Rapid advances in artificial intelligence have renewed concerns that earlier automation forecasts may understate the range of occupations exposed to technological change. This paper provides an ex-post assessment of how U.S. occupational labor-market outcomes evolved during a period of rapid AI advancement. We combine Frey and Osborne’s (2017) estimates of occupational computerization risk with Tomlinson et al.’s (2025) AI applicability scores based on Microsoft Copilot usage, matching both measures to O*NET occupational classifications and U.S. Bureau of Labor Statistics employment and wage data for 2019 through 2024. We first assess how recent developments in robotics, large language models, and generative AI have narrowed the engineering bottlenecks that Frey and Osborne identified in perception and manipulation, creative intelligence, and social intelligence. We then compare employment and wage changes across occupations with high and low AI applicability scores and across occupations classified as having high, moderate, or low probabilities of computerization. The results complicate a simple displacement narrative. Occupations with high AI applicability scores experienced overall employment and wage growth between 2019 and 2024, while occupations with high and moderate automation-risk scores experienced weaker employment performance than low-risk occupations. Wages increased across all automation-risk groupings. These patterns suggest that exposure to AI and automation does not map mechanically onto job loss, at least in the short run, and that task-based measures of technological exposure are better understood as indicators of occupational restructuring than as direct forecasts of employment decline.