Andrew Ng's ground-level observation as of February 2026: the primary mechanism of AI-driven job displacement is not wholesale automation but competency-based replacement. Workers who use AI are replacing workers who don't, within the same roles. This pattern is subtler and more pervasive than the mass-unemployment narrative, and it extends well beyond technical roles into marketing, recruiting, and analysis. Most tech layoffs to date stem from pandemic-era overhiring corrections and general cost-cutting, not AI automation. A handful of highly AI-exposed professions (call center operators, translators, voice actors) face direct displacement, but the broader pattern is intra-role competition where AI-fluent workers demonstrate materially higher productivity than non-adopters. Businesses are beginning to make personnel decisions on this basis. ## Key Principles - "AI won't replace workers, but workers who use AI will replace workers who don't" — the dominant displacement mechanism - Most 2025-2026 tech layoffs are pandemic corrections, not AI automation evidence - AI competency is becoming mandatory in traditionally non-technical roles (marketing, recruiting, analysis) - Direct AI displacement remains concentrated in narrow, highly-exposed professions ## Cross-Domain Connections - [[AI Salary Arbitrage Window]] — The flip side: AI-fluent workers currently earning pre-AI salaries while delivering amplified output - [[K-Shaped Developer Stratification]] — India-specific manifestation of the same AI-capable vs. non-AI-capable bifurcation - [[Credential-to-Career Pipeline Collapse]] — Workers who learned coding for salary rather than curiosity are most vulnerable to this replacement pattern - [[AI Capability Denial Acceleration Pattern]] — Community denial of this replacement pattern amplifies individual risk - [[AI Context-Switching as Workforce Compression Engine]] — How AI enables talented individuals to absorb multiple roles through rapid context-switching - [[AI Task-Level Job Displacement]] — Core mechanism: AI displaces work task-by-task rather than eliminating entire roles at once --- *Source: [[Job seekers in the U.S]] (Andrew Ng, tweet) — https://x.com/andrewyng/status/2021259884709413291*