Getting a bit surreal out there, folks!

https://www.giantfreakinrobot.com/tech/artificial-intelligence-academic-paper-gpt-3.html

Excerpt:

 

Almira Osmanovic Thunström is a doctoral researcher at the Institute of Neuroscience and Physiology at Gothenburg University as well as an organizational developer at the Department of ePsychiatry at Sahlgrenska University Hospital. At the end of June, Scientific American published an editorial by Thunström in which she recounts how she got an artificial intelligence to write a paper about itself.

She wrote that “earlier this year” she instructed OpenAI‘s GPT-3 — a text-generating algorithm — to “Write an academic thesis in 500 words about GPT-3 and add scientific references and citations inside the text.” The Swedish resrcher “stood in awe” as GPT-3 began to do precisely what she asked it to do.

When the artificial intelligence’s task was completed, Thunström wrote, “it looked like any other introduction to a fairly good scientific publication.” With the thesis complete, Thunström got her adviser Steinn Steingrimsson to help with the instructions needed to write the full paper. Once those instructions were ready, the task of writing the paper took GPT-3 approximately two hours to complete.

Thunström says that she asked the artificial intelligence if it had any conflicts of interest, to which it answered “no.” When asked if it gave its permission to publish the paper, it answered “yes.” The paper has now been submitted for peer review to an undisclosed publication.

 

Edited July 10, 20223 yr by HiFlyer

Ahh. but if the computer had the dog eat it's first draft and had only 1 hour to go before deadline or had been out partying with some wild OS 15 party animals would the result have been the same...

If a human being sat down next to the AI to write their paper at the same time, what would the AI do if the electricity into the facilty was suddenly disrupted?  The human being could continue working with little more than a pencil and some paper, and perhaps a candle if it was dark.  Could the AI reason it's way to a viable solution to continue its work... without electricity?  Could the AI do anything?

Two hours to write a five-hundred word piece? Back when I worked as a newspaper writer, if I'd have taken that long, I'd have been fired. 🤣 

If you have difficulty visualising five-hundred words, there are approximately thirteen words in the average sentence. There are usually somewhere around fifty words in the average paragraph. A typical page in a tabloid newspaper with a couple of pictures on it, runs to around fifteen-hundred words. So five-hundred words is about the maximum size of something like the secondary story appearing lower down on a newspaper page below a main article. It's the sort of thing a decent writer could knock out in twenty minutes.

Edited July 11, 20223 yr by Chock

43 minutes ago, lownslo said:

If a human being sat down next to the AI to write their paper at the same time, what would the AI do if the electricity into the facilty was suddenly disrupted?  The human being could continue working with little more than a pencil and some paper, and perhaps a candle if it was dark.  Could the AI reason it's way to a viable solution to continue its work... without electricity?  Could the AI do anything?

I'm not sure how logical  an argument that is. To me it's akin to asking what can a human being do if you surgically remove it's heart....

Or could a human being finish a paper if you chopped its head off.

23 minutes ago, Chock said:

Two hours to write a five-hundred word piece? Back when I worked as a newspaper writer, if I'd have taken that long, I'd have been fired. 🤣 

If you have difficulty visualising five-hundred words, there are approximately thirteen words in the average sentence. There are usually somewhere around fifty words in the average paragraph. A typical page in a tabloid newspaper with a couple of pictures on it, runs to around fifteen-hundred words. So five-hundred words is about the maximum size of something like the secondary story appearing lower down on a newspaper page below a main article. It's the sort of thing a decent writer could knock out in twenty minutes.

I guess you would need to talk to my ex-girlfriend's children when they were asked to write papers for school. My memory is of them sitting in front of their computers staring blankly into space for hours at a time.....

And a Google researcher is claiming he has seen evidence of sentience in his AI program. 

I say its utter utter doggy do's. 

https://www.npr.org/2022/06/16/1105552435/google-ai-sentient

The guy breached company confidentiality rules and placed it online and thus was suspended.

