For me it was the realisation that everything is built atop of assumptions and at the bottom the answer is always 'not sure' or idealised.

Science is a methodology but anything suggested by institutions on behalf of science cannot be trusted unless repeated.

I now look at most of the models as 'good enough' for the modern world. But not truth and it's kind of sludge of useful stuff and falsehoods.

The useful stuff is typically split into numerous topics and never actually taught well in effort to make it commoditized. This most noticeable when looking at technicians, engineers, researchers, and institutions.

Everything comes from the top and is presented down to the lower tiers and they are given their own specialist experience. These continue to make it so that the left hand doesn't know what the right is doing at all levels of the scientific establishment.

Engineers can depend on their assumptions. Techs can depend on theirs, researchers in theirs, etc.

All of them are usually slightly incompatible with one another and would cause too many questions if you were to apply it.

They also like to change the units slightly, and make different rules of thumb. All of this is useful at the level, but not consistent across them.

It allows the flaws to hide more easily because people working in string theory never need to apply it to technology. And engineering will not be questioning it, etc.

Same with climate science, geophysics, geology, oceanology etc. They all take in different assumptions and use them to work, but they don't work across the other disciplines and don't see the relationships.

Instead they are told how to interpret and transform between their results and the other specialists.

That is why science cannot tell the truth these days. People aren't practicing it fully. They have removed themselves from the real practice.