Humanity, we may have a problem.
Computers are getting good at imitating us. How good? Good enough to fool us into accepting their scientific papers and their robocalls. Might they have already passed the Turing Test? And what are bots good for, anyway?
Linguist Daniel Midgley is only human on this episode of Talk the Talk.
Listen to this episode
You can listen to all the episodes of Talk the Talk by pasting this URL into your podlistener.
http://danielmidgley.com/talkthetalk/talk_classic.xmlPromo
Show notes
February 21 is International Mother Language Day
http://www.un.org/en/events/motherlanguageday/
Hazel Sampson, last native speaker of Klallam, dies
http://www.reuters.com/article/2014/02/07/us-usa-klallam-death-idUSBREA1605W20140207
But young people are learning it. Here’s the Klallam Word of the Day on Facebook
https://www.facebook.com/KlallamWOTD
Weirdly realistic bots are denying they’re bots
http://io9.com/freakishly-realistic-telemarketing-robots-are-denying-t-1481050295
but they’re really just pre-recorded voices, activated by humans.
http://newsfeed.time.com/2013/12/17/robot-telemarketer-samantha-west/
So bots haven’t passed the Turing test
http://www.mind.ilstu.edu/curriculum/turing_machines/turing_test_and_machine_intelligence.php?modGUI=240&compGUI=1145&itemGUI=1956
The stakes are high: in The Mind’s I, Dennett and Hofstadter argue that if it passes the Turing test, it’s conscious.
http://en.wikipedia.org/wiki/Consciousness#cite_note-MindsI-55
Publishers are withdrawing 120 papers that are complete gibberish
http://chronicle.com/blogs/ticker/publishers-withdraw-more-than-120-fake-computer-generated-papers/73479
http://www.nature.com/news/publishers-withdraw-more-than-120-gibberish-papers-1.14763
Here’s the site for SCIgen, the program behind the mischief.
http://pdos.csail.mit.edu/scigen/
Here are the Perl files for the SCIgen program
https://github.com/technoskald/scigen
but what you really want to look at is the scirules.in file
https://github.com/technoskald/scigen/blob/master/scirules.in
Try your own Google Scholar searches to find real published gibberish papers! Here’s the one for “have been far-reaching and pervasive”. If the “implications” have been f-r & p, you’ve likely found a fake paper. And that’s just one rule.
http://scholar.google.com.au/scholar?q=%22have+been+far-reaching+and+pervasive%22&btnG=&hl=en&as_sdt=0%2C5
Tesla Motors is building a ‘gigafactory’
http://www.bloomberg.com/news/2014-02-27/musk-s-5-billion-tesla-gigafactory-may-start-bidding-war.html
Origin of giga-
http://www.etymonline.com/index.php?term=giga-
No tera-factories yet, but it’s only a matter of time. And check out the recent interest in the term ‘giga-factory’.
http://www.google.com.au/trends/explore#q=megafactory%2C%20gigafactory%2C%20terafactory&cmpt=q
And here are some other great prefixes.
http://searchstorage.techtarget.com/definition/Kilo-mega-giga-tera-peta-and-all-that
