I was a little worried because for version becuase I had, a couple of Oulivian constructions, you know, constrained literature constructions so it dealt with the n+7 algorithm, which requires choosing a dictionary and applying it to a pre-existing piece of text, so for the talks I had done before, I had done two different versions of this one poem, and for this talk I ended up generating like, four different versions, starting with an Emily Dickinson poem and then replacing the nouns in the poem according to what various dictionaries say and I had four or five different dictionaries that I applied to this poem and I, because TED Talks are not supposed to have a whole lot of background, like visuals, I memorized all the poems so I didn’t have to display them on the screen or something like that, so it was kind of hard keeping four or five short, but similar, but nonsensical poems. Because one, I used an etymological dictionary. One of them I used a maritime law dictionary. One of them, I used a fairly standard Webster’s Dictionary, and I don’t remember what the other one was, but of course, they're all very different.
