Markovbot creates eerily plausible Drumpf speeches


#1

[Read the post]


#2

Good job, Frostworks! Much more efficient than my method of transcribing backwards metal tracks.


#3

Drumpf’ing the Christie, as it was called in the old country.


#4

A Markov chain generator seems like overkill. Random selections from a short list of words or phrases would probably simulate Drumpf speeches even better, given his incredibly limited vocabulary and content-less utterances.


#5

But I don’t think he repeats himself too much, so a Markov chain seems fine. Now that random repeat a limited vocabulary might work for the Rubio bot.


#6

I dunno, I think just "loser, “dumb,” “moron,” “huge,” “great,” “classy,” and “tremendous” would cover about half of his actual word-count…


#7

Maybe next they can create one that generates Bernard Gutman speeches, or William Blythe speeches, even. Names which were changed post-17th-century.


#8

They seem to almost but not quite make sense, like the mad speeches of a dream-Trump.


#9

If a bot convinces an audience that it is Drumpf, does that mean that it passes the Turing Test, or that Drumpf fails?


#10

Electing Trump to run America line a business, eh?

So he’ll fire anybody he thinks isn’t contributing to profits, and then declare bankruptcy after maxing out his stock value?


#11

On the plus side, he can outsource Congress. Maybe the new one will do something productive once in a while.


#12

low hanging fruit, I know


#13

Well, considering the GOP debates have devolved into discussing the candidate’s twig and berries, I’d say low hanging fruit is fair game.


#14

This topic was automatically closed after 5 days. New replies are no longer allowed.