Summary of Key Points
This article highlights the paradoxical phenomenon where Silicon Valley tech companies (such as Meta, Microsoft, and Google) are conducting large-scale layoffs even as their performance is growing—especially targeting employees who are adept at using AI. The underlying logic is that companies are distilling employee skills to train AI to replace them. The criteria for valuing tech companies by capital have shifted from “talent density” to “fewer people, more AI.” Workers must abandon the notion of relying solely on AI to increase productivity and instead develop intangible abilities (such as industry expertise and interpersonal trust) to remain competitive.
Detailed Explanation
1. Are You Being Laid Off Because You Use AI Well? You’ve Become an “Free Trainer” for AI
Do you think being proficient in AI is a advantage? Wrong! Companies are now packaging employees’ work experience, habits, and decision-making processes into AI skill sets (e.g., the popular “Colleague Skill” projects on GitHub). Once the employee leaves, the AI takes over their tasks—this is known as “employee distillation.”
Meta’s “Model Capability Initiative (MCI)” aims to teach AI how to perform tasks like clicking buttons, writing documents, and following approval processes. What AI lacks most in training is not computing power, but the data from your daily interactions with company computers (which are valuable training materials). The more you use AI efficiently and provide standardized data, the faster it learns, and the sooner it can replace you. Employees jokingly say that Meta’s workers are training their own AI replacements.
2. A Shift in Capital Values: Tech Companies Now Compete on “Fewer People, More AI”
Previously, tech companies valued “talent density”—the more graduates they hired from top universities, the higher their valuation. Now, capital’s criteria have reversed: fewer employees and more GPUs are more valuable.
For example, Meta cut 8,000 jobs, saving $3 billion annually, and invested all that money in AI hardware (doubling its capital expenditure to $135 billion in 2026, purchasing H100 GPUs). Microsoft’s net profit increased by 27% after laying off 15,000 employees, and its stock price rose, with a market value exceeding $4 trillion. Capital is voting with its feet: fewer employees lead to higher valuations. Human resources have shifted from being an asset to a cost, with employee salaries replaced by the cost of hardware.
3. Which Jobs Will Be Taken Over by AI? Which Can You Protect?
- Jobs That Are Likely to Be Automated: Those that are measurable and standardized (e.g., writing 10 articles per day, high code output, fast editing). These skills are essentially “high-quality training data” that can be used to train AI. Once AI learns your methods, it can do the work on its own.
- Jobs That AI Cannot Take Over: Intangible knowledge (that cannot be documented). For example, trust built with clients, intuitive understanding of industry trends, and influential networks within the industry. These cannot be “distilled” by AI.
4. How Workers Can Survive This Change: Stop Being Training Data, Become Rule Makers
Instead of worrying about being laid off, focus on “upgrading” your skills from using AI to improve productivity to using AI to reinvent your work.
Industry elites don’t just use AI to write articles; they integrate it with their expertise to drive innovative AI applications for the company. They become drivers of AI transformation rather than targets of replacement. The difference lies in whether you are a “data provider” for AI or a “rule maker.”
In conclusion, in the AI era, your value does not depend on your ability to use tools but on what tools cannot learn from you—judgment, trust, networks, and industry influence. These are the core competencies that will help you navigate through layoffs.
This article clearly warns workers: don’t let yourself become a source of “high-quality training data.” Invest more time in developing skills that AI cannot replace. After all, tools will continue to improve, but the intangible value of humans is irreplaceable.