There are different degrees of certainty. Unfortunately much of that is lost when it gets communicated to the world beyond those closely involved. Lots of things get reported as fast but omitting all the caveats and clarifications about the scope. If you want an actually relevant recent example, the way we think about depression from a neurobiological perspective has been changing dramatically. The focus on serotonin because of SSRIs really led the field in the wrong direction imo and in the opinion of many people who are in that field (I'm not even in neuro let alone that subfield so my opinions are just opinions and I'm by no means an expert)
But, for the most part, responsible scientists have been clear about the lack of certainty. And even when not stated explicitly in every single paper, it's something that comes up about a field as a whole.

Homeopathy having an effect would require our most fundamental understanding of some very basic concepts to be completely wrong. Water molecules having memory might as well be magic. I can't say it's impossible, but it's the same way I can't say it's impossible that the world is controlled by faeries.

Empirically speaking, homeopathy has not demonstrated an effect over placebo. So there's no reason to believe it works in practice or in theory. There's no compelling evidence whatsoever, or any unexplained phenomenon being explained.