While substantial scholarship has focused on estimating the susceptibility of jobs to automation, little has examined how job contents evolve in the information age despite the fact that new technologies typically substitute for specific job tasks, shifting job skills rather than eliminating whole jobs. Here we explore the process and consequences of changes in occupational skill contents and characterize occupations subject to the most re-skilling pressure. Recent research suggests that high-skilled STEM and technology-intensive business occupations have experienced the highest rates of skill content change. Using a dataset covering the near universe of U.S. online job postings between 2010 and 2018, we find that when the number and similarity of skills within a job are taken into account, the re-skilling pressure is much higher for workers in low complexity, low education and low compensation occupations. We use high-dimensional embeddings of skills estimated across all jobs to precisely assess skill similarity, and characterize occupational skill transformations, demonstrating that skills requiring machine-operation and interface rise sharply in importance in the past decade, much more than human interface skills in low and mid-education occupations. We establish that large organizations buffer jobs from skill instability and obsolescence, especially low-skilled jobs with unstable skill requirements. Finally, the gap in re-skilling pressure between low/mid-education and high-education occupations is smaller in large organizations, suggesting that by controlling the surrounding skill environment, such organizations reduce the rate of required re-skilling and sustain short-term productivity for those occupations.