

Depends on what kind of mentor. A lot of times, having someone who knows tips and tricks or has seen the problem before can be really helpful. Even just watching how someone else does it can be helpful.
For example, at work we have tests written in C in the 90s to 2000s to control equipment and test our products. The way some of the new test engineers and interns would fix things would be to change a parameter and recompile the test, instead of adding a variable to the watch window, using breakpoints, stepping through functions, etc. Showing them that those tools existed was huge for troubleshooting problems whenever something breaks (which is often).
The bad kind of mentor is one who doesn’t/cant explain why doing it one way may have tradeoffs as opposed to the way you would do it. Their job is to 1) set an example for good practices and 2) guide you when you are doing something in a less efficient or self sabotaging way.
To be more specific, the fuel tank was placed between the rear bumper and rear differential. In a rear end collision, the tank would get sandwiched by the bumper and differential, which had bolts protruding out the back and would pierce the tank, spilling fuel onto the road.
Additionally, rear end collisions would bend the frame in a way that jammed the doors so you couldnt get out.
They figured that people would die and their cost benefit analysis assumed a certain number of deaths and lawsuits. The resulting recall and larger than expected number of deaths and lawsuits made it a huge loss for them.