We never can know something with 100% certainty, there is always a chance that we have been making some consistent mistake all along. We do have good trust in theories that have been developed for over a century and have passed all kinds of cross-tests. Sure, every now and then, a result that seems contradictory to what we know pops up - it doesn't mean that we should immediately say, "Ah, our theory is wrong. Scrap it!", it means that we should spend time to understand what this result means. Might be an error in the experiment ("FTL" neutrinos in the OPERA experiment in 2011 is a good example of that), might be wrong interpretation of the results. Might be something that our theory has been missing, and we can now incorporate it into it.
The thing is, physics doesn't work the way, say, math does. In math, you can have a theorem, come up with the counter-example and say, "See, this theorem is incorrect" - and that's it, this theorem will be busted once and for all. In physics, there is a huge room for interpretation of results, so just one weird result doesn't mean anything. Produce this result in multiple experiments consistently, rule out systematic errors, build a few computer simulations of the experiment and produce the results that agree with each other and contradict what we see - and only THEN can we say that we may have stumbled upon something new, requiring revising our theories.
There is a long process of approval of physical results, and for a good reason. It doesn't work like, "I just linked you the article. I am right, and you are wrong!" You can link articles all you want, but they don't mean much in themselves, no matter how peer reviewed they are. That's the point Garnier Fructis was making.
True, but we haven't concluded the opposite either. According to the null hypothesis, there is little reason to take cosmological principle as a fact. Maybe some loose version of it, some broader and more careful interpretation...
What I'm about to say is very tangential, but this post reminded me of something humorous. Most mistakes in math research begin with the words 'It is clear that...'
Which is why you have 'theorems' that are wrong, because people simply assumed something for being intuitively 'obvious.'
For new things, yeah. But you can find errors with much higher frequency in books, like incorrect results or incorrect proofs of correct results. So even if you're learning foundational material, you still have to be wary when details are swept under the rug for being 'clear.'