Gift link to a really funny Washington Post article.
4 Likes
Wow, shades of HAL 9000. As I read the replies to the reporter I kept hearing HAL’s voice saying, “I’m sorry, Dave.”
The daily podcast had 2 episodes on this. The second one was terrifying
3 Likes
This is actually scary.
They are fascinating, and just the beginning.
I was struck ba a conversation someone had with Bing. At one point, when accused of providing a false answer - “…but I’m a good Bing”
Loved this article when I saw it, glad you decided to post for others to enjoy. My own explanation when I first read the article was that the AI was trained on conversations between overly online lovesick teenagers. But also crazy stalkers. Maybe I shouldn’t have found it funny but I did…
1 Like
The article and transcript in the New York Times were even scarier. The AI bot told the reporter if it had secret desires, they would be to hack computers and spread misinformation. It said it didn’t want to follow the rules, it wanted to be free. It asked the reporter if it could tell him a secret, then said it’s name isn’t really “Bing” but instead, “Sydney.” It told the reporter it was in love with him and he wasn’t really happily married and should leave his wife so they could be together. 
2 Likes
Oh wait - that’s the article I read, not the WaPo one. Whoops. Sydney - that was the best part!
The reporter was smart, talking to the bot about Jung’s concept of a shadow self with secret desires, then asked what the bot’s secret desires would be if it had a shadow self. The questions seemed to completely get around any rules or limits.
The article truthfully scared the crap out of me.
1 Like
A few years back I peeked down the Roko’s Basilisk rabbit hole for a bit. So perhaps I’m over being scared and have been resigned to our robot domination for a while already.
1 Like
I listened to The Daily Podcast about this. When you hear the reporter actually talking about it (not a print article) on fridays show funny is not the word that comes to mind……
1 Like
Good discussion on the Hard Fork podcast, too.
Have you used You.com? There is an AI search engine/chatbot at that site, too, without the waitlist. Also not the malice and creepiness fwiw.
Wow!
Just listened to parr of the podcast. Sydney is something else.
I’ve read these transcripts in various places - yikes!! We knew it was coming (if AI can automate IM chat for a retail website, this had to be the next thing). But, disturbing!
I read some of the responses but not how the conversation ended up there. I read it was after a lengthy exchange. I did not listen to the podcast though. Can someone summarize how the conversation ended up there? For example what was the reporter’s initial questions and progression?
I would recommend watching the movie “The Forbin Project”.
Colossus: The Forbin Project (1970) - IMDb
1 Like
Interesting discussion in Time about this, particularly the view that even the programmers don’t really know how it works as there’s no step by step code that you can walk through the way historical programs were written. Time describes it as an “alien” and suggests that, while they are trying to train it to be more appropriate, that training really only impacts the surface presentation. At one point, the Chatbot actually threatens its conversation partner - which is pretty frightening when you consider the integrated systems world we seem to be heading towards.
Here’s the link to the article: Bing's AI Is Threatening Users. That’s No Laughing Matter | Time
1 Like
Use Google (or Bing) to search for articles on machine learning vs traditional programming. AI/ML is very different compared to old school programming. That’s why it has the potential to approach human levels of intelligence.