I wouldn't say it's that bad. Obviously, training set is very small, much smaller than Obama-RNN's one. Hence all the repetitiveness, it just doesn't have enough examples of what more could be said. But I would say that it has even more sentences that actually make sense than Obama-RNN's output, probably because of average sentence being shorter.
This is good, but nothing compares to Borat warning Jimmy Kimmel that the nincompoop Sacha Baron Cohen has fooled the world with his fictional and offensive characters Ali G, Bruno and Donald Trump. Then when pointed out by Jimmy Kimmel that Trump is actually real, still disbelieves Kimmel because "the only person who would ban Muslims is a person with a brain like a female chicken. It is a clearly a parody of a rich, American racist".
The most interesting thing about Trump's popularity is that the Greeks had invented a word for the style and method he uses because people were already stooping to lowest common denominators 2500 years ago when democracies came about:
and for those who think there is no way this country would elect someone like Trump, I have another wikipedia entry for you to read about the elected leader of Italy with scary parallels to Trump that the news never talks about:
I can't help to hear the voice of the Admiral Ackbar now in my head: "It's a Trump!!, it's a Trump!!". Are Mon Calamari guys officially banned to enter USA yet?
10 comments
[ 3.9 ms ] story [ 36.5 ms ] threadPerhaps the Trump corpus could be fed into char-rnn for comparison and figure out why the output is so nonsensical?
How much sample text do you need, and is there even a way of answering that question?
https://wikipedia.org/wiki/Demagogue
and for those who think there is no way this country would elect someone like Trump, I have another wikipedia entry for you to read about the elected leader of Italy with scary parallels to Trump that the news never talks about:
https://wikipedia.org/wiki/Silvio_Berlusconi