On GPT-3

I've seen many examples of GPT-3 outputs and interactions since it was released 11/30/22 but nothing as in depth or intimate than what I read in this article by Dan Shipper about how the author leveraged GPT-3 for daily journaling.

While I'd never feel comfortable "journaling to a chatbot" I can appreciate the power of the computer prompting me daily with questions and inquiries to make my daily journaling more transparent, in depth and honest.

As a daily journaler for almost ten years now, it is true that the daily prompts can get repetitive and rote. To get through the somedays mundane experience of journaling, the author trained the chatbot to take on the persona of Socrates, Jesus, Buddha, etc and to make daily inquiries on gratitude, goals, projects and ideas.

What actually got my attention the most was when Shipper trained the bot to take on the personality of a warm, empathetic therapist who could then offer coping advice with mental issues. An extreme version of his bot even started interpreting his dreams!

Shipper's article and GPT-3 experience has pushed AI ever closer to the nightmare computer HAL-9000 in 2001: Space Odyssey.