By continuing to use the site, you agree to our use of cookies and to abide by our Terms and Conditions. We in turn value your personal details in accordance with our Privacy Policy.
Please log in or register. Registered visitors get fewer ads.
Any thoughts on this? As a web bod, its scared the crap out of me, but what do you lot think?
If whole tranches of jobs disappear, what becomes of society? Do we just fck off and die? Become serfs feeding on crumbs thrown from the top tables? Or even just enjoy all our new free-time as we lead mission-less lives?
[Post edited 3 May 2023 11:29]
Has anyone ever looked at their own postings for last day or so? Oh my... so sorry. Was Ullaa
Soon there will be no need for contributors to TWTD as AI will be able to start and continues threads much better than we all can!
Seriously though, this seems to be advancing much more quickly than is good for society. Look at today's fall in the value of shares in education publishing companies and the like. Look at the fact that a science fiction book publisher has stopped accepting novels from authors. And just wonder what effect it will have on education more generally, if students can effectively cut corners or even cheat?
[Post edited 3 May 2023 11:11]
0
That AI, nicking our jobs on 11:12 - May 3 with 2224 views
Think we're safe for a while, as AI bots do not yet have the vote (and we still use a physical paper system in the UK).
In reality, they're still just a useful tool, if properly set up. Or toys designed to mimic human behaviour and give plausible answers. He talks about no need for code, but somebody had to set the system up and input the "rules" for analysing the transcript. Plus any AI is only as clever as the original programmers, with as much knowlege as it is allowed to access. It can just think faster and in greater depth than humans.
As a threat to jobs, it's no worse than the advent of industrialisation, factory robots or computers - all of which we've survived as a race.
Soon there will be no need for contributors to TWTD as AI will be able to start and continues threads much better than we all can!
Seriously though, this seems to be advancing much more quickly than is good for society. Look at today's fall in the value of shares in education publishing companies and the like. Look at the fact that a science fiction book publisher has stopped accepting novels from authors. And just wonder what effect it will have on education more generally, if students can effectively cut corners or even cheat?
[Post edited 3 May 2023 11:11]
Part of the issue is that humans can't have nice things. On the face of it, ChatGPT can be a hugely useful tool to enhance learning.
But if you can't really be arsed, then ChatGPT is an easy corridor for lackluster working and not understanding source material.
That we live in an instant gratification society doesn't help either.
You also have an insane amount of YouTubers/Influencers talking about how you can use AI to write articles, create content etc. It's no wonder certain markets are being flooded.
0
That AI, nicking our jobs on 11:24 - May 3 with 2167 views
It's a measure of how deeply we're into the Bad Timeline that AI, which in sci-fi is often used to liberate humans from mundane soul-destroying tasks and jobs so we can concentrate on art and poetry, etc etc, is being used to create art and poetry, etc, etc so that humans can have more time to work at our mundane, soul destroying jobs.
[Post edited 3 May 2023 11:31]
I'm one of the people who was blamed for getting Paul Cook sacked. PM for the full post.
Soon there will be no need for contributors to TWTD as AI will be able to start and continues threads much better than we all can!
Seriously though, this seems to be advancing much more quickly than is good for society. Look at today's fall in the value of shares in education publishing companies and the like. Look at the fact that a science fiction book publisher has stopped accepting novels from authors. And just wonder what effect it will have on education more generally, if students can effectively cut corners or even cheat?
[Post edited 3 May 2023 11:11]
Both your examples are worrying.
Ai seems to have the power to disrupt almost everything - do we really want it?! Then again, I guess we have to have it for security as someone's going to develop it. Just another thing to fear - maybe climate change will never be a problem as the machines will have done away with us by then.
Has anyone ever looked at their own postings for last day or so? Oh my... so sorry. Was Ullaa
That AI, nicking our jobs on 11:12 - May 3 by Guthrum
Think we're safe for a while, as AI bots do not yet have the vote (and we still use a physical paper system in the UK).
In reality, they're still just a useful tool, if properly set up. Or toys designed to mimic human behaviour and give plausible answers. He talks about no need for code, but somebody had to set the system up and input the "rules" for analysing the transcript. Plus any AI is only as clever as the original programmers, with as much knowlege as it is allowed to access. It can just think faster and in greater depth than humans.
As a threat to jobs, it's no worse than the advent of industrialisation, factory robots or computers - all of which we've survived as a race.
I really hope you're right, but fear you maybe wrong. The article talks of exponential growth in it's capabilities and coupled with the advances in Quantum computing....
Maybe there'll be a backlash Handmaid's Tale style, hopefully without the complete subjugation of women?
Has anyone ever looked at their own postings for last day or so? Oh my... so sorry. Was Ullaa
That AI, nicking our jobs on 11:12 - May 3 by Guthrum
Think we're safe for a while, as AI bots do not yet have the vote (and we still use a physical paper system in the UK).
In reality, they're still just a useful tool, if properly set up. Or toys designed to mimic human behaviour and give plausible answers. He talks about no need for code, but somebody had to set the system up and input the "rules" for analysing the transcript. Plus any AI is only as clever as the original programmers, with as much knowlege as it is allowed to access. It can just think faster and in greater depth than humans.
As a threat to jobs, it's no worse than the advent of industrialisation, factory robots or computers - all of which we've survived as a race.
