I know that at the end of lecture we talked about using one kernel across all of the different scales of the image but I am still confused about what this does for us. How would the same kernel be able to detect the sunflower center when they are different sizes (based on what scaled image you are looking at)? And also once we identify a local maximum as this sunflower center across all scales, what do we do with this?

The Laplacian kernel used in this example is only designed to detect sunflowers of a particular size. By convolving the Laplacian kernel of a certain size with an image, sunflowers of the "same" size will produce a strong response. To detect sunflowers of different sizes, we can either (i) change the scale of the kernel, or (ii) change the scale of the image.

So this helps us detect not only the spatial location of a feature in an image (in this case, a sunflower), but it also allows us to get the scale of the feature. We can use both scale and positional information as part of a descriptor---as discussed in the lecture on Feature Detectors and Descriptors.