When you can personally verify all the facts about biology, physics, and distant places that you rely on to make decisions every day, then you can throw stones about arguments from authority. The rest of us use a little bit of faith.
Too much reductionism can be a dangerous thing. This is pretty much the Sophist argument and it is not enough to build a scientific foundation on.
No, you don't need to personally verify all facts. But you need to be able to observe several indepedent validations. Yes, somewhere you have to take someone's word for fact, but multiple independent observations is a close approximation to the perfect scientific method. Related to that is the fact that science is not a democracy, and these observations have to be weighed against how they fit in our generally accepted worldview. Theory is more than just facts.
Every little detail of science has to be questioned, just not by everyone personally. Where that limit for healthy skepticism is drawn is not something can answer logically, but we can recognize if it is completely wrong.
It's perfectly okay to stay agnostic on things you don't know about. It's always okay to say "I don't know." If you were forced to take a position on something you don't know about yourself, then looking to relevant authorities may be a decent enough plan, but who's forcing you to take positions on things?
You can trim that down quite a lot by assuming things aren't magic and use some minimal trust. If people say that cars are designed using newtonian physics, and you see your car is working, then you can say you have tested newtonian physics even if you don't understand the calculations.
Basically you test what you have and trust the description of the implementation. If it works you can trust the whole chain. This is of course not 100%, but it is about as reliable as it can get for an individual.
What you shouldn't do is assume that stuff that aren't 100% for sure are equally bad.
It's quite easy to test this very assumption against chaos theory in a controlled environment and demonstrate why a system involving a complex chain of dependencies is increasingly unpredictable down the chain.
As an uninformed consumer at or the bottom of the chain, you cannot trust that an interface that works today, will work tomorrow, for you do not have access to all of the chaotic variables which constrain the interface.
This is exactly why the appeal to authority is such a useful heuristic: when it comes to physics, I trust the man who teaches the engineers that car companies employ to make cars - because I can see cars work every day, so at the end of the chain there must exist a person who has fundamental insights about how the world works.
So while I can't prove something is right by appeal to authority, it's a very practical way for me to swift though the unbounded mass of claims and information that I couldn't possibly verify or prove in my lifetime, not even to a 19th century level of scientific knowledge.
For example, I saw an onion cell once at a school microscope, but I have never seen a bacteria. So as far as I can independently prove, I live in times before the germ theory of disease - but that doesn't mean I shouldn't trust my doctor when he recommends vaccines and antibiotics. The entire antivaxx movement is just an attempt to challenge the medical authority by people who are not intellectually equipped to verify the scientific claims.
