Wordy Wednesday: You Look Like a Thing and I Love You by Janelle Shane
I found Janelle Shane on Instagram (@janelle.shane) via an artist who used Shane’s neural-net generated lists of drawing prompts. The #botober lists served as an alternative project for the #inktober drawing challenge, and included separate lists of things, concepts, and even a list Shane entitled “Terrible,” to prompt a drawing each day in October. Here are a few examples:
Something that is both a bunny and a shark
Pretzels as a being
Tiny chocolate pants
Exploring Shane’s Instagram further, I found absurd and hilarious bot-generated hilarity, such as candy heart sayings (“okay bro,” “doomed,” “fart booby”) and candle scent names (“volcanoes comfort,” “freshly scrutinized orange”).
You Look Like a Thing and I Love You is Shane’s book about artificial intelligence (AI). I read this book because I wanted to keep laughing, and I was able to learn about AI while having a lot of fun.
Of the AI-generated content in You Look Like a Thing, I found most memorable the knock-knock jokes, recipes. and perfume names. But what I found most elucidating was the chapter in which Shane breaks down neuron by neuron, with helpful illustrations, how AI actually works.
I especially enjoyed learning about how AI has infiltrated the world of art, in which humans and bots collaborate: humans train the bots, then curate the content that the bots generate.
Shane defines computer science terms throughout the book, which is why I’m highlighting it on Wordy Wednesday. The usual format for Wordy Wednesday is to share the word, it’s Oxford definition, the sentence from the book in which I found it, and a sentence of my own using the word. For today’s word, the sentences of my own summarize a point Shane makes in her description of it. Enjoy!
Noun: a test for intelligence in a computer, requiring that a human being should be unable to distinguish the machine from another human being by using the replies to questions put to both.
“A computer program passes the standard Turing test if it can chat with humans and convince approximately one-third of them that it’s a human being rather than a computer” (36).
Shane points out that passing the Turing test, unlike as depicted in movies, is not a sign of self-awareness. Plenty of bots used in customer service (think live chat online assistance) pass the Turing test, only getting tripped up for complicated issues or if a person veers way off topic.
You can check out more of Janelle Shane’s writing on her blog, aiweirdness.com.