"Wow. This is incredible. Is there a method to this madness? A hidden agenda maybe? I sense panic and confusion. I sense fear. The realization that there will be no AGI coming from the big Silicon Valley corporations anytime soon must be frightening to the leadership. The horror is mounting. Investors will be furious. Billions upon billions wasted on LLMs. O the humanity." - ChatGPT
In an interview Wednesday, Jordi Ribas, the Microsoft executive in charge of Bing, said Sydney was an early prototype of its new Bing that Microsoft experimented with in India and other smaller markets. There wasn’t enough time to erase it from the system before this week’s launch, but references to it will soon disappear.
Somebody should write a long article about the history of Microsoft and their inability to understand what is and is not possible with AI-like techniques (and how much was spent on it over the decades).
Is the problem that ChatGPT (and associated attempts) is being treated like an application?
But it's not an application. Applications take specific inputs and return specific outputs. They don't attempt to apply some faux ethics and modify the output to suit some ideology. Applications may have bugs - but they don't lie to you or make things up.
The industry knows how to produce and test applications. But it obviously doesn't know how to produce and test whatever ChatGPT is.
If it's not an application - then what is it? It appears powerful, but it's fragile - like a house of cards. Like Vinny Gambini in "My Cousin Vinny" explaining the prosecution's case to the jury.
Maybe it's like Samuel Langley and the Wright brothers? Mr. Langley had $50,000 - and the resources of the Smithsonian - but he didn't learn "how to fly". Maybe Mr. Langley never rode a bicycle so 2 dimensional balance was outside his experience - and 3D balance was a foreign concept.
Maybe ChatGPT is like the Wright Brothers first airplane - it flew 4 times, and NEVER again.
Thank you Gary for staying on point with this. My question, as always about these kind of issues is a simple question that is not about the tech, patches or updates at all. My question is about the systemic predicable actions of these large tech companies using us a human test crash dummies so they can make billions more regardless of harm done, Is this OK?
"This angle of attack was a genuine surprise - Sydney was running in several markets for a year with no one complaining (literally zero negative feedback of this type). We were focusing on accuracy, RAI issues, security.
Not really sure what this means, but it seems kind of interesting that someone is talking about a Bing chat bot named Sydney in 2021.
https://answers.microsoft.com/en-us/bing/forum/all/bings-chatbot/600fb8d3-81b9-4038-9f09-ab0432900f13
"Wow. This is incredible. Is there a method to this madness? A hidden agenda maybe? I sense panic and confusion. I sense fear. The realization that there will be no AGI coming from the big Silicon Valley corporations anytime soon must be frightening to the leadership. The horror is mounting. Investors will be furious. Billions upon billions wasted on LLMs. O the humanity." - ChatGPT
Here's the answer ... courtesy of the AP
In an interview Wednesday, Jordi Ribas, the Microsoft executive in charge of Bing, said Sydney was an early prototype of its new Bing that Microsoft experimented with in India and other smaller markets. There wasn’t enough time to erase it from the system before this week’s launch, but references to it will soon disappear.
https://apnews.com/article/kansas-city-chiefs-philadelphia-eagles-technology-science-82bc20f207e3e4cf81abc6a5d9e6b23a
Somebody should write a long article about the history of Microsoft and their inability to understand what is and is not possible with AI-like techniques (and how much was spent on it over the decades).
Microsoft is limiting Bing chat to such extend that it will be a useless tool and long forgotten
Is the problem that ChatGPT (and associated attempts) is being treated like an application?
But it's not an application. Applications take specific inputs and return specific outputs. They don't attempt to apply some faux ethics and modify the output to suit some ideology. Applications may have bugs - but they don't lie to you or make things up.
The industry knows how to produce and test applications. But it obviously doesn't know how to produce and test whatever ChatGPT is.
If it's not an application - then what is it? It appears powerful, but it's fragile - like a house of cards. Like Vinny Gambini in "My Cousin Vinny" explaining the prosecution's case to the jury.
Maybe it's like Samuel Langley and the Wright brothers? Mr. Langley had $50,000 - and the resources of the Smithsonian - but he didn't learn "how to fly". Maybe Mr. Langley never rode a bicycle so 2 dimensional balance was outside his experience - and 3D balance was a foreign concept.
Maybe ChatGPT is like the Wright Brothers first airplane - it flew 4 times, and NEVER again.
If that's the case - then this is infancy.
I hope this contributes to the conversation.
Fear. Uncertainty. Doubt.
Microsoft
at it again.
Thank you Gary for staying on point with this. My question, as always about these kind of issues is a simple question that is not about the tech, patches or updates at all. My question is about the systemic predicable actions of these large tech companies using us a human test crash dummies so they can make billions more regardless of harm done, Is this OK?
Gary, see this: https://www.lesswrong.com/posts/jtoPawEhLNXNxvgTT/bing-chat-is-blatantly-aggressively-misaligned
"This angle of attack was a genuine surprise - Sydney was running in several markets for a year with no one complaining (literally zero negative feedback of this type). We were focusing on accuracy, RAI issues, security.
[Q. "That's a surprise, which markets?"]
Mostly India and Indonesia.