Find all $4<n,m\in\mathbb N$ for which
$\cos^2\left(\frac{2\pi}n\right)=\cos\left(\frac{2\pi}m\right).$
I think it has one non-trivial solution, $n=8,m=6$. Intuitive idea for a proof: $\cos(x)\approx1-\frac{x^2}{2}$.
$\left(1-\left(\frac{2\pi}{n}\right)^2\right)^2\approx1-\left(\frac{2\pi}{m}\right)^2$.
$1-2\left(\frac{2\pi}{n}\right)^2+\left(\frac{2\pi}{n}\right)^4\approx1-\left(\frac{2\pi}{m}\right)^2$; $\sqrt2 m\approx n$.
On the other hand, $\mathbb Q\to\mathbb Q\left[\cos\left(\frac{2\pi}{m}\right)\right]\to \mathbb Q\left[\sqrt{\cos\left(\frac{2\pi}{m}\right)}\right]$ is of degree $2\varphi(m)$. So, $2\varphi(m)=\varphi(n)$.
From here, I think it is possible to formalize these arguments and arrive at a contradiction for large enough $m,n$.
Start by observing that $\cos(2 \pi/r)$ is an algebraic number and its conjugates are exactly the numbers $\cos(2 \pi k/r)$ for $(k,r) = 1$.
Since the conjugates of $\alpha^2$ are the squares of the conjugates of $\alpha$, it follows that the conjugates of the LHS are all positive real numbers.
OTOH, the conjugates of the RHS include $\cos(2 \pi k/m)$ for all $(m,k) = 1$. Hence, if there exists a $k$ with $(k,m)=1$ and
$$m/4 < k < 3m/4,$$
then $\cos(2 \pi k/m) < 0$ which would be a contradiction, implying that there cannot be any solutions to the original equation.
Define $k$ as follows:
It's easy to verify that $(k,m)=1$, since $2k-m \in \{1,2,4\}$ so the only common factor could be $2$, but $k$ is odd when $m$ is even.
It's also easy to verify the required inequality holds unless $m=1$, $m=2$, or $m=6$. For $m=1$ one has $n=1,2$, for $m=2$ the RHS is negative, and for $m=6$ there is the solution $n=8$.