ChatGPT it's incredible and scary at the same time

[nzoomed]

I've never paid alot of attention to AI in the past

Most previous efforts have never really been anything exciting, but ChatGPT is a gamechanger.
We are still only in its early development phases, who knows where it will go in another 5 to 10 years.
I wish I had this when I was at school, could have done my homework in 10 minutes!
We have just witnessed a huge disruption in the tech sector, if anyone has ever used siri or Alexa  this chatbot can do far more complex tasks.
My friend has been using it to write arduino code and it actually works and gives you a description on how to assemble it.
It's nuts, but still needs a way to go, some answers can at times be garbage but shows you what this is tech is capable of.

Try it out yourself at
chat.openai.com

[SueDonim]

It's overloaded at the moment. I did see an article about it the other day. Tutors and exam markers will need to stay on the lookout for student cheats but apparently if they are ontoit they should have no problem recognising the "work". Trouble is teachers don't seem to be as ontoit as they should be and so many people don't understand that the purpose of homework is to teach the skills needed to be disciplined, work independently and search for and analyse information. The last couple of years have highlighted how poor people are at those basics.

Personally I hate even the chatbots that companies use to provide "support". ChatGPT is clearly a whole new world and I find it pretty freaky, but I guess it will settle in to its useful purposes like helping your friend write code. Its power is pretty scary though.

[nzoomed]

Yes it has tutors in the universities worried, looks like they are going to switch to more oral assessments rather than written ones.
Apparently its costing 3 million dollars a day to run ChatGPT.

[zqwerty]

Here's a quote from the State of the World Conference I posted about earlier January about Chat-GDP:

"inkwell.vue.522 : State of the World 2023: Bruce Sterling and Jon Lebkowsky
permalink #332 of 332: Michael Brockington (jonl) Tue 24 Jan 23 14:38


(Via email from Michael Brockington - posted late, originally
received January 15)

Thank-you, everyone, for an ever-stimulating start to the year.

As a post-script, I thought an experimental investigation of
Chat-GPT's sense of humour might be worthwhile. I can report that
comedians (and poets?) have little to fear for now.
Humour is subjective, of course, but these results could at best be
characterized as lame. They bear some statistical resemblance to
humour, without actually being funny. On a par with dad-jokes,
perhaps?

Note also that a google search turns up some of these jokes
verbatim...

The dismal results:

>>Tell me a joke about iguanas
Why did the iguana break up with his girlfriend? Because she was
always too cold-blooded for him.

>>make a pun about bananas
Why did the banana go to the doctor? Because it wasn't peeling well.
[plagiarized verbatim from the internet]

>>demonstrate wordplay about the cold war
Why did the Soviet Union and the United States have a cold war?
Because they were too chilly to warm up to each other.

>>tell me a joke about an onion and a drainpipe
Why did the onion go to the drainpipe? To get a little privacy while
it cried.

>>make a joke
Why don't scientists trust atoms? Because they make up everything.
[plagiarized verbatim from the internet]

>>demonstrate witty repartee about dishes
Person 1: "I can't believe I have to do dishes again, it's such a
chore."
Person 2: "I guess you could say it's a plate full of trouble."
Person 1: "Ha, at least it will be a clean slate when we're done."
Person 2: "Yeah, but it's a lot of work to get to that point, it's a
real grind."

>>make a joke about death
I'm sorry, but making a joke about death may be considered offensive
or insensitive to some people. It's important to be respectful and
considerate of others' feelings and beliefs. Is there anything else
I could help you with?


Further investigation is required, to see how chat-GPT copes with
heckling."

[Oldfellah]

Yes it has tutors in the universities worried, looks like they are going to switch to more oral assessments rather than written ones.
Apparently its costing 3 million dollars a day to run ChatGPT.

If it costs that much where is the money coming from? And why is it free to use, I dont know if they have a paid version or not but having to pay for it wouldnt interest me.

[Kenj]

Wow, I've been playing around with it today. It's pretty darned good. Of course I asked it "The meaning of life the universe and everything"

It was considerably quicker than Hitchhikers Deep Thought. It took that computer 7.5 million years to come up with the answer, which was 42

Chat-GPT did it in seconds

K;

[king1]

just having a look at it now - seems to work pretty good...

might be a challenger for that minor Google problem here
https://tmmb.co.nz/forums/showthread.php?tid=1869

here was my Q & A

quantum mechanics and how it relates to theory of relativity

Quantum mechanics and the theory of relativity are two pillars of modern physics that describe the behavior of matter and energy on very different scales. While quantum mechanics provides a highly successful framework for understanding the microscopic world of atoms and subatomic particles, the theory of relativity describes the behavior of objects and events at macroscopic scales, such as stars and galaxies.

