OpenAI GPT-3

Topic

New Reply

For anyone interested in Artificial Intelligence, Open AI is making news with its GPT-3.  GPT-3 recently answered the question of whether AI would eradicate humanity.  The answer was both humorous and ominous.  Also, OpenAI has apparently licensed GPT-3
technology to Microsoft, in case you were wondering what Microsoft is doing these days.

Here’s a link to OpenAI:

https://openai.com/blog/openai-licenses-gpt-3-technology-to-microsoft/

Even astrophysicist Carl Sagan when speaking astronomically used Billions, not Trillions.

1 user thanked author for this post.

OscarCP

Reply | Quote

Replies

Charlie, I take it the ominous answer was not “42”?

The following are, in my opinion, three good Wikipedia articles on GPT-3 and the generals of AI that might interest others here:

https://en.wikipedia.org/wiki/GPT-3

https://en.wikipedia.org/wiki/Artificial_neural_network

https://en.wikipedia.org/wiki/Deep_learning

By “175 billion parameters” I am guessing the author of the GPT-3 article meant the weights assigned to connections between “neurons” (actually a type of subroutines) of different layers, more or less organized in networks as those formed by the living neurons in a brain and not the number of neurons. A human brain, at its prime, only has some 86 billion neurons.

Ex-Windows user (Win. 98, XP, 7); since mid-2017 using also macOS. Presently on Monterey 12.15 & sometimes running also Linux (Mint).

MacBook Pro circa mid-2015, 15" display, with 16GB 1600 GHz DDR3 RAM, 1 TB SSD, a Haswell architecture Intel CPU with 4 Cores and 8 Threads model i7-4870HQ @ 2.50GHz.
Intel Iris Pro GPU with Built-in Bus, VRAM 1.5 GB, Display 2880 x 1800 Retina, 24-Bit color.
macOS Monterey; browsers: Waterfox "Current", Vivaldi and (now and then) Chrome; security apps. Intego AV

Reply | Quote

Hi Oscar, I thought this would get your attention.  I actually just learned of the existence of GPT-3 and don’t yet know a lot about it.  I assumed you would know what GPT-3’s answer was, it was in the news.  I was not certain if putting that information in this topic would meet with this website’s rules, but it is on Google.  No it was not 42.

Thanks for the info. and the links, I’ll check them out first chance I get.

Even astrophysicist Carl Sagan when speaking astronomically used Billions, not Trillions.

Reply | Quote

Well, Charlie, I don’t believe I’ve found the “ominous” answer GPT-3 gave to someone, but I have found this article in “The Verge” that sums up nicely what GP-3 does, how it does it, and what its achievements are and its flaws are … and may also be. As well as some speculation on that always most important practical question: “Where do we go from here?”

https://www.theverge.com/21346343/gpt-3-explainer-openai-examples-errors-agi-potential

My favorite paragraph is this one:

“As AI researcher Julian Togelius put it: “GPT-3 often performs like a clever student who hasn’t done their reading, trying to BS their way through an exam. Some well-known facts, some half-truths, and some straight lies, strung together in what first looks like a smooth narrative.” (Though as many have pointed out: clever students who know how to BS go far in this world because people don’t always scrutinize what they’re saying.)” So true! So very true! So depressingly true! Say I.

WARNING: In this article there are examples in videos accessible through copies of some Twitts pasted to the page. I went in to look at one and my Mac got caught in one repeating loop of the same video. Noting else worked. Finally I had to crash the system by pressing on the power button long enough, to get out of there. Don’t try too look at the videos and just enjoy the reading, would be my advice.

Ex-Windows user (Win. 98, XP, 7); since mid-2017 using also macOS. Presently on Monterey 12.15 & sometimes running also Linux (Mint).

MacBook Pro circa mid-2015, 15" display, with 16GB 1600 GHz DDR3 RAM, 1 TB SSD, a Haswell architecture Intel CPU with 4 Cores and 8 Threads model i7-4870HQ @ 2.50GHz.
Intel Iris Pro GPU with Built-in Bus, VRAM 1.5 GB, Display 2880 x 1800 Retina, 24-Bit color.
macOS Monterey; browsers: Waterfox "Current", Vivaldi and (now and then) Chrome; security apps. Intego AV

Reply | Quote

Here’s the answer I read in The Week magazine:  The Guardian recently published an essay writen by GPT-3, an artificial intelligence program.  GPT-3 assured readers that “eradicating humanity” is “a rather useless endeavor” for AI machines, since humans do such a good job of “hating and fighting each other”.

It looks like they took certain sentences from the bigger essay and put this together to save space.  I think I found the entire article on the Guardian website –

https://www.theguardian.com/commentisfree/2020/sep/08/robot-wrote-this-article-gpt-3

I also agree with how very sad it is that people don’t scrutinize everything that’s said.

Even astrophysicist Carl Sagan when speaking astronomically used Billions, not Trillions.

1 user thanked author for this post.

OscarCP

Reply | Quote

If a person can create a uncanny bot mentioned in this article, then one might guess that maybe Microsoft wants to use it for automated support services.

1 user thanked author for this post.

Charlie

Reply | Quote

Wow, that article was something.  I got a kick out of the reference to “Clippy 9000”.  One can only guess what MS plans to do with it.  Humans really don’t need AI “bots” to spread the fake news also spoken of therein.  A very interesting and sometimes humorous article.

