The anthropic principle is useful for rationally debunking "... but surely..." arguments, but it's definitely a last resort, because it doesn't tell you anythin

The anthropic principle is useful for rationally debunking "... but surely..." arguments, but it's definitely a last resort, because it doesn't tell you anything about how or why. It's a form of proof by reductio ad absurdum. You assume the counterfactual ("what if the universe couldn't support human life?"), derive a contradiction ("the universe can't support human life, yet here we are discussing it"), and conclude the phenomenon you're trying to explain is obviously true ("therefore we shouldn't be surprised that the universe supports human life"). In mathematics, reductio is a valid proof technique, but one of the least interesting, because it doesn't teach you much about the domain, whereas a proof by construction (for example) may suggest related theorems or techniques to explore. You can explain gravity by the anthropic principle (without gravity, we wouldn't have stars or planets, so human life as we know it could never have evolved; yet here we are discussing it), but we much prefer an explanation like "matter curves spacetime so that nearby matter feels an acceleration toward it". It gives us tools to visualise and calculate the phenomenon. It also gives us a narrative, albeit a very limited one (we don't know why matter curves spacetime!), and human understanding is very dependent on narratives.