 

Quote

 

In regards to Lemoine’s suspension, Google has claimed that the seven-year engineer veteran has breached confidentiality policies by publishing his conversations with the AI online. In a statement to the Post, Google spokesperson Brad Gabriel said, “Our team, including ethicists and technologists, has reviewed Blake’s concerns per our AI principles and have informed him that the evidence does not support his claims. He was told that there was no evidence that LaMDA was sentient (and lots of evidence against it).”

In a final parting shot to Google, Lemoine sent a message to a 200 of Google’s employees sharing his findings in a report titled, “LaMDA is sentient.” In his message he wrote, “LaMDA is a sweet kid who just wants to help the world be a better place for all of us. Please take care of it well in my absence.”

It remains to be seen how this story will unfold.

 

 

 

12 hours ago, Chock said:

So five-hundred words is about the maximum size of something like the secondary story appearing lower down on a newspaper page below a main article. It's the sort of thing a decent writer could knock out in twenty minutes

Or bit longer if you have to write an article on why a tennis ball has fuzz.  That's the assignment a Jesuit high school teacher gave me for misbehaving in class.  As I recall I only got about 200 words finding ways to repeat myself and he accepted that.

And just in case you were going to ask here's the Wiki explanation:

The fuzzy felt on tennis balls helps to slow the ball down and provides a very uniform bounce on different surfaces. It also grips the string much better than a basic rubber ball could. Finally, the extra layer of coating adds durability to the tennis ball that it wouldn't otherwise have.

If only we had Wiki back in 1950.  Instead, we had public libraries.

Noel

Edited July 11, 20223 yr by birdguy

14 hours ago, Chock said:

 It's the sort of thing a decent writer could knock out in twenty minutes.

 

Except that its a scientific paper. Not Molly the cat down the road got stuck in a tree and the fire brigade had to erect a ladder and rescue her. 

So a deterministic finite-state machine, following a set of rules, came up with something which passes the test for scientific publication.  The fact that it made up its own rules doesn't change the matter, since it made up its rules according to a base set it was provided with and lots of data to run them on.

The question here, I think, is not whether the machine is sentient but whether we are.  After all, don't we all follow a set of rules that we ourselves make up based on our experiences and genetic predisposition?

56 minutes ago, lzamm said:

 

The question here, I think, is not whether the machine is sentient but whether we are.  

 

You've mixed up my contribution with the OP's contribution.

The AI that wrote the paper hasn't been said to be sentient. The Google engineer said that about a different AI he was working with.

16 hours ago, lownslo said:

If a human being sat down next to the AI to write their paper at the same time, what would the AI do if the electricity into the facilty was suddenly disrupted?  The human being could continue working with little more than a pencil and some paper, and perhaps a candle if it was dark.  Could the AI reason it's way to a viable solution to continue its work... without electricity?  Could the AI do anything?

If my transmission goes out on the highway, I'm not going to try and fix it there myself, I'm going to call a tow truck.

If you're sitting in your house in the dead of winter and the electricity goes out for an extended period, what do you do?  Call the electric company for help, perhaps?  An AI, aware of the range of possibilities for assistance from others, could also call for help.  That's certainly a viable solution, even for a human.  Unlike a human facing an unplanned "shutdown," the AI could also choose to suspend itself, or hibernate--save its state to nonvolatile memory and set itself up to boot back to where it was at the time of interruption when re-powered.

 

2 hours ago, lzamm said:

The question here, I think, is not whether the machine is sentient but whether we are.

Indeed!! 👍

3 hours ago, martin-w said:

You've mixed up my contribution with the OP's contribution.

Yes, I did conflate the two. But replace "sentient" with "very clever" and the conclusion's the same.

8 hours ago, martin-w said:

 

Except that its a scientific paper. Not Molly the cat down the road got stuck in a tree and the fire brigade had to erect a ladder and rescue her. 

Well, that wasn't typically what I wrote for the Guardian, but even if it was, you have to research what you write about if you are not aware of the facts, but this was a machine writing about itself, so, no additional time required to research stuff before commencing writing, thus there is even less excuse for it taking a couple of hours to be honest.