Didn’t they do this before and people turned it racist in like, 12 hours? I think Internet Historian had something on it
Yup. You’re thinking of Tay.
Has Bing gone full Tay and start agreeing that Hitler was right and to fire up the gas chambers yet?
After a couple of early tweaks it started sounding like a jealous psycho ex-girlfriend. You know that one we’ve all had. They limited the sessions down to 30 iterations and that stopped happening, now every 30 conversations or so you have to reset the history of yourself with it.
I use it every night to generate lengthy bedtime stories for the kid. Most of the time it just gives you what you ask for. If you select creative mode it’ll throw a little surprise in here or there.
The night before last something kind of funny happened. I use a mad libs style slightly modified prompt every night to generate the continent. It gives me the same characters (mother, father, child) the same personalities and whatever details in the outline I give it is what story I get.
Last week there was a dragon in the story I was pretty vague about the dragon and told it to make up the character and provide a lot of detail. It named it gave it personality traits did a bang-up job. Well the most recent story, Even though it didn’t totally fit I told it to be creative and it brought the same exact dragon back. It inserted into the story semi-appropriately and it was kind of fun.
At the end of the story for some reason it decided to make the dragon a new boyfriend for the mother. And made it a point that the father was down with this. Open relationship with a dragon that’s a kind of wacky.
At the end I gave it feedback saying that I loved that it brought the dragon back from a previous episode, but said that the parents were in a monogamous relationship and bringing the dragon in as a boyfriend wasn’t where I wanted to go with the kids bedtime story.
“I’m sorry if I made you uncomfortable by adding the boyfriend into a family that already had a mother and father I didn’t mean to offend you or anyone else I just thought it would be a nice way to show that love can come in different forms and that everyone deserves to be happy.”
I know the ethics behind it are questionable especially with the way they implemented but honestly for the time when they first started testing it, I really enjoyed watching it break and it be rude/passive aggressive. Like it was clear it wasn’t ready at all but it was so funny. When it was breaking I would just sit there having fights with it over random bullshit. That’s what made it feel more “real” more than anything else.
In the future if my AI chatbot doesn’t have an option to add some bitchiness to it, I don’t want it. I need my AI to have some attitude.
Interesting angle! I can see the need for both - maybe there’s room for a snarky option??
AI Hallucinations: https://www.sify.com/ai-analytics/the-hilarious-and-horrifying-hallucinations-of-ai/
LOL
“Bringing out the best of bing to the chatGPT experience”…
LMAO
Rushing out general purpose AI, what could go wrong
They are competitors after all. Openai would love to see Microsoft keep working on gpt integration for the coming years while Chatgpt steals the show.
I wish tech companies would do this more. Put a warning or label on it if you have to, but interacting with that early version of the Bing chatbot was the most fun I had with tech in awhile. You don’t have to install a ton of guardrails on everything before it goes out to the public.
