I’m curious, because it was an interesting question, and would like someone to please clarify / help me determine if my statement is actually true.
So by subtraction of co-ordinates and applying Pythagoras, they could use the co-ordinate values to ensure the distances are greater than s. Thanks. Are there any better ways of doing this?


As rightly pointed out by @TobyMak one condition is to determine if any vertex of a triangle is, or is not, inside another triangle.
However, if all vertices of a triangle are outside another, this will also include the case that the two triangles intersects.
So we shall counter-test also for this condition, if overlapping shall be excluded.
Now to test if a point is external to a given triangle, we have better use the barycentric coordinates approach.
Given the triangle $ABC$ with vertices $(x_A, y_A),\, (x_B, y_B), \, (x_C, y_C)$ we construct the matrix $$ {\bf T} = \left( {\matrix{ {x_A } & {x_B } & {x_C } \cr {y_A } & {y_B } & {y_C } \cr 1 & 1 & 1 \cr } } \right) $$
A point $P=(x,y)$ will be internal to the triangle if its barycentric coordinates wrt $\triangle ABC$ are all positive, i.e. iff $$ \left( {\matrix{ {\lambda _{\,1} } \cr {\lambda _{\,2} } \cr {\lambda _{\,3} } \cr } } \right) = {\bf T}^{\, - \,{\bf 1}} \left( {\matrix{ x \cr y \cr 1 \cr } } \right) \quad \Rightarrow \quad 0 < \lambda _{\,1} ,\lambda _{\,2} ,\lambda _{\,3} $$
The $\lambda _k$ sum to $1$, so they cannot be all negative.
If one at least is negative, then the point $P$ is external.
And if it is $\lambda_1$ to be negative, then you know that $P$ is on the side of the edge $BC$ opposite to $A$.
So, if you take the $\triangle XYZ$ and modify its coordinates by a translation and a rotation $$ \eqalign{ & {\bf X} = \left( {\matrix{ {x_X } & {x_Y } & {x_Z } \cr {y_X } & {y_Y } & {y_Z } \cr 1 & 1 & 1 \cr } } \right)\quad \Rightarrow \cr & {\bf X}' = \left( {\matrix{ {\cos \alpha } & { - \sin \alpha } & 0 \cr {\sin \alpha } & {\cos \alpha } & 0 \cr 0 & 0 & 1 \cr } } \right)\left( {\left( {\matrix{ {x_X } & {x_Y } & {x_Z } \cr {y_X } & {y_Y } & {y_Z } \cr 1 & 1 & 1 \cr } } \right) + \left( {\matrix{ u & u & u \cr v & v & v \cr 0 & 0 & 0 \cr } } \right)} \right) \cr} $$ then by taking the product $$ {\bf \Lambda } = {\bf T}^{\, - \,{\bf 1}} \;{\bf X}' $$ you can adjust the parameters $\alpha, u, v$ in such a way that at least one component in every column of $\Lambda$ be negative, to ensure that all the vertices of $\triangle X'Y'Z'$ are external to $\triangle ABC$.
Finally, to test that the two triangles do not overlap, we take a generic point $Q$ internal to $\triangle X'Y'Z'$ $$ Q = {\bf X}'\, \left( {\matrix{ {\mu _{\,1} } \cr {\mu _{\,2} } \cr {\mu _{\,3} } \cr } } \right) \quad \left| \matrix{ \,0 \le \mu _{\,1} ,\mu _{\,2} ,\mu _{\,3} \left( { \le 1} \right) \hfill \cr \,\mu _{\,1} + \mu _{\,2} + \mu _{\,3} = 1 \hfill \cr} \right. $$ and test that not all the components of $$ {\bf T}^{\, - \,{\bf 1}} Q = {\bf T}^{\, - \,{\bf 1}} {\bf X}'\left( {\matrix{ {\mu _{\,1} } \cr {\mu _{\,2} } \cr {\mu _{\,3} } \cr } } \right) = {\bf \Lambda }\left( {\matrix{ {\mu _{\,1} } \cr {\mu _{\,2} } \cr {\mu _{\,3} } \cr } } \right) $$ are positive for any of the allowed values of $\mu _k$.