Here ∠A=∠DAB, ∠B=∠ABC, ∠C=∠BCD, ∠D=∠CDA. Since ABCD is non-planar, D is not in the plane of triangle ABC.
Key fact (angles at a corner in space). If rays AB, AC, AD do not lie in one plane, then
∠DAB<∠DAC+∠CAB
Why (an intuitive argument that follows the hint; a full proof needs ideas from higher classes): keep △ABC fixed and turn △ADC about the hinge AC. The lengths AB, AD and the angles ∠CAB, ∠DAC do not change, but the distance BD does. BD is largest when the two triangles lie flat in one plane on opposite sides of AC; then ∠DAB is exactly ∠DAC+∠CAB. (If this sum is 180∘ or more, the inequality is clear anyway, since ∠DAB<180∘.) For every other position (D out of the plane) BD is shorter, and in △DAB, with AB and AD fixed, a shorter third side BD means a smaller angle between them — just as the tips of a pair of compasses come closer when you close it. So ∠DAB<∠DAC+∠CAB. The same holds at C: ∠BCD<∠BCA+∠ACD.
The sum is less than 360∘.
-
∠A<∠CAB+∠DAC and ∠C<∠BCA+∠ACD — key fact.
-
∠CAB+∠BCA+∠B=180∘ — angle sum of △ABC.
-
∠DAC+∠ACD+∠D=180∘ — angle sum of △ADC.
-
Adding: ∠A+∠B+∠C+∠D<180∘+180∘=360∘.
What happens during the rotation. As triangle ADC turns about AC, ∠B and ∠D stay the same (each triangle is rigid), while ∠A and ∠C change: they are largest when the figure is flat with B and D on opposite sides of AC (sum 360∘), and they shrink steadily as D swings round towards the side of B. If the two triangles are congruent mirror images (a kite), ∠A and ∠C shrink towards 0∘ as D approaches B.
A sum of 2∘. Take a very thin rhombus: AB=BC=CD=DA=10 cm with ∠B=∠D=0.5∘ (so AC is very short, less than 1 mm). Flat, its angle sum is 360∘. Now fold it along AC. ∠B and ∠D stay at 0.5∘, so ∠B+∠D=1∘, while ∠A+∠C shrinks from 359∘ towards 0∘ as the fold closes. So the sum moves steadily from 360∘ down towards 1∘, and at some fold it equals exactly 2∘ — when ∠A+∠C=1∘, that is ∠A=∠C=0.5∘.
Measured with coordinates, this happens when the two triangles are opened to an angle of about 0.5∘ from each other: then ∠A=∠B=∠C=∠D≈0.5∘, so the sum is 2∘. At that position D is not in the plane of ABC, so ABCD is a non-planar quadrilateral with ∠A+∠B+∠C+∠D=2∘. Yes, such a quadrilateral exists.