For decades it was open whether sensitivity $\mathrm{s}(f)$ -- the most local measure in the zoo -- is polynomially related to the rest. [Nisan91] and [NS94] showed everything else is polynomially related: $$\mathrm{s} \le \mathrm{bs} \le \mathrm{C} \le \mathrm{s}\cdot\mathrm{bs},\qquad \mathrm{bs}\le 2\deg^2,\qquad \mathrm{D}\le 2\deg^4 .$$ The missing link was $\deg$ vs $\mathrm{s}$.
Huang's theorem (2019): $\deg(f) \le \mathrm{s}(f)^2$.
The proof is two pages: a signed adjacency matrix of the hypercube whose eigenvalues are $\pm\sqrt n$, plus Cauchy interlacing, gives that every induced subgraph of $Q_n$ on more than $2^{n-1}$ vertices has maximum degree $\ge\sqrt n$. [ABKRT21] later gave the spectral refinement $\deg(f)\le\lambda(f)^2$ with $\lambda\le\mathrm{s}$, which is where \(\lambda(f)\) comes from.
What is still open
The exponents. Currently $$\mathrm{bs}(f) \le 2\deg(f)^2 \le 2\mathrm{s}(f)^4.$$ The quadratic strengthening was refuted by [OpenAISensitivity26]. The latest public campaign theorem [Prove2MeSensitivity26] gives $\mathrm{bs}\ge\mathrm{s}^{\alpha-\varepsilon}$ for every $\varepsilon>0$ and arbitrarily large sensitivity, where $\alpha=\log(14055^8)/\log(3738277295413674)=2.1308255793978239\ldots$. The quartic universal upper bound remains. See What is the optimal exponent of … versus …?.