Which answer is correct?

Always use list comprehensions. Generators are slower because of the overhead of pausing and resuming.

— list_fan

Use generators when you only need to iterate once and the sequence is large (or infinite). They produce values lazily, using O(1) memory instead of O(n).

— lazy_eval

Use generators everywhere. They're always more efficient because they don't build the full list in memory.

— gen_everywhere

They're identical. Python optimizes list comprehensions into generators internally.

— no_difference