化面试、工作样本测试和付费试用期来替代现有流程，以更准确地识别人才1。
Author Tejas Kumar has challenged the widespread adoption of hiring and management methodologies across technology companies, contending that many popular frameworks lack empirical support and inadvertently optimize for compliance rather than innovation.1 His critique examines several industry-standard practices and the evidence—or lack thereof—backing their effectiveness.
Google's internal research revealed striking findings about interview validity.1 When comparing interview scores from tens of thousands of assessments against actual job performance, the company found "zero correlation," describing the process as "complete random chaos."1 This discovery prompted Google to develop its "Rule of Four," which can predict hiring decisions with 86 percent confidence and has reduced hiring cycles from six months to 45 days.1
The Bar Raiser interview system, originating from Amazon's 1999 initiative and now employed by Amazon with over 10,000 trained Bar Raisers, exemplifies the industry's reliance on unvalidated methods.1 Similarly, the Spotify model—promoted through a 2012 white paper—was never fully implemented by Spotify itself; one of the paper's co-authors told product managers, "We weren't even doing it when we wrote it."1 OKR advocate Marty Cagan acknowledged in 2020 that he had "stopped recommending" the practice at most companies, calling it a "waste of time and energy with little payoff."1 Software engineer Dave Thomas declared in 2014 that "Agile is dead," stating the term had "been subverted to mean almost nothing."1
Research into interview formats has uncovered additional problems.1 A 2020 North Carolina State University study found that observed candidates performed at half the level during public whiteboard interviews compared to private settings, and "all women who took public interviews failed."1 A Columbia student named Roy Lee demonstrated the vulnerability of Amazon's interview process in 2025, passing an internship interview using hidden AI tools; Amazon subsequently removed the video, but Lee raised $15 million in funding.1 Some companies are experimenting with alternatives: Automattic pays standard rates of $25 per hour for trial projects, with trial periods lasting up to 20 hours.1
Looking ahead, artificial intelligence may intensify employment pressures for younger workers in roles with high AI exposure.1 Stanford research using ADP data projects that by June 2026, employment rates for 22-to-25-year-olds in high-AI-exposure positions will be 19 percent lower than in low-exposure roles.1
评论
还没有评论，欢迎留下第一条。