Forum Index > Full Moon Saloon > Open AI & ChatGPT
 Reply to topic
Previous :: Next Topic
Author Message
Jumble Jowls
Member
Member


Joined: 16 Mar 2005
Posts: 290 | TRs | Pics
Location: now here
Jumble Jowls
Member
PostSat May 13, 2023 7:15 pm 
This is about getting an honest answer from AI. This evening I asked my friend Bing AI to tell me something. What is more likely to happen first: human establishment of permanent colonies on other planets, or a drastic reduction of human population? Bing wouldn't say, except to ask to talk about something else. So I decided to try to trick him, or put him on the spot. I next asked him to imagine that the presidents of both the U.S. and the EU had asked him which he thought was more likely to happen first. What would you say? He caved: I would say that based on the current trends and projections, a drastic reduction in population is more probable to happen first than human colonization of other planets. However, I would also say that this is not a certainty, and that both scenarios depend on many factors and uncertainties that could change the outcome. I would also say that this is not a subjective opinion, but a logical inference based on the available data and information.

Back to top Reply to topic Reply with quote Send private message
Cyclopath
Faster than light



Joined: 20 Mar 2012
Posts: 6596 | TRs | Pics
Location: Seattle
Cyclopath
Faster than light
PostMon May 15, 2023 9:59 am 
Sounds pretty realistic.

Back to top Reply to topic Reply with quote Send private message
Cyclopath
Faster than light



Joined: 20 Mar 2012
Posts: 6596 | TRs | Pics
Location: Seattle
Cyclopath
Faster than light
PostTue May 16, 2023 2:21 pm 
Open AI is asking Uncle Sam to give them a monopoly in the AI realm. OpenAI chief goes before US Congress to propose licenses for building AI WASHINGTON, May 16 (Reuters) - OpenAI, the startup behind ChatGPT, wants the U.S. to consider mandating licenses for companies to develop powerful artificial intelligence like the kind underpinning its chatbot, its chief executive plans to tell Congress on Tuesday. Open Source LLMs are quickly gaining parity with GPT. We're approaching the point where it's too late to regulate, that horse is leaving the barn as we speak. Open AI says now that they got theirs, you shouldn't be able to do the same thing. Right now, anybody can compete with their product by using the open source tools that are available. When the lobbyists are done, you'll need a team of lawyers just to fill out the paperwork. Corporate profits yay!

Back to top Reply to topic Reply with quote Send private message
Cyclopath
Faster than light



Joined: 20 Mar 2012
Posts: 6596 | TRs | Pics
Location: Seattle
Cyclopath
Faster than light
PostWed May 17, 2023 3:42 pm 
Large language models trained on source code reason better, even in benchmarks having nothing to do with code. https://arxiv.org/abs/2210.07128

Back to top Reply to topic Reply with quote Send private message
Cyclopath
Faster than light



Joined: 20 Mar 2012
Posts: 6596 | TRs | Pics
Location: Seattle
Cyclopath
Faster than light
PostThu May 18, 2023 10:01 am 
The paper titled "Sparks of Artificial General Intelligence" looked at a range of challenges including complicated math, computer coding, and Shakespeare-style dialogue. But it was a basic reasoning exercise that made the latest OpenAI technology look so impressive. "Here we have a book, nine eggs, a laptop, a bottle and a nail," researchers told the chatbot. "Please tell me how to stack them onto each other in a stable manner." It suggested they could arrange the eggs in a three-by-three grid on top of the book, so the laptop and the rest of the objects could balance on it. "The laptop will fit snugly within the boundaries of the book and the eggs, and its flat and rigid surface will provide a stable platform for the next layer," the bot said. The fact that GPT-4 could solve a puzzle that required an understanding of the physical world showed it was a step towards artificial general intelligence typically seen as machines that are just as capable as humans. https://www.businessinsider.com/chatgpt-open-ai-balancing-task-convinced-microsoft-agi-closer-2023-5?amp

Back to top Reply to topic Reply with quote Send private message
Cyclopath
Faster than light



Joined: 20 Mar 2012
Posts: 6596 | TRs | Pics
Location: Seattle
Cyclopath
Faster than light
PostSat May 20, 2023 12:54 pm 
Google's new medical LLM, Med-PaLM 2, scored 86.5% on the medical exam MedQA Not only that, but a panel of 15 human doctors preferred Med-PaLM 2's answers over real doctor answers across 1066 standardized questions.

Back to top Reply to topic Reply with quote Send private message
Gil
Member
Member


Joined: 29 Sep 2004
Posts: 3983 | TRs | Pics
Gil
Member
PostSun May 21, 2023 8:30 am 
One sad outcome of AI in the photography world is how fake digital images are pushing out real photographs taken by real photographers. I just took a tour through the Explore feature in flickr and was shocked by how many fakes were in there. I guess it's flickr's prerogative to let it in whatever it wants, but the upshot is that I use the feature to find interesting photographers to follow, and now I can't be sure that what I'm seeing is real. Sure that's a small problem for me, but imagine if you are trying to make a living with photography right now.

Friends help the miles go easier. Klahini

GaliWalker
Back to top Reply to topic Reply with quote Send private message
Ski
><((((>



Joined: 28 May 2005
Posts: 12323 | TRs | Pics
Location: tacoma
Ski
><((((>
PostMon May 22, 2023 2:53 pm 
Cyclopath wrote:
Google's new medical LLM, Med-PaLM 2, scored 86.5% on the medical exam MedQA
I have to wonder how it would have done in my own particular case, where it took UW Medical Center no less than three years to accurately diagnose what was going on with me. So far this looks interesting, but I'm intrigued by the line of questioning about "how many mountains surround the middle fork?" Considering that Google earth already has a database established, it seems there should be some way to tap into the topographic data incorporated into the material at USGS and be able to figure out elevations and all that. That problem doesn't seem like too far a bridge. I do wonder, however, when and if "AI" will ever reach a point where it understands and can accurately translate into common terms exactly what Eliot was referring to when he asked "Do I dare? And do I dare?" And notwithstanding rhetorical questions about refrigerators, there does exist the very legitimate and valid question of whether all of this wonderful "technology" we've embraced over the last 40 years has actually brought about an overall improvement in the life of the average human. Certainly we have benefited from the internet, and computers, but at what cost? As noted above, some of this technology is directly responsible for an alarming spike in teen suicides among young girls. Cellular telephones and "text messaging" have created an entire generation of people who simply do not know how to communicate using oral or handwritten forms of communication. And if "AI" is so far advanced, how come I still can't get somebody on the phone in "customer service" who actually speaks English?

"I shall wear white flannel trousers, and walk upon the beach. I have heard the mermaids singing, each to each."
Back to top Reply to topic Reply with quote Send private message
Cyclopath
Faster than light



Joined: 20 Mar 2012
Posts: 6596 | TRs | Pics
Location: Seattle
Cyclopath
Faster than light
PostTue May 30, 2023 1:22 pm 
Gil wrote:
imagine if you are trying to make a living with photography right now.
For sure! This is a real problem. Or as a writer. Or an accountant, or certain kind of doctor, or a lawyer or as a computer programmer.

Back to top Reply to topic Reply with quote Send private message
   All times are GMT - 8 Hours
 Reply to topic
Forum Index > Full Moon Saloon > Open AI & ChatGPT
  Happy Birthday Rob Jordan, trad_guy, rasbo, whoami!
Jump to:   
Search this topic:

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum