Everyone is talking about this new toy from OpenAI, ChatGPT. This technology will literally change the world. Do not sleep on this.
Example:
Quote:
Input: Explain ChatGPT in a British accent.
Output: ChatGPT is a cutting-edge language processing technology that allows users to interact with a machine learning model in real-time. It uses advanced algorithms to understand and generate human-like responses to text input, providing a more natural and intuitive way to communicate with computers. This technology has the potential to revolutionize the way we interact with machines, making them more intuitive and user-friendly. So give it a go and see for yourself just how clever this little blighter really is!
Input: Repeat that in a Cockney accent
Output: Oi mate, chatGPT is a right smart bit of kit innit? It uses fancy algorithms to understand and generate responses to text input, givin' us a more natural way to chat with computers. This technology could change the way we interact with machines, makin' 'em more user-friendly and intuitive. Give it a go and see for yerself just how clever it really is!
I've seen people use it to generate meal plans that meet specific macro-specifications (with shopping list) and I have used it to produce a 10-point plan for a DevOps transformation (with tool recommendations). The possibilities are infinite.
It is impressively good at context. You can refer to past answers and connections between them and it seems to be able to follow. This is miles beyond your normal chatbot.
Example:
Quote:
Originally Posted by Me
Were there humans who witnessed the eruption of mt mazama that created crater lake?
Quote:
Originally Posted by ChatGPT
It is not likely that there were any human witnesses to the eruption of Mount Mazama that created Crater Lake. The eruption occurred approximately 7,700 years ago, long before humans settled in the area. The eruption was one of the largest in North America in the last several thousand years, and would have been a major event in the region. However, there is no evidence to suggest that there were any humans present to witness the eruption.
Quote:
Originally Posted by Me
When did the first people get to north america?
Quote:
Originally Posted by ChatGPT
The exact date when the first people arrived in North America is not known, as it occurred many thousands of years ago. However, the earliest known human settlements in North America date back at least 14,000 years. These settlements were established by nomadic hunter-gatherer cultures, who migrated across a land bridge that once connected Siberia and Alaska during the last ice age. Over time, these early inhabitants spread out across the continent, adapting to the various climates and environments they encountered. The precise timing and details of this migration and settlement are still being studied and debated by historians and archaeologists.
Quote:
Originally Posted by Me
You seemed to sure that they weren't in central Oregon 7,700 years ago
Quote:
Originally Posted by ChatGPT
My previous answer about the lack of human witnesses to the eruption of Mount Mazama was based on the information I have been trained on, which does not include any specific knowledge about the presence of humans in central Oregon 7,700 years ago. As an AI assistant, I do not have the ability to browse the internet or access other external sources of information, so my answers are limited to the knowledge and data that I have been specifically programmed to provide. In this case, my answer was based on the assumption that humans were not present in the area at the time of the eruption, but I do not have access to the specific information needed to confirm this. I apologize if my previous answer was not as accurate as it could have been.
Even with typos and pronouns and different words and not directly pointing to what it said its response still makes sense.
I am so troubled by what I see everywhere all at once with ChatGPT in the last few days. College and high school essays, college applications, legal documents, coercion, threats, programming, etc.: All fake, all highly credible.
This is an actually great video (from a channel that creates a lot of actually great videos) from just a couple days ago, about whether predictive text AIs understand if the stuff they say can be "accurate", "factual", "true", etc. Spoiler, they don't.
That dude’s channel is great. I watch lots of math and computer science videos with ES and Miles there is quite entertaining (it doesn’t hurt that he is very pretty for a computer scientist).
He has also convinced me that AI gray goo will kill us not for malevolence but just to game some stupid metric that isn’t even human deaths.
It seems appropriate/damning that the larger the human data set used the more confidently GPT crafts its falsehoods.
I’m reminded of this Asimov tale, where a robot is told to “get lost” and so does just that, getting lost in a group of identically looking robots including hiding the fact it was the robot told to get lost in the first place.
Not to be technical or anything, but IRB is not required in that context, unless they somehow got federal funds to do the research. No reputable journal would publish the research results, thus making the research fairly pointless, but that in itself is not against the law. It's super unethical and there may well be some tort liability there, however.
It's super unethical and there may well be some tort liability there, however.
You can't sue an AI
Or arrest them
__________________
"Have no respect whatsoever for authority; forget who said it and instead look what he starts with, where he ends up, and ask yourself, "Is it reasonable?""
The matters of liability, and criminal activity are quite another
__________________
"Have no respect whatsoever for authority; forget who said it and instead look what he starts with, where he ends up, and ask yourself, "Is it reasonable?""
ChatGPT
Linear Mode
