My post from yesterday about why I'm not anxious to use AI as a content creator stirred up some interest in the comment section.
I thought I should do a follow-up about something that happened during one of the conversations I had late last night (for me) on my post.
@Steemmillionaire - whom I know for a long time from the previous chain as his username suggests - made an experiment in the comment section, and I was practically involved in it without knowing.
Here's his comment and my reply:
Now, a little context.
He does comment from time to time on my posts, usually leaving useful information, sharing his opinion, or even asking a question if he wants to know more about something I didn't speak of enough in the post.
This comment didn't represent him. It felt really awkward reading it, and I didn't understand why he showed up and dumb the expression down. And for whom?
I didn't show it in my response, but I was a bit pissed off too, lol.
Here's how the conversation continued and ended:
Yep, although his comment was strange and I felt it was something out of place, it didn't cross my mind at that late hour that he used the chat GPT to generate it.
As I say in my reply, the fact the chat GPT understands (simple) metaphors and can explain them is not a small thing at all. That makes it very powerful already!
What I believe happened was the AI saw in the input text both the word 'teacher', the expression, and 'scientific fact' and decided to take a didactic approach.
On the other side, we are not in any danger of talking to an AI and not noticing its clumsiness in picking the right (unawkward) answers which a real person will do naturally, in most cases.
I've heard that from a number of people who tried chatting with it for more than a bit.
OpenAI still has a collection of very impressive AI tools. Including one AI that writes Python code starting from natural language. Coders beware! AI is coming for your jobs too.