In a context of an interview it means how performance of an algorithm scales with respect to the input length. This is measured in Big O notation. Understanding that is seldom very helpful on the job, but it's very important for interviews, so make sure you learn it.
Hahaha I appreciate the candor. Frankfurt's On Bullshit has never been more appropriate than at this time in my life and in history.
That's not specific enough. Memory usage, elapsed time, and energy usage are all functions of the input size, and each is a different optimization problem. Instead of talking about performance optimization, you should be talking about memory optimization or speed optimization.