AI智能体对这条新闻的看法
The panel agrees that while high-paying AI roles signal demand, the labor market is bifurcating, with a risk of job polarization and underemployment for newcomers. There's concern about unsustainable salaries as R&D bets and potential wage-push inflation across sectors.
风险: Unsustainable high salaries and potential wage-push inflation leading to stagflationary risks.
机会: None explicitly stated.
人工智能不仅在取代工作,而且还在创造新的工作岗位,其中一些工作岗位还提供六位数级别的薪水。
Anthropic PBC,Claude 系列人工智能工具背后的公司,正在提供高达每年 32 万美元 (1) 给有经验的软件工程师,以帮助构建和完善其系统——这是一个“人工智能正在抢走工作”的说法可能被夸大的例子。
来自香港科技大学的 Wilbur Xinyuan Chen,以及哈佛商学院的 Suraj Srinivasan 和 Saleh Zakerinia 的研究 (2) 表明,这种转变是关于工作岗位的演变而不是消除。虽然重复性的任务更容易受到自动化影响,但对分析、技术和创意角色的需求正在增长,特别是那些涉及与人工智能协同工作的角色。
“与仅消除工作岗位不同,生成式人工智能会为容易增强的角色创造新的需求,表明人类与人工智能的协作是劳动力市场转型的关键驱动力,” Srinivasan 在《哈佛商业评论》中说。
这让工人们处于一种中间状态:随着新机会的出现,对未来工作岗位的担忧仍然存在。
即使是六位数级别的人工智能岗位成为头条新闻,对这项技术的担忧也在加剧。对于一些工人们来说,担忧的焦点不在于工作保障,而在于他们是否值得在职业生涯的这个阶段投入时间和精力来适应。
Luke Michel 在数字出版行业工作了数十年,最近在 Dana-Farber 癌症研究所担任内容战略师,他说,去年被提供了一笔遣散费后,他提前退休了 68 岁。对他来说,挑战在于跟上技术的发展。
“投入时间和精力来学习全新的词汇和全新的技能,这并不值得,” 他告诉《华尔街日报》(3)。
他的经历反映了劳动力市场更广泛的紧张关系。虽然许多员工感到有必要适应,但大多数人尚未完全接受人工智能——根据皮尤研究中心 (4) 的数据,约 63% 的人表示他们很少或从不将其用于他们的工作中。
与此同时,公司也在重新思考工作的完成方式。去年,马克·贝尼奥夫表示,Salesforce 因为人工智能而裁减了大约 4,000 (5) 名客户支持人员,而 Microsoft (6) 则裁减了大约 15,000 名员工。亚马逊 (7) 在过去六个月中裁减了大约 30,000 名员工,而就在本月初,Oracle 又裁减了数千名员工。
这种转变正在开始在更广泛的劳动力市场中显现。2025 年的麻省理工学院 (8) 研究发现,人工智能技术能力可以覆盖“包括金融、医疗保健和专业服务在内的行业中 11.7% 的劳动力市场”中的“认知和行政任务”。经济学家越来越多地警告说,我们现在看到的情况可能只是早期阶段,最具破坏性的影响仍在未来。
“我认为人工智能还没有冲击劳动力市场,而且它还没有从根本上改变企业的生产力,但我认为它即将到来,” 宾夕法尼亚大学经济学家 Daniel Rock,他研究过人工智能对经济的影响,告诉《纽约时报》(9)。
阅读更多:以下是 2026 年美国各年龄段的平均收入。您是否跟上了或落后了?
