I just attended a sharing session with a leading VC firm and saw some thought-provoking data and perspectives: AI is reshaping the personnel structure of large companies, and at a faster pace than imagined. Several key figures: ① If 50% of entry-level jobs are replaced, the US unemployment rate will surge from 4% to 10%. ② A 20% layoff at a large company → EPS increase of ~10%; a 20% layoff at a labor-intensive software company → EPS increase of over 70%. ③ Another 10% layoff would bring the workforce of large companies back to pre-pandemic levels of 2019. The underlying logic of layoffs has changed: - AI investment is consuming cash flow; cloud vendors' FCF may turn negative by 2027—layoffs are not an option, but a financial necessity. - AI is compressing the R&D cycle from 2-3 months to 2 weeks; traditional roles like QA and PM have transformed from assets into bottlenecks. - Meta's core AI team has only 3,000-4,000 people; the value of the remaining 70,000+ employees is being re-evaluated. But history gives us confidence: Every technological revolution initially presents employment panic, but the final result is often a double increase in both job numbers and income. Radiologists are a prime example—after the widespread adoption of AI-assisted diagnosis, their income and the number of practitioners actually increased. Implications for businesses: → Organizations should become flatter and more atomized, dynamically forming cross-functional teams as needed. → AI should handle standardized tasks, allowing humans to focus on creative ones. → Short-term profits can be maintained through efficiency; long-term value can be unlocked through restructuring. In short: Layoffs are a short-term cost of technological iteration, not the end result. The real risk is not AI replacing humans, but companies failing to restructure processes and clinging to old models.