I'm blanking on the exact phrase, but it's something like "never believe a number with unreported error".

To get further into the weeds there is a significant difference in approach between theoretical and experimental science. In experimental science it's not only enough to communicate what you "know" but to communicate the underlying biased, tolerances and precisions of the thing being measured and modeling approach being used.

these represent the threshold of the known.

I would argue that those representations are inherently bad science because they do not communicate the margin of error. Grue, I believe you are spot on with a concept in how you would make those drawings more scientifically accurate, but ultimately they are artistic renderings of scientific understandings, but not scientific themselves.

While I don't disagree with WoodScientist that modern scientific institutions are inherently conservative, the process of science is not, nor should it be. Apologizing for the inherent conservatism in science is unscientific, harms belief in vetted resulted, conflates institutions for processes and projects a people problem onto the inanimate.