The two theories appear to be incompatible with each other, as quantum mechanics is based on the principles of indeterminacy and wave-particle duality, while the theory of relativity is based on the idea that the laws of physics are the same for all observers and that the speed of light is constant. The unification of these two theories into a single framework, known as quantum field theory, is a central goal of theoretical physics and has been the focus of much research in the last century.

While a fully consistent and coherent theory of quantum gravity has yet to be established, the marriage of quantum mechanics and the theory of relativity has provided a deeper understanding of the physical world and has led to many practical applications, such as the development of quantum computers and the description of black holes.

[Kenj]

just having a look at it now - seems to work pretty good...

might be a challenger for that minor Google problem here
https://tmmb.co.nz/forums/showthread.php?tid=1869

here was my Q & A

The workings of Google would be to much for any intelligent being, silicon or carbon!

Ken

[Oh_hunnihunni]

But haven't we all been talking to chatbots for an awful long time anyway? Seems to me you cannot phone any substantial organisation for assistance these days without having an AI triage your call, often deciding you do not merit human help until you become incomprehensible to the damned thing. Surely this version is simply the currently cleverest of the evolving form?

Perhaps we should test it not by being sensible communicators, but by being nonsensical ones...

[king1]

But haven't we all been talking to chatbots for an awful long time anyway? Seems to me you cannot phone any substantial organisation for assistance these days without having an AI triage your call, often deciding you do not merit human help until you become incomprehensible to the damned thing. Surely this version is simply the currently cleverest of the evolving form?

Perhaps we should test it not by being sensible communicators, but by being nonsensical ones...

i can sometimes get through to a human by pushing the phones number/*/# buttons at random times during the call...

[nzoomed]

Yes it has tutors in the universities worried, looks like they are going to switch to more oral assessments rather than written ones.
Apparently its costing 3 million dollars a day to run ChatGPT.

If it costs that much where is the money coming from? And why is it free to use, I dont know if they have a paid version or not but having to pay for it wouldnt interest me.

It might be free in monetary terms, just as Facebook and Gmail are also free.
Right now it's in the testing phase and they are likely using you as a user to help tweak and improve the system.
I believe they are launching a paid version that gives you priority when it's overloaded.

[Oh_hunnihunni]

But haven't we all been talking to chatbots for an awful long time anyway? Seems to me you cannot phone any substantial organisation for assistance these days without having an AI triage your call, often deciding you do not merit human help until you become incomprehensible to the damned thing. Surely this version is simply the currently cleverest of the evolving form?

Perhaps we should test it not by being sensible communicators, but by being nonsensical ones...

i can sometimes get through to a human by pushing the phones number/*/# buttons at random times during the call...

Exactly. I wonder if AI can act irrationally? Because until it can it will never grasp what being human is all about...

[king1]

i can sometimes get through to a human by pushing the phones number/*/# buttons at random times during the call...

Exactly. I wonder if AI can act irrationally? Because until it can it will never grasp what being human is all about...

We'll know we're in trouble when AI responds to irrational behaviour like HAL

[harm_less]

If it costs that much where is the money coming from? And why is it free to use, I dont know if they have a paid version or not but having to pay for it wouldnt interest me.

It might be free in monetary terms, just as Facebook and Gmail are also free.
Right now it's in the testing phase and they are likely using you as a user to help tweak and improve the system.
I believe they are launching a paid version that gives you priority when it's overloaded.

In the words of Andrew Lewis, “If you are not paying for it, you're not the customer; you're the product being sold.”

[nzoomed]

Exactly. I wonder if AI can act irrationally? Because until it can it will never grasp what being human is all about...

We'll know we're in trouble when AI responds to irrational behaviour like HAL

Most definitely!

[zqwerty]

Google Invests Almost $400 Million in ChatGPT Rival Anthropic

https://www.bloomberg.com/news/articles/...erify+wall

[harm_less]

Very in-depth analysis of the development and likely future of AI in this World In Data piece.