The study explores routine and non-routine content of job tasks across occupations in the Russian labor market. Each occupation contains a bundle of job tasks, which can be routine or non-routine, cognitive or manual. Occupations, in which most job tasks are routine, are under the risk of automation, and therefore their future is of concern. Our empirical analysis uses three main data sources that contain information in the detailed occupational breakdown: on employment, on the extent of routinization, and on annual labor incomes. This data set allows to estimate task indices for disaggregated occupations and different socio-economic groups and also present empirical estimates of wage penalties and premiums for workers with different job content. The calculations suggest that the proportion of jobs in which routine tasks tend to prevail is not large and hardly exceeds 10%. From this follows that a massive substitution of labor by machines or AI in foreseeable future does not look like a plausible option. Content of job tasks is expectedly associated with the level of pay: workers with non-routine cognitive tasks are better rewarded, and non-routine manual tasks are the most penalized. A polarization scenario feared by some observers does not seem a likely option in Russia for years to come.