www.lesswrong.com/posts/8Pz2RddtmuCLjxe2s/is-there-such-a-thing-as-an-impossible...

The cited paper, “Distinguishing Proteins From Arbitrary Amino Acid Sequences,” is about a computational sequence test, not a definition of proteins as entities with a well-structured 3D shape.

In the paper’s abstract and introduction, the authors say they are developing a method for determining whether an arbitrary amino acid sequence can be a protein sequence. They represent sequences as points in a 60-dimensional “natural vector” space and test whether those points fall within convex hulls derived from known proteins. The paper explicitly says:

• “Structurally, proteins are sequences of amino acids.”
• “In this paper we will describe a 99.69% accurate computational test for determining if an arbitrary amino acid sequence can be a protein.”

Nothing in the paper’s abstract or main methodological description defines a protein as “anything that has a well-structured 3D shape.” That description is therefore a mischaracterization of the cited source.

So the problem here is not a debatable interpretation of protein science; it is that the post attributes a definition to the paper that the paper does not give.