ChatGPT - artificial intelligence.

[Agent_zed]

The whole point of technology is to increase productivity i.e. to reduce the amount of work needed.

Assuming you believe this to be a good thing then it completely contradicts your previous statements about how the rich are too rich. The reason they are rich is because they have increased productivity at the expense of workers. Bezos is hugely rich as he has automated a huge amount of the system and funnels the excess money into his bank account.

AI will only continue to take jobs and sadly the vision of the future, that has been touted for decades, where everyone works 2 days a week and plays tennis and goes for a jog on the other days has simply not appeared. The lowest people on the rung now work 2 jobs instead at great speed to ensure 'efficiency' whilst a few people near the top get paid a fortune to keep the system running and design new ways of getting rid of those pesky inefficient humans. Once AI is good enough it should be able to not rely on those technicians either as it will be able to diagnose and fix itself, it will just need someone to manually change a part that it has already ordered.

distopian future? most of this is already happening.

edit - Purely by chance a story about it this morning Amazon strikes: Workers claim robots are treated better

 

[paulrbarnard]

Funny how so many threads turn in to a quagmire like this. There does however seem to be a common trigger point. I’m starting to get fed up with it. Perhaps it is time to go elsewhere.

 

[artie]

Funny how so many threads turn in to a quagmire like this. There does however seem to be a common trigger point. I’m starting to get fed up with it. Perhaps it is time to go elsewhere.

Deja vu.

This keeps coming up in comments.

No one is compelled to read or participate in any thread.

On a side note which might even be on topic.

What is the driving force and perceived benefit of AI.

Surely not that students can have a perfect answer to any question in milliseconds and thereby ace any exam.

 

[paulrbarnard]

Deja vu.

This keeps coming up in comments.

No one is compelled to read or participate in any thread.

On a side note which might even be on topic.

What is the driving force and perceived benefit of AI.

Surely not that students can have a perfect answer to any question in milliseconds and thereby ace any exam.

The problem is that the thread is about AI but has turned in to a political/ social argument that is completely off topic for the thread. Sure I don’t have to read it but when so many threads go off topic like this it is difficult to not see it. There is a specific forum for this type of discussion. I wish the moderators would stamp on this type of exchange in the open fora.

 

[paulrbarnard]

Deja vu.
On a side note which might even be on topic.

What is the driving force and perceived benefit of AI.

Surely not that students can have a perfect answer to any question in milliseconds and thereby ace any exam.

AI is a far wider topic than cheating in exams. My entire job is dependant on AI. My contract prevents me from engaging in discussion on the topic though.

There is an excellent body called “AI for good” run by the ITU that hold very regular seminars and discussions on AI applied to many areas and the wider implications of those. I highly recommend signing up for notifications of forthcoming talks. The next one is “The Future of Robots for Good – The Quest for Embodied AI”. It free.

 

[Agent_zed]

The problem is that the thread is about AI but has turned in to a political/ social argument that is completely off topic for the thread. Sure I don’t have to read it but when so many threads go off topic like this it is difficult to not see it. There is a specific forum for this type of discussion. I wish the moderators would stamp on this type of exchange in the open fora.

Sorry I'm guilty of this, although for this topic it is such a big and important topic that effects everyone that it is almost impossible for it not to be political/social.

If i ask what Chisel is good or bad it won't make much difference to many people as it won't affect their life if I choose a narex or a stanley.

AI reaches into all of our lives in a constantly evolving way, from automated telephone assistants to agriculture, medicine and businesses across the globe. This isn't just a neat website that can answer questions it is the beginnings of a completely new way of automation.

 

[Jacob]

Assuming you believe this to be a good thing then it completely contradicts your previous statements about how the rich are too rich. The reason they are rich is because they have increased productivity at the expense of workers

Not quite that cut and dried - the "workers" too are party to the increased productivity. Gained productivity and lost work yes, but then begs the question of livelihoods of the now unemployed

.....

distopian future? most of this is already happening.

Has been happening since before the start of the industrial revolution. The Luddites were not opposed to increased productivity they were concerned about losing their livelihoods, though they tend to get conflated.

edit - Purely by chance a story about it this morning Amazon strikes: Workers claim robots are treated better

 

[Spectric]

I dont think you can look at AI as being a single concept, I think it has various subdivisions and depends how a person interprets it. A lot will think of AI as an intelligent self thinking android but it could equally be a non mobile product. A topic such as AI will go off topic because it has so many implications that will involve everyones lives and many will perceive it as a threat to jobs and even the need for humans. So from what angle do people think AI is ?

