The OpenAI organization has now made the GPT-3 language model open to developers. I’ve signed up and started to experiment with it.
I listened to a demo of the remarkable potential of this technology on the “Clearer Thinking” podcast and here’s a short excerpt from the 2-hour podcast; it’s really fascinating to notice both its flaws and its flexibility:
This episode from the “A.I. Nation” podcast is entertaining. Below is an excerpt.
An AI researcher with access to GPT3 (the language understanding neural network) feeds in the words of a character in Jurassic Park and GPT3 improvises what the character might have said next. It’s uncanny. (6.5 minutes)
It doesn’t matter to me whether AI “understands” per se. Airplanes don’t flap their wings but we still call it flight. AI works differently than our brains and has different strengths and weaknesses. But it’s getting better and better at extending our minds into a new medium, through a new tool. Question is: what can we do with it?