据Rohan Paul引用a16z及Hebbia CEO观点，AI并未取代人类工作，反而创造了更多就业。数据显示，到2026年5月，最顶尖1%的公司每年AI支出已达约9万美元。若保持14.1%的月增长率，该数字将在2026年底超过软件工程师平均年薪19.2万美元。同期，高AI投入公司员工数增长10.2%，低投入公司基本不变。AI正成为增长引擎而非降本工具，入门级岗位增加12%，人类比软件更便宜。
It may not be long before a firm is paying more for AI tokens per employee than it is to employ that worker.
The biggest AI adopters are combining more people with far more compute.
In May 2026, the most advanced 1% of companies were already spending around $90,000 per year on AI.
If the current monthly growth rate of 14.1% continues, that number could surpass the average $192K a year a software engineer makes by the end of 2026.
During the same 24-month period, companies that spent heavily on AI added 10.2% more employees, while companies that spent a little on AI stayed roughly the same.
AI is a huge enabler of growth than a tool for cutting costs.
Companies invest a lot of money in tokens because each worker can do more work, help more customers and support a bigger business.
Entry-level jobs increased by 12% and low-intensity adopters’ numbers did not change in a statistically significant way.
"AI was supposed to replace human labor. It did the opposite. For the first time in history, humans are cheaper than software. And AI is creating more jobs than it eliminates." Hebbia CEO George Sivulka on what the core lessons of human management mean for agent workforces: https://www.a16z.news/p/the-next-ai-goldrush-tokens-loops在 X 查看被引用的帖子
来源：Rohan Paul · x.com