To start AI is artificial intelligence, so define intelligence: From Wiki

Intelligence has been defined in many ways: the capacity for abstraction, logic, understanding, self-awareness, learning, emotional knowledge, reasoning, planning, creativity, critical thinking, and problem-solving. More generally, it can be described as the ability to perceive or infer information, and to retain it as knowledge to be applied towards adaptive behaviors within an environment or context.

So now you just need to produce this in a machine. Not going to be easy, self awareness seems a long way off and emotional knowledge, never seen any sensors that can measure these parameters in order to provide an input to a system. Adaptive behavior should be a lot easier as we already have adaptive strategies in digital systems, but it is the ones we as humans take for granted will be the showstoppers with current technology.

Then what I see as a big difference is that humans and the world we live in are analogue, a world of continous data unlike digital that works on just samples at points in time, ie discrete so it is technology that needs to change to allow the inclusion of human traits like emotion and awareness.

 

[Thingybob]

You can have the AI view on it

Are children being indoctrinated to exist in a world that won't exist in 15 years?

It is difficult to predict with certainty what the world will look like in 15 years, but it is important for children to be educated about a wide range of topics so they are prepared to navigate a constantly changing world. This includes learning about environmental issues, such as climate change, as well as developing critical thinking skills and learning about different cultures and perspectives. Additionally, it is also important for children to learn about technology and how it is rapidly advancing, as this will likely play a significant role in shaping the future.

When I was at school I used to watch Tomorrows World on a Thursday with Ramond Baxter and I'm still waiting for most of the things shown to materialise 60 odd years later

 

[Jacob]

When I was at school I used to watch Tomorrows World on a Thursday with Ramond Baxter and I'm still waiting for most of the things shown to materialise 60 odd years later

We still haven't had our atomic powered cooking stove. Just one little pellet of uranium to keep it going for life! I thought it was too good to be true even back then.

 

[selectortone]

As an aside - the subject of AI got me thinking about processor chips and transistor density and last night I went down a proper google rabbit-hole on the subject.

I'm old enough to remember the first germanium transistors replacing valves. I trained as a TV and radio service engineer back in the early 70s, just as the introduction of easily swapped-out modular, plug in transistorised printed circuit boards made fault-finding a thing of the past. Remember the Rediffusion guy with his toolbox and soldering iron who came to your house and fixed your rented, valve powered TV? That was me!

The first widely available microprocessors consisted of a few thousand transistors. The first IBM PCs used in homes and offices, powered by the Intel 8088 had 29,000 transistors. By the time of the Intel 80386 we were up to 275,000.

This may be old news to some, but I was staggered to read that the latest Apple chip, the Apple M1 Ultra contains 114 billion transistors. Yep, 114,000,000,000. Wow.

The highest transistor count in a single chip processor is of the deep learning processor Wafer Scale Engine 2 by Cerebras, it has 2.6 trillion MOSFETs

This is what is powering AI. And now AI is designing the next generation of chips, with parallel processing on stacks of chips, each with trillions of transistors. There's food for thought.

And another little titbit: it has been estimated that a total of 13 sextillion (1.3×10 to the 22) transistors have been manufactured worldwide between 1960 and 2018.

All in my lifetime (72 years and counting). Amazing.

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

 

[jetsetwilly]

I dont think you can look at AI as being a single concept, I think it has various subdivisions and depends how a person interprets it. A lot will think of AI as an intelligent self thinking android but it could equally be a non mobile product. A topic such as AI will go off topic because it has so many implications that will involve everyones lives and many will perceive it as a threat to jobs and even the need for humans. So from what angle do people think AI is ?

To start AI is artificial intelligence, so define intelligence: From Wiki

Intelligence has been defined in many ways: the capacity for abstraction, logic, understanding, self-awareness, learning, emotional knowledge, reasoning, planning, creativity, critical thinking, and problem-solving. More generally, it can be described as the ability to perceive or infer information, and to retain it as knowledge to be applied towards adaptive behaviors within an environment or context.

So now you just need to produce this in a machine. Not going to be easy, self awareness seems a long way off and emotional knowledge, never seen any sensors that can measure these parameters in order to provide an input to a system. Adaptive behavior should be a lot easier as we already have adaptive strategies in digital systems, but it is the ones we as humans take for granted will be the showstoppers with current technology.

Then what I see as a big difference is that humans and the world we live in are analogue, a world of continous data unlike digital that works on just samples at points in time, ie discrete so it is technology that needs to change to allow the inclusion of human traits like emotion and awareness.

