So whats going to happen...

[Tedx]

Look, I don't pretent to understand any of this stuff....but I take more than a passing interest.

There seems to be 2 major technologies 'on the brink' so to speak (going by the general press - which probably means they're some way ahead of what they actually tell us (and going by the shennanigans at Open AI recently, maybe that's true)

1. Artificial General Intelligence via Open AI/ChatGPT and others

and

2. Quantum Computing - IBM and others are (apparently) making great strides in this area.

The 2 seem related to me (maybe they're not?)

But let's say they crack both fairly soon....what does it mean if they marry them together? Exponential growth in machine learning - exporing possibilities in seconds? Solving the greatest problems of mankind in the blink of an eye? A huge uptick in efficiency, growth? A massive increase in living standards across the globe?

Or......?

[Urbandreamer]

They are not related, one is hardware and the other software.

In addition ChatGPT et-al are not "Artificial General Intelligence", they are "Large language learning models".

Not to claim that they are unimpressive or useless, but they are not what you claimed them to be.

ChatGPT, as impressive as it is, belongs to a category of AI known as Narrow AI.

These systems are designed to excel in specific tasks but lack the ability to understand, learn, and apply knowledge across a wide range of tasks and domains like humans do.

Specifically ChatGPT is "Chat Generative Pre-Trained Transformer".
OpenAI works in the same space.
Both work well on language, but can't produce an image (digital art) unlike AI designed for that.

I'm not too surprised that anyone believes that these programs are more general purpose than they are, that is the story supplied by the media. It's just not true.

BTW, before someone points out that ChatGPT can write software, can I re-phrase the point. ChatGPT can use computer languages to express a desire. Err which bit of the word language did the journalists miss?

However to return to your point. The careful use of the likes of ChatGPT will indeed make society richer. Quantum computing "may" do so as well.

For quantum computing to make society richer it needs to solve computationally difficult problems the result of which benefit society above and beyond possible with current computation. Solutions to the traveling salesman problem have the possibility of saving fuel, thereby cutting cost and environmental harm. However the likes of Amazon have found methods of bypassing the problem. Statistically find out what is going to be ordered in a given area and store it locally.

It took us a significant amount of time from the invention of the maser in 1955 before we found uses for the principle everywhere in lasers. The same could be true with quantum computing.

[ReformedCharacter]

Specifically ChatGPT is "Chat Generative Pre-Trained Transformer".
OpenAI works in the same space.
Both work well on language, but can't produce an image (digital art) unlike AI designed for that.

Since there is wide debate about LLM and AGI, I'm not going to disagree with you. But it appears that some people believe that LLMs are approaching AGI. For example, I recently posted a link showing the huge improvements in GPT4's ability to play chess, despite the fact that it is not a chess computer.

https://www.youtube.com/watch?v=9LDaY7X2qGk

'Sparks of AGI' - Bombshell GPT-4 Paper: Fully Read w/ 15 Revelations...

GPT4 can output detailed images based on a text prompt.

https://www.youtube.com/watch?v=Mqg3aTGNxZ0

RC

[mc2fool]

For quantum computing to make society richer it needs to solve computationally difficult problems the result of which benefit society above and beyond possible with current computation. Solutions to the traveling salesman problem have the possibility of saving fuel, thereby cutting cost and environmental harm. However the likes of Amazon have found methods of bypassing the problem. Statistically find out what is going to be ordered in a given area and store it locally.

It took us a significant amount of time from the invention of the maser in 1955 before we found uses for the principle everywhere in lasers. The same could be true with quantum computing.

Yeah, but it's also the case that practical uses can go in different directions to the "obvious" that's predicted...

"The guard returned with a bulky long-barreled object cradled in his arms. The woman wrote: “Do you recognize this?” Koskinen shook his head. He couldn’t see the thing very well. “A maser gun. It amplifies radiation by stimulating atoms to re-emit in a highly collimated beam. Call it a heat ray.”" https://archive.org/details/Fantastic_v11n06_1962-06/page/n75/mode/2up?q=maser

Plenty of sci-fi of the time was predicting maser/laser ray guns and the like. I don't know of any that predicted CD players and light shows...

[Infrasonic]

Cisco using AI for Enterprise Firewalls...https://www.youtube.com/watch?v=n_-QuGvQXso

[Grumpsimus]

One existing use for Quantum computers is picking the random numbers for the Premium Bond Prize draw. It can pick out all three million winning numbers in 12 minutes.

This means winners are notified far more quickly than they used to be.

[DrFfybes]

But let's say they crack both fairly soon....what does it mean if they marry them together?

Would

you

like

to

play

a

game?

[Tedx]

Global Thermonuclear Warfare by any chance?

[88V8]

Artificial General Intelligence via Open AI/ChatGPT and others
and
Quantum Computing - IBM and others are (apparently) making great strides in this area.
...let's say they crack both fairly soon....what does it mean if they marry them together?

A chatbot that's harder to bypass so you get to speak with a human... but you only think it's a human...

V8

[CliffEdge]

What's going to happen? Things'll get worse.

[MuddyBoots]

I did a cryptography course on Futurelearn and it predicted a great leap forward in both encryption and decryption using quantum computers because of their greater speed. So what worries me is that soon your data will only be safe if you have a quantum computer; with traditional machines and current encryption methods your passwords will be easily cracked.

[XFool]

...That reminds me. I must reset all my passwords to: IfYouCanReadThisYourComputerWillSelfDestructInTenSeconds

[Tedx]

They are not related, one is hardware and the other software.

Right.

So my PC is hardware, my copy of Windows 10 is software. Neither can work without the other. And they work well together (most of the time!). They allow me to do things that only a couple of decades ago were either very difficult or thought impossible.

So what I'm getting at is if you bring together the most advanced hardware (i.e. a Quantum Computer) and the most 'intelligent' AGI software (ChatGPT26 or whatever Google are cooking at the moment), where will we be/what will happen? Intelligent 'life'? The ability to solve anything in minutes/seconds? Intelligent automation on a grand scale? New materials/drugs/processes that have never been thought of before?


Not to claim that they are unimpressive or useless, but they are not what you claimed them to be.

I dont think I claimed them to be anything. I don't know enough about them to claim anything. Hence my question......

[Infrasonic]

...I dont think I claimed them to be anything. I don't know enough about them to claim anything. Hence my question......

Just search for things like current real world use implementations of AI / ML (machine learning) / Quantum computing - there's plenty out there.

It's early days so it is far from perfect but the combination of smart people and all the above is going to change the world as we know it at an exponential rate, some for the better some for the worse. I just hope the former outweighs the latter...

On the quantum / encryption point. Google and others already have 'quantum proof' solutions...
https://www.theregister.com/2023/08/12/ ... hrome_kem/

Google has started deploying a hybrid key encapsulation mechanism (KEM) to protect the sharing of symmetric encryption secrets during the establishment of secure TLS network connections.

Devon O'Brien, technical program manager for Chrome security, explained on Thursday that starting in Chrome 116 – due August 15 – Google's browser will include support for X25519Kyber768, an alphanumeric salad that desperately needs a catchy name.

The unwieldy term is a concatenation of X25519, an elliptic curve algorithm that's currently used in the key agreement process for establishing a secure TLS connection, and Kyber-768, a quantum-resistant KEM that last year won NIST's blessing for post-quantum cryptography...

Cont.

[Tedx]

Yeah, i know I could google it and probably understand some of it....but I think its a worthy thread to have on LF given the implications.

And you lot have the ability to explain it to me, shall we say....in simple terms

[Tedx]

A brand new substance, which could reduce lithium use in batteries, has been discovered using artificial intelligence (AI) and supercomputing.

It goes on to say:

The two organisations achieved this by using advanced AI and high-performance computing which combines large numbers of computers to solve complex scientific and mathematical tasks.

