Why people ask it
Because it is a genuinely serious proposal that many physicists take seriously, not a dodge.
Short answer
It is a legitimate hypothesis. Its difficulties are that it is currently untestable, and that a mechanism producing universes would itself need to be finely tuned to do so.
Deeper answer
Treat this one with respect. Some multiverse models arise from serious physics rather than being invented to escape design, and dismissing it makes you look uninformed.
Two difficulties are worth raising. First, testability: if other universes are in principle unobservable, the hypothesis sits outside the empirical method it is meant to defend. Second, the generator problem: whatever produces universes with varied constants would need specific properties to do so, which relocates the fine-tuning rather than removing it.
There is also a methodological point. Postulating an enormous number of unobserved entities to avoid one inference is a substantial cost. That does not make it wrong, but it means both explanations are paying a price and the comparison should be made honestly.
Common mistake
Calling it a desperate invention. Some versions come out of serious physics, and the sneer costs you the room.
Conversation-ready response
""That is a real hypothesis and I take it seriously. Two questions I would want answered: how would we test it, and what fine-tuned the universe generator?""
Source notes coming soon.