Totally agree, AI is so broad as to be nearly meaningless. I work in the industry, and we don't often use the term 'AI', instead either referring to things by the task they perform or the tech they employ. Nobody really talks about intelligence in my field, but then we are applying AI, not studying it. We do talk about the ability of a model to 'reason', 'infer' etc, but these are narrowly defined.

I don't really agree about the continuous vs discrete problem - that challenge has been met adequately many times in many ways eg in audio and video. A maybe related point is that human perception is generally multimodal (lots of simultaneous signals of different types) and human cognition is multichannel (you’re never really just doing one thing). I don’t know if anyone is trying to model that.

 

[Spectric]

that challenge has been met adequately many times in many ways eg in audio

But many will still say that a pure analogue sound system is purer than a digitised one, but then the sampling on newer media is far superior to the original Cd quality and any compression techniques are much improved.

Really all this shows just what an amazing organ a human brain can be, nothing is really as simple as you would initially think and as you say we are multimodal in that we are not just viewing with our eyes as many of our other senses come into play, our vision actually has a feel to it and so why a given view is wonderful on a nice sunny day but not so when its cold and damp.

 

[jetsetwilly]

But many will still say that a pure analogue sound system is purer than a digitised one, but then the sampling on newer media is far superior to the original Cd quality and any compression techniques are much improved.

Really all this shows just what an amazing organ a human brain can be, nothing is really as simple as you would initially think and as you say we are multimodal in that we are not just viewing with our eyes as many of our other senses come into play, our vision actually has a feel to it and so why a given view is wonderful on a nice sunny day but not so when its cold and damp.

It's a great point that there's more to the perception of a sunny day than just the visual signal. There's also more to perception than just perception - we're also simulltaneously fitting the current experience into our memory, thereby triggering and replaying other memories, and all the while continuing to process other recent experiences (and even some not-so-recent, if they are important enough) and anticipate future ones, all at the same time and in the same brain. Some of that happens in artificial neural networks (one of the design principles is that activation of a pathway increases its probability of future activation) but to me that's like comparing a burning stick to an F1 car.

 

[BucksDad]

@paulrbarnard This week I found the feature that lets you ignore complete threads or even the entire sub-forum. Sometimes I turn on ignore for Off Topic

 

[Terry - Somerset]

The usual definition applied to intelligence relates to the rational - the ability to understand and learn well, and to form judgments and opinions based on reason. Most exams test precisely this.

This avoids "softer" attributes - emotional, creative, perceptions, moral, ethical, etc.

AI is already, or shortly will be the equal of, or better than, homo sapiens in the rational. It can more completely store, access, and analyse data to find optimal solutions. It is potentially stronger, faster, more reliable than flesh and bone.

Rational analysis applied to the emotional may not get a consistently better outcome. Beauty, taste, colour, feelings, are individual and variable. Rational analysis may ultimately seek to categorise responses rather than embracing diversity. This could be a source of extreme social stress.

 

[SDev]

I really recomend the BBC Reith lectures 2021 in which Stuart Russell, a British computer scientist working at UC Berkeley, gives an excellent overview of AI and it’s recent and possible future impacts on our lives. Brilliantly delivered, clear and totally terrifying. Also an excellent Q&A session at the end. Should be required listening for our politicians. I listened to them whilst building some bookcases in the workshop … Think that should count as “on topic”

 

[Scruples]

You'd compare exam results?
How could we know about a machine's sentience? Aren't they are either just 'on' or 'off' ?

That alone would be the deciding factor, Jacob. When a machine can react to a situation in a way that appears to be considered, and without any specific programming in that specific area, then maybe it could be an early indication of sentience. There's a long way to go but the development of neural networks and artificial organic memory is on the up.

 

[Jacob]

That alone would be the deciding factor, Jacob. When a machine can react to a situation in a way that appears to be considered, and without any specific programming in that specific area, then maybe it could be an early indication of sentience. There's a long way to go but the development of neural networks and artificial organic memory is on the up.

But intelligent animals are heavily programmed by evolution, into being what they are.
I think "sentience" doesn't actually have much meaning - a thermostat is "sentient" in that it senses, and is "intelligent" if it has been programmed to make choices.

 

[Spectric]

a thermostat is "sentient" in that it senses, and is "intelligent" if it has been programmed

It cannot be both sentient and programmed, if it was sentient it would FEEL the cold rather than just send a signal to a boiler when a bi-metalic contact closes. Have you ever seen a thermostat shiver !