AI job panic and internal metrics
Public discussion in Silicon Valley is showing rising ‘AI job panic’, with some arguing that interpersonal and humanities skills may become more valuable as automation spreads. Separately, a Meta employee built an internal leaderboard of the company’s top 250 AI token users — a sign firms are experimenting with ways to quantify AI adoption, and MoneyControl reports Zuckerberg did not make that top 250. (channelnewsasia.com) (moneycontrol.com)
Silicon Valley’s AI debate has split into two tracks at once: fear that artificial intelligence will erase jobs, and pressure inside companies to measure who is using it most. (channelnewsasia.com) (moneycontrol.com) At the HumanX conference in San Francisco, an ad at the entrance said “Stop hiring humans,” and the four-day event drew about 6,500 investors, founders and executives. Writer chief executive May Habib said Fortune 500 bosses were having a “collective panic attack” over the technology. (channelnewsasia.com) Speakers at HumanX argued over what workers should do next. Some said interpersonal judgment and humanities training could gain value as automation spreads, while DeepLearning.AI founder Andrew Ng said on April 8 that treating coding as obsolete would be “some of the worst career advice ever given.” (channelnewsasia.com) The anxiety is no longer abstract. Channel NewsAsia, citing the conference discussion, reported that Salesforce cut 4,000 customer-support jobs after saying artificial intelligence now handles about 50 percent of that work, while Block chief executive Jack Dorsey said “intelligence tools” changed how companies operate as he outlined plans to shrink headcount. (channelnewsasia.com) Inside Meta, the same shift showed up as a dashboard. Moneycontrol reported that an employee built an internal leaderboard ranking the company’s top 250 artificial-intelligence token users, with labels including “Token Legend” and “Cache Wizard,” and Mark Zuckerberg did not appear in that top 250. (moneycontrol.com) A token is the basic unit a chatbot reads and writes, so counting tokens is one way to count how much employees use large language models. Reports on the Meta tool said the leaderboard tracked usage across more than 85,000 employees and turned internal adoption into a visible ranking. (finance.yahoo.com) (moneycontrol.com) The leaderboard did not last. Multiple reports published on April 9 and April 10 said Meta took the dashboard down after details were shared outside the company. (finance.yahoo.com) (timesofindia.indiatimes.com) Meta had already moved toward formal measurement before the dashboard surfaced. Reports in late 2025 said Janelle Gale, Meta’s head of people, told staff that “AI-driven impact” would become a core expectation in 2026 performance reviews. (hrgrapevine.com) (news.abplive.com) The two stories point to the same workplace question: companies are not only asking whether artificial intelligence can replace tasks, but also how to score the workers who use it. In San Francisco, that question was argued from a conference stage; at Meta, it briefly appeared as a numbered list. (channelnewsasia.com) (moneycontrol.com)