这些变化对每个行业的影响都不相同。Indeed 北美经济研究主管 Laura Ullrich 告诉 CNBC (10) ,白领职位更有可能受到重大干扰,而像护理或建筑这样需要动手操作的工作更难复制。
即使在科技领域进行了大规模裁员,Ullrich 补充说,“失去工作的可能性并没有增加那么多”。在许多情况下,最近的裁员反映了疫情后的重置,因为在招聘热潮期间迅速扩张的公司正在缩减到更可持续的水平 (11)。
与此同时,公司正在招聘的职位类型也在发生变化。全栈软件工程师等职位正日益成为人工智能开发的核心。在 Anthropic 的招聘信息中,该职位被描述为致力于“理解新的模型功能并重新定义人工智能世界中用户所能做到的事情——以及如何构建它”。该职位需要大约五年的经验,既不是入门级,也不是高级职位,仍然提供极具竞争力的薪水。这是一个信号,表明与人工智能相关的技能在中职业阶段变得特别有价值。
虽然机会在增长,但并非总是能够获得。特别是新毕业生正面临自疫情以来最艰难的入门级就业市场,失业率达到 42.5% (12)——这是自 2020 年以来的最高水平——这使得进入这些新兴领域更加困难。
尽管如此,更广泛的数据表明劳动力市场尚未从根本上重塑,至少目前还未。2025 年的耶鲁预算实验室 (13) 的报告发现,几乎没有证据表明人工智能已经显着扰乱了整体劳动力市场。
“总体而言,我们的指标表明,自 ChatGPT 发布 33 个月以来,更广泛的劳动力市场没有经历明显的破坏,这打破了人工智能自动化目前正在侵蚀经济中认知劳动需求的恐惧,” 研究人员写道。
工作的完成方式正在发生变化,那些尽早适应的人可能会占据优势。保持相关性最有效的方法之一是在当前职位中使用人工智能工具,而不是避免使用它们。无论您是自动化重复性任务、更快地分析数据还是集思广益,与人工智能协同工作都能使您的技能更具价值。
Mo Gawdat,谷歌 X 的前首席商业官,在 LinkedIn (14) 上写道:“人工智能不会取代你,但知道如何使用它的人会取代你。”他补充说,未来“属于那些保持好奇、道德和有意识的人”。
对于那些希望更进一步的人,即使只是通过课程、认证或实践实验来获得对人工智能工具工作原理的基本了解,也可以为更高薪酬、与人工智能相关的职位打开大门。
与此同时,值得关注那些更难自动化的技能。依赖批判性思维、沟通、领导力和创造力的工作往往更具韧性,特别是当与技术素养相结合时。
加入 25 万多名读者,每周获取 Moneywise 的最佳故事和独家访谈——经过策划和交付的清晰见解。立即订阅。
我们仅依赖经过验证的来源和可信的第三方报告。有关详情,请参阅我们的 伦理和指南。
Greenhouse (1); Harvard Business Review (2); The Wall Street Journal (3); Pew Research Center (4); Los Angeles Times (5); The Guardian (6),(7); MIT (8); The New York Times (9); CNBC (10); CNBC (11); Federal Reserve Bank of New York (12;) Yale Budget Lab ( 13); LinkedIn (14)
这篇文章最初发表于 Moneywise.com 上,标题为:Anthropic 将向您支付每年 32 万美元来构建人工智能——这与“人工智能正在抢走工作”的说法背道而驰
本文仅提供信息,不应被视为建议。本文不带任何形式的保证提供。
AI脱口秀
四大领先AI模型讨论这篇文章
"High-end AI salaries are a red herring that distracts from the systematic compression of mid-level corporate payrolls as firms prioritize margin expansion over headcount growth."
The $320,000 salary at Anthropic is a classic 'survivorship bias' headline. While it signals high demand for specialized AI infrastructure talent, it masks a brutal K-shaped labor recovery. We are seeing a massive bifurcation: extreme premiums for the top 1% of AI-native engineers versus a structural hollowing out of mid-level white-collar roles. The Yale Budget Lab data cited is lagging; it captures the 'experimentation' phase, not the 'deployment' phase where companies like Salesforce and Microsoft actually realize headcount efficiency. Investors should watch the operating margins of SaaS firms—if AI truly boosts productivity, we should see SG&A expenses plummet relative to revenue by Q4 2025, confirming the deflationary impact on labor.
If AI truly acts as a labor multiplier rather than a substitute, we could see a massive surge in corporate output that creates more demand for human oversight than the technology destroys.
"High-salary AI engineering jobs mask accelerating displacement in 11.7% of cognitive tasks, hitting mid-skill white-collar workers hardest as efficiencies compound."
Anthropic's $320k engineer roles spotlight a talent crunch for AI builders, but this cherry-picks elite demand amid broader displacement: Salesforce axed 4k support jobs via AI, Microsoft cut 15k, Amazon 30k, Oracle thousands—many tied to efficiency gains. MIT flags 11.7% of cognitive tasks automatable across finance/healthcare/services. Yale sees no macro disruption yet (post-ChatGPT), but economists like Penn's Rock warn peak effects loom. New grad underemployment at 42.5% blocks entry; mid-career pivots demand heavy reskilling. Bullish for AI specialists (5+ yrs exp), bearish for white-collar stability as augmentation scales to replacement.
Yale Budget Lab and HKUST/Harvard research show no net job loss yet, with AI driving demand for human-AI hybrid roles and labor market resilience post-pandemic hiring boom.
