[Lex Computer & Tech Group/LCTG] Re. AI as a visual geometry problem solver

[Jon Dreyer]

Surely we can stretch the problem in all sorts of ways: maybe it's not 
planar, maybe apparently straight lines aren't straight, maybe it's on 
the surface of a sphere, maybe the x is a Roman numeral 10, etc. But 
there is a fairly standard language of geometry diagrams. Typically, 
geometry diagrams are assumed planar unless we're told otherwise, and 
lines that appear straight are assumed straight unless we're told 
otherwise. But angles that look right cannot be assumed to be right 
angles unless they have that little right-angle-indicator square at the 
vertex.

So according to this fairly standard interpretation, there's only one 
right answer. Of course the problem is (probably intentionally) 
misleading because the 80° angle does look closer to 90°.

-

Jon Dreyer
Math Tutor/Computer Science Tutor <http://www.passionatelycurious.com>
Jon Dreyer Music <http://music.jondreyer.com>

On 2/21/26 17:28, Richard Wagner via LCTG wrote:
>
> Just saw this emailed problem and accompanying thread, etc. I am in 
> Bob's corner to the extent that the problem is not simple. Not enough 
> information is given in the problem and assumptions must be made where 
> different assumptions can be correct!
>
> Is the object planar or are we seeing a 3D projection of it. If the 
> latter, are we viewing it where one side is planar from our vantage? 
> If the right side is planar to us, then the answer is 30 degrees from 
> the law of sum of angles in a triangle and the angle at the bisection 
> of the vertical line and horizontal line _appears_ to be  90 degrees. 
> (From the Pythagorean Theorem {a squared + b squared = c squared}, the 
> ratio of the sides is 1:square root of 3:2 for opposite sides of a 
> 30:60:90 degree triangle. I'll never forget this ratio and formula 
> from high school /ad nauseam)./ If we're seeing the left triangle as a 
> planar projection of a 3D object, the angle of the left triangle at 
> the vertical/horizontal bisection must be 80 degrees for the given 
> isosceles triangle. This leaves the angle on the other side of this 
> bisection open to question. Naively, one would assume it to be 100 
> degrees, if one assumes the "horizontal" line to be straight, and x to 
> therefore be 20 degrees. However, one can't assume the bottom line to 
> be horizontal if the object is a projection, and from the information 
> given we can't determine angle at the vertical/horizontal bisection 
> for the triangle on the right side of the projection.
>
> Thus, I think the problem is unsolvable from the information given, 
> and I'm not surprised that AI is confused and gave conflicting 
> results. I agree with Olga: be cognizant to what AI is doing and don't 
> summarily assume it is giving correct answers to your queries and 
> conversations.
>
> -Dick
>
-
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.toku.us/pipermail/lctg-toku.us/attachments/20260221/02e38a33/attachment.htm>