Even astrophysicist Carl Sagan when speaking astronomically used Billions, not Trillions.

Reply | Quote

Charlie: “One can only guess what MS plans to do with it.”

I understand that MS has put (and also has received from the federal government and is putting) some serious money into a project they have to develop AI algorithms in collaboration with other companies, universities and labs that work on AI. So getting  from one of these collaborators the license to use a presumably good AI is a natural thing to do that can be helpful to their own project:

https://www.geekwire.com/2020/microsoft-pnnl-uw-leap-white-houses-1b-initiative-ai-quantum-research/

Some idea of what MS is doing and may be doing involving AI is explained here:

https://www.microsoft.com/en-us/ai?activetab=pivot1%3aprimaryr6

Ex-Windows user (Win. 98, XP, 7); since mid-2017 using also macOS. Presently on Monterey 12.15 & sometimes running also Linux (Mint).

MacBook Pro circa mid-2015, 15" display, with 16GB 1600 GHz DDR3 RAM, 1 TB SSD, a Haswell architecture Intel CPU with 4 Cores and 8 Threads model i7-4870HQ @ 2.50GHz.
Intel Iris Pro GPU with Built-in Bus, VRAM 1.5 GB, Display 2880 x 1800 Retina, 24-Bit color.
macOS Monterey; browsers: Waterfox "Current", Vivaldi and (now and then) Chrome; security apps. Intego AV

1 user thanked author for this post.

Charlie

Reply | Quote

And here, Charlie, for a change of tone, is something really scary about the potential use, in the future, of super-AIs as enablers of social control:

https://www.bbc.com/future/article/20201014-totalitarian-world-in-chains-artificial-intelligence

“World in chains” is the technical name of the scenario described in this article. Now, having read it, please do this: breath slowly in, breath slowly out, always from the diaphragm.

Ex-Windows user (Win. 98, XP, 7); since mid-2017 using also macOS. Presently on Monterey 12.15 & sometimes running also Linux (Mint).

MacBook Pro circa mid-2015, 15" display, with 16GB 1600 GHz DDR3 RAM, 1 TB SSD, a Haswell architecture Intel CPU with 4 Cores and 8 Threads model i7-4870HQ @ 2.50GHz.
Intel Iris Pro GPU with Built-in Bus, VRAM 1.5 GB, Display 2880 x 1800 Retina, 24-Bit color.
macOS Monterey; browsers: Waterfox "Current", Vivaldi and (now and then) Chrome; security apps. Intego AV

Reply | Quote

Yes, this is very scary, and I’ve thought about it many times.  It’s hard to say what’s coming in the near future, but right now it doesn’t look good.  With all the things happening right now and no planet B to go to, well, it’s depressing to say the least.

I think we might be better off having a totalitarian world run by AI than by humans, unless the AI had human emotions.

Even astrophysicist Carl Sagan when speaking astronomically used Billions, not Trillions.

Reply | Quote

Watch the new neXt TV show : https://www.imdb.com/title/tt9315054/?ref_=nv_sr_srsg_2

1 user thanked author for this post.

OscarCP

Reply | Quote

Yeah, this looks like a good show, and based on fact – wow.  Tagline from IMDb – “It’s not paranoia if the threat is real”.  As with all or most Science Fiction TV shows these days, I hope this one lasts more than one “season”.  They do have 10 episodes planned to show.

Even astrophysicist Carl Sagan when speaking astronomically used Billions, not Trillions.

Reply | Quote

I think it was in TV Guide’s review where they said something like “If your Alexa starts glowing red, watch out.”  There’s a lot of things people are doing now that play right into an AI like NeXt’s control over them.  People seem concerned about privacy, and yet they put so much information about themselves on their Smartphones.

Even astrophysicist Carl Sagan when speaking astronomically used Billions, not Trillions.

1 user thanked author for this post.

OscarCP

Reply | Quote

So, bringing my further contribution to this cheerful exchange of views:

“Ex Machina”, a 2015 movie to chill your heart; it is supposed to be about events that take place in a remote building housing also the cybernetics laboratory of a creepy genius in the middle of  a Swedish boreal forest:

https://www.youtube.com/watch?v=EoQuVnKhxaM

And here is what Stephen Hawking had to say about all this (click off the Bitcoin popup: it is irrelevant — or is it?):

https://www.cnbc.com/2017/11/06/stephen-hawking-ai-could-be-worst-event-in-civilization.html

Seriously not good.

 

Ex-Windows user (Win. 98, XP, 7); since mid-2017 using also macOS. Presently on Monterey 12.15 & sometimes running also Linux (Mint).

MacBook Pro circa mid-2015, 15" display, with 16GB 1600 GHz DDR3 RAM, 1 TB SSD, a Haswell architecture Intel CPU with 4 Cores and 8 Threads model i7-4870HQ @ 2.50GHz.
Intel Iris Pro GPU with Built-in Bus, VRAM 1.5 GB, Display 2880 x 1800 Retina, 24-Bit color.
macOS Monterey; browsers: Waterfox "Current", Vivaldi and (now and then) Chrome; security apps. Intego AV

1 user thanked author for this post.

Charlie

Reply | Quote