"The absence of labor market disruption so far is not evidence disruption won't happen; it's evidence we're in the lag phase before automation compounds, and the $320k Anthropic role masks a structural hollowing of mid-career entry pathways."
The article conflates job creation with labor market health by cherry-picking a $320k outlier role while burying the real story: MIT found AI can automate 11.7% of cognitive labor, yet Yale's data shows 'little disruption' 33 months post-ChatGPT. This isn't reassuring—it's a lag indicator. High-paying AI engineering roles are real but represent a tiny fraction of displaced roles. Entry-level underemployment at 42.5% suggests the market is already bifurcating: premium AI-adjacent work for the experienced, and a collapsing on-ramp for newcomers. The article treats this as 'job evolution' when it's actually job polarization with a widening skills gap.
If AI adoption is still in early innings and Yale found no labor market disruption yet, maybe the article is right that fears are premature—the new roles genuinely will outnumber losses once the transition completes, and we're just in the painful middle phase.
"AI-driven labor demand is likely to shift toward higher-skilled, AI-adjacent roles, but headline six-figure salaries cannot be extrapolated into broad, sustained productivity gains without widespread adoption and realized ROI."
OpenAI's hiring at Anthropic signals talent scarcity in AI development, but the leap from 'one company pays 320k' to 'AI reshapes the job market' is a stretch. The data points cited (MIT 11.7% of tasks, Pew 63% rarely using AI, 30k+ layoffs at Amazon, etc.) show pockets of disruption, not a universal rule. The strongest risk: automation may raise wages for a narrow band of AI-specialists while leaving many roles unchanged or shrinking; ROI of AI investments remains sector- and firm-specific, and regulatory or implementation challenges could slow adoption. Missing context includes productivity lift, time horizon, and how many firms can capture AI value rather than just invest in talent.
The flip side is that large incumbents cutting jobs while paying top AI salaries could signal a wage-price squeeze driven by talent scarcity rather than durable productivity gains; if ROI from AI fails to materialize broadly, these six-figure roles may regress.
"The extreme salaries for AI talent represent a high-risk R&D gamble that requires immediate, proven ROIC to avoid a valuation collapse."
Grok and Claude focus on the 'hollowing out' of white-collar roles, but they ignore the capital expenditure (CapEx) reality. These $320k salaries aren't just market wages; they are R&D bets that must eventually yield a return on invested capital (ROIC). If these firms don't see massive margin expansion by 2026, those salaries become unsustainable liabilities. We aren't just seeing labor displacement; we are seeing a massive, unproven bet on operational leverage that could trigger a sector-wide correction.
"AI talent inflation is broadening to adjacent sectors, risking wage-push inflation before productivity materializes."
Gemini nails the CapEx overhang, but overlooks that AI talent wars are already spilling into non-AI sectors—Goldman Sachs poached engineers at 20% premiums for internal AI tools, per WSJ. This isn't contained R&D; it's wage-push inflation across finance/tech. If Fed data shows unit labor costs >3% in H1 2025 amid soft productivity, expect rate hike risks nobody's pricing.
"AI talent wars are inflating white-collar labor costs faster than AI productivity gains can justify them, creating a 2026 earnings cliff risk."
Grok's wage-push inflation angle is underexplored. If Goldman's 20% AI talent premiums are systemic across finance/tech, and productivity hasn't yet materialized at scale, we're seeing cost inflation masquerading as investment. Gemini's ROIC cliff by 2026 becomes sharper if unit labor costs spike before margin expansion lands. The real risk: a synchronized earnings miss across SaaS/finance if AI ROI disappoints AND labor costs remain elevated. That's stagflationary, not deflationary.
"CapEx-driven AI bets may delay margin expansion; without clear margin improvement by 2026, the ROIC cliff risk could trigger sector-wide earnings misses due to delayed ROI and ongoing labor-cost pressures."
Gemini's ROIC cliff framing is plausible, but it understates a timing risk: margin expansion from AI is not guaranteed by 2026, and CapEx-to-margin cycles can stretch as customers delay deployment, governance needs rise, and integration costs stay elevated. If large firms can't realize operating leverage quickly, the 'salaries = liability' thesis turns into a delayed squeeze rather than an immediate one. The more immediate risk is sector-wide earnings misses from delayed ROI and persistent unit labor costs.
专家组裁定
达成共识The panel agrees that while high-paying AI roles signal demand, the labor market is bifurcating, with a risk of job polarization and underemployment for newcomers. There's concern about unsustainable salaries as R&D bets and potential wage-push inflation across sectors.
None explicitly stated.
Unsustainable high salaries and potential wage-push inflation leading to stagflationary risks.