Code reviews where some coworker looks at the changes, types "LGTM", and proceeds to approve it do next to nothing to prevent bugs and bad decisions. Code revi

Code reviews where some coworker looks at the changes, types "LGTM", and proceeds to approve it do next to nothing to prevent bugs and bad decisions. Code reviews where some coworker runs the code on their machine and has plenty of time and power to express the ways it could break or it could be improved are useful. When the former review happens more often than the latter, if often has to do with soft skills rather than hard skills. Some requirements for a good review are honesty, mutual respect, and ability to handle criticism. Other than that, some engineers struggle to understand there may be more than one good approach to a problem. Ultimately, engineers are not entirely to blame for bad reviews. In my experience it is often a company culture issue. Authoritarian leadership trickles down to middle management, senior engineers, etc. If people don't feel empowered to challenge others code regardless of their seniority or perceived status, weird or downright bad technical decisions will be made.