I agree with his advice, and it feels like I'm giving that advice on a weekly basis to first and second year PhD students that think they can go to industry for 3-5 years and then get a professorship. Some students have told me it works that way in China, but it definitely isn't true in the United States. Unless you can publish prolifically at your company (Microsoft Research, FAIR, Google Research, etc.) or work at a government laboratory, your CV will quickly lose its competitive edge after you leave the university.<p>As an aside, just managing to be competitive for professorships after one's PhD or postdoc is an enormous challenge due to the scarcity of positions and the qualifications of competitors. Based on some limited analysis I did previously, the number of publications required to get an interview is often in the range of 10-30 [1,2].<p>[1] <a href="http://www.chriskanan.com/statistics-on-ucsds-computer-science-faculty-candidates/" rel="nofollow">http://www.chriskanan.com/statistics-on-ucsds-computer-scien...</a><p>[2] <a href="http://www.chriskanan.com/planning-for-life-after-your-phd/" rel="nofollow">http://www.chriskanan.com/planning-for-life-after-your-phd/</a>