How can I show that $A=\begin{pmatrix}a&b\\b&d\end{pmatrix}$ with $b\neq 0$ is diagonalizable?
Sei $b\neq 0$, dann gilt:$$\det(A-\lambda E)=\det \begin{pmatrix} a-\lambda & b \\ b & d-\lambda \end{pmatrix}=(a-\lambda)(d-\lambda)-b^2=\lambda ^2-\lambda (a+d)+ad-b^2 $$ $$\lambda_{1,2}=\frac{a+d}{2}\pm\sqrt{\frac{(a+d)^2}{4}-ad+b^2}=\frac{a+d}{2}\pm \sqrt{\frac{1}{4}(d-a)^2+b^2}$$ Since $\frac{1}{4}(d-a)^2\geq 0$ for all $d,a\in \mathbb{R}$ and $b\neq 0$ (and thus $b^2>0$), we have $\sqrt{\frac{1}{4}(d-a)^2+b^2}>0$. Hence, we have to different Eigenvalues.
Do I have to brute force this problem? Or can I show it without going through all these tedious computations?
Here's a different option.
Instead of saying to yourself "Those were tedious computations", you could instead say "What have I really proved?"
Proof: (tedious but suddenly, perhaps, more interesting computations)
Proof: $\text{trace}^2(M) - 4 \, \text{det}(M)$ is the discriminant of the characteristic polynomial, therefore if it is positive then there are two distinct characteristic roots. QED
Proof: $\text{trace}^2(M) - 4 \, \text{det}(M) = (a+d)^2 - 4 (ad-bb^2) =$ (still one last tedious, but perhaps more enlightening, computation) $= (a-d)^2 + 4b^2 > 0$.
And now you can start to ask yourself: What happens with $3 \times 3$ matrices?.... or $4 \times 4$? ............