Attention is just Kernel Smoothing?

I was reading this blog post, where the author explains that attention is just a re-invention of kernel smoothing.