A huge breakthrough in approximation algorithms was announced recently as asymmetric travelling salesman problem was shown to allow a constant approximation scheme. See discussion in an article by R.J. Lipton. One of the co-authors was Jakub dj3500 Tarnawski. It always pleases me to see competitive programmers achieving heights beyond CP, in academia and "real-life" problems (recall that OpenAI's bot recently beat human players at 1v1 Dota2, with meret and Psyho in the developers team). Congratulations on an outstanding achievement, Jakub!