Military conscription is widespread, with many men around the world performing months of service. While the effects of service on a variety of outcomes have been studied, little is known about what skills are learned during service. We use the Danish draft lottery in combination with skills data from the Program for the International Assessment of Adult Competencies (PIACC) to estimate what is learned from military service. We observe measures of abilities from the military conscription board measured at age 18 and a rich set of background variables through administrative registers. Some men who are assessed fit for military service are randomly assigned to serve. We exploit this random assignment to credibly estimate the causal effect of peacetime military service on skills. To circumvent difficulties about measurement errors, we use plausible values from the PIACC data. While military service is associated with reduced skills, we find a positive causal effect on skills measured 10 years after service – problem solving, numeracy and especially literacy – for men forced to serve who otherwise would not have served. The positive service effect on skills is largest for high ability men. Our findings have three important implications. First, the average costs of mandatory service, in terms of delayed human capital accumulation and reduced earnings, is hiding some positive effects on skills. Second, the skill formation is unequally distributed, with high ability men benefiting the most through higher skills. Third, our findings have some policy implications showing that intensive training around age 20 can improve skills.