>It won't speed up the singularity. All techs seem to follow some form of exponential curve

I understand what you mean. But when you zoom in on an exponential curve it's often made up of smaller S curves.

Increasing the efficiency, accuracy, and speed of programming, while reducing the cost should theoretically give a boost to the S curves across different sectors.

The nature of an S curve is it will plateau for a while before accelerating again, a large enough upward S curve near the end of the exponential curve could be a sign of an impending take off.

Sorry this is hard to explain without drawing a diagram.

I feel as though this could be a signal toward much closer transformative AI than expected by most people. Maybe not the singularity yet, but a revolution nonetheless.