Executive vice president of Microsoft, Jason Zander, told the BBC one of the tech giant's missions was to "compress 250 years of scientific discovery into the next 25".

"And we think technology like this will help us do that. This is the way that this type of science I think is going to get done in the future," he said.

Ok, supercomputers not Qantum computers (which I assume will be the next step) - but this is the kind of thing I was getting at.

What's next? The AI+Super/Qantum computer discovering materials that can be used inside a fusion reactor? Or a new drug that can cure dying? Or a new foodstuff compound that could feed the world?

https://www.bbc.co.uk/news/technology-67912033

[dionaeamuscipula]

A brand new substance, which could reduce lithium use in batteries, has been discovered using artificial intelligence (AI) and supercomputing.

It goes on to say:

The two organisations achieved this by using advanced AI and high-performance computing which combines large numbers of computers to solve complex scientific and mathematical tasks.

Executive vice president of Microsoft, Jason Zander, told the BBC one of the tech giant's missions was to "compress 250 years of scientific discovery into the next 25".

"And we think technology like this will help us do that. This is the way that this type of science I think is going to get done in the future," he said.

Ok, supercomputers not Qantum computers (which I assume will be the next step) - but this is the kind of thing I was getting at.

What's next? The AI+Super/Qantum computer discovering materials that can be used inside a fusion reactor? Or a new drug that can cure dying? Or a new foodstuff compound that could feed the world?

https://www.bbc.co.uk/news/technology-67912033

Computational chemistry is already being used in drug discovery, using techniques which I suspect are virtually identical to those used in the example above. I think (but I don't know, in fact I don't really know anything about this stuff beyond listening to a couple of low level talks by university chemistry departments) that AI is a big word here, lots of things are being described as "AI" when in reality they are nothing of the sort, but calling them AI releases funds that aren't triggered by boring old but large scale modelling.

DM

[servodude]

A brand new substance, which could reduce lithium use in batteries, has been discovered using artificial intelligence (AI) and supercomputing.

It goes on to say:

The two organisations achieved this by using advanced AI and high-performance computing which combines large numbers of computers to solve complex scientific and mathematical tasks.

Executive vice president of Microsoft, Jason Zander, told the BBC one of the tech giant's missions was to "compress 250 years of scientific discovery into the next 25".

"And we think technology like this will help us do that. This is the way that this type of science I think is going to get done in the future," he said.

Ok, supercomputers not Qantum computers (which I assume will be the next step) - but this is the kind of thing I was getting at.

What's next? The AI+Super/Qantum computer discovering materials that can be used inside a fusion reactor? Or a new drug that can cure dying? Or a new foodstuff compound that could feed the world?

https://www.bbc.co.uk/news/technology-67912033

Computational chemistry is already being used in drug discovery, using techniques which I suspect are virtually identical to those used in the example above. I think (but I don't know, in fact I don't really know anything about this stuff beyond listening to a couple of low level talks by university chemistry departments) that AI is a big word here, lots of things are being described as "AI" when in reality they are nothing of the sort, but calling them AI releases funds that aren't triggered by boring old but large scale modelling.

DM

AI is getting a bit "buzzy" these days
At one level I was doing AI 30 years back in my PhD ('machine listening using generic algorithms for heuristic analysis")
which meant we didn't have to explain what the computer was looking for... but we couldn't talk to it as if it was a twat in the pub... ...and it wouldn't lie to us in a plausible way either.
I've done loads of stuff since that extracts patterns in a heuristic way from data without getting my knickers twisted because it was trendy.
Can we please just be let alone to get on with it?

[1nvest]

HAL ... 2001 Space Odyssey were the pre-letters of IBM
NT were the subsequent letters of MS
AI ... subsequent letters are BJ

... a ring of potential chaos.

[servodude]

HAL ... 2001 Space Odyssey were the pre-letters of IBM
NT were the subsequent letters of MS
AI ... subsequent letters are BJ

... a ring of potential chaos.

Here's the thing.
It's not funny if you have to explain it...
..and I really think you need to here