Can Yang-Baxter imply Lie algebra?

Quantum knot invariants (like colored HOMFLY-PT or Kauffman polynomials) are a distinguished class of non-perturbative topological invariants. Any known way to construct them (via Chern-Simons theory or quantum R-matrix) starts with a finite simple Lie algebra. Another set of knot invariants - of finite type - is related to quantum invariants via a perturbative expansion. However can all finite type invariants be obtained in this way? Investigating this problem, P. Vogel discovered a way to polynomially parameterize the expansion coefficients with three parameters so that, at different specific values, this reproduces the answers for all simple Lie (super)algebras. Then it is easy to construct a polynomial $P_{alg}$ that vanishes for all simple Lie algebras, and the corresponding Vassiliev invariant would thus be absent from the perturbative expansion. We review these Vogel claims pointing out at least two interesting implications of his construction. First, we discuss whether infinite-dimensional Lie algebras might enlarge Chern-Simons theory. Second, Vogel's construction implies an alternative axiomatization of simple Lie algebras - when we start from knot invariants and arrive at Lie algebras and their classification, which is opposite to conventional logic that we mentioned at the beginning.