Small error in a video (I think?)

Hi, I’m auditing the specialization and in course 1 week 3 video: Decision boundary, I found an error at 9:34

Prof Andrew says and demonstrates multiple times in this video that values ‘outside’ of error boundary are considered 1 and ‘inside’ the error boundary are considered 0 (which makes sense because inside the boundary, we are closer to the axes intersection therefore closer to 0).
But at about 9:34, he has drawn an ellipse and a weird shape depicting complex error boundaries. Then he continues to say that the y hat (predicted) values outside these boundaries are considered 0 and inside are considered 1 which contradicts what he said earlier.

If this is a misunderstanding on my part, please help me understand more clearly.

Thanks,
Prakhar

That’s not an error.
Which side of the boundary has the 0’s or 1’s depends on the signs of the weight values.