Second hint: if you can't "see" the solution from the hint above and/or find the problem too difficult, try a warm-up with the problem in 2D instead of 3D, as follows.
2D version of the problem: Consider, in the $y,z$ plane, the set of points $S=\{(y,z):z\leq \min(\beta y, \delta y)\}$. Suppose then that for some $s\in\mathbb{R}$ it happens that, for every $(y,z)\in S$, we have $z\leq sy$. Prove that in this case there is a number $t\in\mathbb{R}$ with $0\leq t\leq 1$ such that $s=t\beta + (1-t)\delta$.
N.B. do not just prove it. See it. In particular, draw the two lines $z(y)=\beta y$ and $z(y)=\delta y$, see that $S$ is the region of the plane "below" both lines, and see where the generic line $z_t(y)=\left(t\beta + (1-t)\delta\right)y$ lies. Then see how any line in the form $z(y)=sy$ with no portion of $S$ above it, must have $s=\left(t\beta + (1-t)\delta\right)y$, for $0\leq t\leq 1$.
From 2D to 3D: As in the first hint above, consider the two planes $\pi: z=\alpha x + \beta y$ and $\sigma: z=\gamma x + \delta y$, both including the origin. In this case, $S$ is the portion of the Euclidean space the falls "below" both planes. If the two planes are distinct, their intersection, which obviously contains the origin, must be a line. If you rotate the axes so that the new $x$ axis coincides with this line, and look at the $y,z$ plane, you are back to the the 2D version of the problem (and the proof of the 2D version yields a proof of the 3D version).
Bonus points! How can you extend the problem to $4$ dimensions? How can you extend it to $k$ dimensions for any $k\in\mathbb{N}$?