René’s prompt for OpenAI’s GPT-3:

Write an email to tell Mark that I need him to wake up and work

alt text

If you are like me, reading GPT-3 generated text makes your eyes glaze over. But when the text is about me, it seizes my attention. I simply observe this as a fact, more with curiosity rather than shame.

I’m not so bullish on AI but I haven’t thought about it deeply. I once read from someone I respect (but don’t remember the author) that AI was modern alchemy, because it made very smart people detour into an intractable problem which could not be resolved, and I thought that in all my dullness my mind would be sucked into a never ending rabbit hole, and so I decided AI was too sophisticated for me.

But today I took a look at OpenAI’s API. It’s amazingly simple. It is wonderful that I can stand on the shoulders of giants and implement sophisticated behavior with very little effort, so why not make a prototype about it? I’ve always been skeptical about AI completing my text but you must experience things first hand.

Wake up and work!