The force vectors are along the rope which could be represented as the sum of force towards the other column, and a force to the side at 90 degrees...
That is the key point. Hence start with the rope taut for maximum "pull in vector".
As an alternative to starting with the rope taut try starting with the rope at 45
o. That could lead to a more intuitive grasp for a lay person. Try this rough sketch:
Now quantify the vectors. The logical "easy way" way through is:
Applied weight 2U (I even made it 2U AND set the rope at 45
o to make the outcomes simpler.

)
Therefore Av and Bv are each equal to U.
For 45
o the horizontal and vertical vectors are equal.
Therefore the inwards pull vectors Ah and Bh are equal and also equal to U
and that is for 45o rope inclination. Two points for the more pedantic members:
Point 1: The four rope tension vectors are "internal" to the system AND
The four attachment point vectors Av, Bv, Ah and Bh PLUS the weight vector 2U are "external" to the system. So:
Point 2: Those five external vectors must vector add to net zero - otherwise the whole system would be subject to an "external impressed force" and would go merrily accelerating in the direction of the resultant non zero vector.
Some bloke named Newton said that. Sadly he often gets misrepresented on these forums.
Progress with "real people testing".
I tried the issue out on a mature age ex builder now a "building certifier" this AM over coffee:
A) He had no doubt that there was "pull in";
B) Needed to think before accepting that the pull in could be far more than the applied load.
I still have difficulty accepting that an engineer would legitimately question the concept.
AND Newly made family arrangements will see my 6 yo grandson visiting this weekend. I may try the concept on him as suggested in my earlier post.
