I’ve been flipp-flopping between being AI-positive, like Joanna Penn, and a complete AI-alarmist, depending on whether I’m thinking about this when the sun’s up or at 3am when I can’t sleep.
On one hand, AI is cool, ChatGPT is fun to play with, and Midjourney is amazing.
On the other, I’m a writer, and reading the news, you’d think ChatGPT is about to do with creative wordsmiths what Microsoft Excel did to accountants in the 1980’s and 90’s: slaughter us wholesale.
Then I read Ted Chiang’s thoughts on the matter.
If you don’t know Ted Chiang, he’s only one of the greatest new SF writers of the past two decade or so. He’s won four Hugos, four Nebulas, six Locusts, and the John W. Campbell. He’s also a futurist and a technical writer working in IT. He knows what he’s talking about.
And Ted’s not worried.
Basically, he sees ChatGPT as a broken JPEG version of the Internet, something that uses an ingenious compression algorithm to create an almost-faithful (but still lossy) copy of the knowledge on the web. And it’s this “still lossy” part that makes it fascinating to us. We humans interpret it as creativity.
It’s not.
ChatGPT has no idea what it outputs. It doesn’t understand our input. It can’t give us better ideas, better stories, better explanations. It can only return a smoothed, washed out version of what we’re already saying.
ChatGPT is a worse version of Google, nothing more.
But Ted Chiang says it so much better: ChatGPT is a Blurry JPEG of the Web in The New Yorker.
