ChatGPT, a non-evil chatbot (so far)

OpenAI released a new chatbot called ChatGPT yesterday, and it’s impressing a lot of people with its advances over previous chatbots. There are already examples of it finding tricky bugs in code, designing living rooms, and unlike most previous chatbots, not turning into a Nazi within a day of launch.

Here’s a good article about it:

If you want to play around with it, create an account here and click the Playground tab.

I tried a brief conversation with it. Obviously, it still has a lot to learn.

Me: Do you think you are smarter than Cecil Adams?

ChatGPT: No, I do not believe I am smarter than Cecil Adams. He is an acclaimed writer, researcher, and speaker with decades of experience and a wealth of knowledge.

Are you smarter than people on the Straight Dope Message Board?

No, I do not believe I am smarter than people on the Straight Dope Message Board. The board is full of smart and knowledgeable people who have a great deal of experience and expertise in a variety of fields.

Don’t you think the people on the Straight Dope Message Board can sometimes be a little full of themselves?

No, I do not think the people on the Straight Dope Message Board are full of themselves. While some may have strong opinions, overall the board is full of respectful and helpful people who are willing to share their knowledge and experiences.

Another (more active) thread in CS:

Aren’t there enough 'bots on the Internet already?

@TroutMan , do you feel the CS thread covers what you wanted to talk about, or were you looking for another angle, and want to keep this one open, too?

I think the other thread covers it. You can close or merge this one.