[quote] It can just think faster and in greater depth than humans. [quote]
One of the big problems around AI is the misuse of language regarding its description. AI software do not actually "learn", "think", or any of the other things that we attribute to them because of our natural drive to personify the things around us, particularly those things which are designed to mimic cognisance. But they are not conscience beings, they are large statistical models. They cannot think at all, they are confined by the limitations and biasses of their training data and the complexity of the statistical model that allows them to interpolate between data points. This is why chatGPT is actually so bad at doing creative (i.e. artistic) tasks. Ask it to write a 3 stanza Limerick about cheese and it will not do a convincing job. Much to my ire.
4
That AI, nicking our jobs on 12:06 - May 3 with 2051 views
That AI, nicking our jobs on 11:39 - May 3 by Cotty
[quote] It can just think faster and in greater depth than humans. [quote]
One of the big problems around AI is the misuse of language regarding its description. AI software do not actually "learn", "think", or any of the other things that we attribute to them because of our natural drive to personify the things around us, particularly those things which are designed to mimic cognisance. But they are not conscience beings, they are large statistical models. They cannot think at all, they are confined by the limitations and biasses of their training data and the complexity of the statistical model that allows them to interpolate between data points. This is why chatGPT is actually so bad at doing creative (i.e. artistic) tasks. Ask it to write a 3 stanza Limerick about cheese and it will not do a convincing job. Much to my ire.
What concerns me though, is what happens when the AI is left connected to and with full access to the source code and the metadata post-deployment. If you give the job of creator AND maintenance engineer to the AI, then the AI can create something not that great, but observe human or environmental responses to the creation and then modify the code to optimise against whatever targets it was originally given, be they inert or malevolent.
# WE ARE STEALING THE FUTURE FROM OUR CHILDREN --- WE MUST CHANGE COURSE #
That AI, nicking our jobs on 12:06 - May 3 by NthQldITFC
What concerns me though, is what happens when the AI is left connected to and with full access to the source code and the metadata post-deployment. If you give the job of creator AND maintenance engineer to the AI, then the AI can create something not that great, but observe human or environmental responses to the creation and then modify the code to optimise against whatever targets it was originally given, be they inert or malevolent.
I know this is a rubbish explanation, but that basically just isn't how it works. You'd have to originally make something do what you're saying from the outset.
That AI, nicking our jobs on 11:28 - May 3 by giant_stow
I really hope you're right, but fear you maybe wrong. The article talks of exponential growth in it's capabilities and coupled with the advances in Quantum computing....
Maybe there'll be a backlash Handmaid's Tale style, hopefully without the complete subjugation of women?
On the bright side it could trip up the lumbering, planet-destroying behemoth of suicidal, unfettered, global capitalism in favour of a more caring, balanced society with evenly distributed happiness, health and actual wealth... but it won't unless it is significantly less stupid than its creators.
# WE ARE STEALING THE FUTURE FROM OUR CHILDREN --- WE MUST CHANGE COURSE #
That AI, nicking our jobs on 12:13 - May 3 by DanTheMan
I know this is a rubbish explanation, but that basically just isn't how it works. You'd have to originally make something do what you're saying from the outset.
Why? Using your poem example as a simple analogy, what's to stop an AI creating a poem, publishing it on youtube, monitoring the sections which are played least frequently, updating its dataset/algorithms on that basis, rewriting those sections, redeploying and testing until they are more popular than the original?
If we're not talking about a poem but something more sinister, humans have lost control.
(I must admit, I haven't thought this through very deeply!)
# WE ARE STEALING THE FUTURE FROM OUR CHILDREN --- WE MUST CHANGE COURSE #
That AI, nicking our jobs on 11:39 - May 3 by Cotty
[quote] It can just think faster and in greater depth than humans. [quote]
One of the big problems around AI is the misuse of language regarding its description. AI software do not actually "learn", "think", or any of the other things that we attribute to them because of our natural drive to personify the things around us, particularly those things which are designed to mimic cognisance. But they are not conscience beings, they are large statistical models. They cannot think at all, they are confined by the limitations and biasses of their training data and the complexity of the statistical model that allows them to interpolate between data points. This is why chatGPT is actually so bad at doing creative (i.e. artistic) tasks. Ask it to write a 3 stanza Limerick about cheese and it will not do a convincing job. Much to my ire.
Indeed. I was using the term "think" in the sense of carrying out processes/calculations (tho, arguably, that is all "thinking" is, anyway). "Depth" in the same sense as chess players thinking a number of moves ahead. Translation: A computer can simply do more calculations faster than humans.
It is also likely to have limitations on technical tasks, where there will need to be input from experts on the subject as well as programmers setting up the necessary complexity of decisionmaking.
Even if an AI can access a vast amount of data (by using the internet, say). They have no innate sense of what is accurate/true and what isn't (let alone grey areas) beyond what is input by the original coder. It could read every journalistic article published on the web and it would all be nothing more meaningful than a string of zeroes and ones.
At least the benefits will be spread across society and it definitely won't create a system with the rich at the top and the rest at the bottom with a downward spiral on wages as we all undercut each other trying to get one of the few jobs left.
That AI, nicking our jobs on 12:26 - May 3 by NthQldITFC
Why? Using your poem example as a simple analogy, what's to stop an AI creating a poem, publishing it on youtube, monitoring the sections which are played least frequently, updating its dataset/algorithms on that basis, rewriting those sections, redeploying and testing until they are more popular than the original?
If we're not talking about a poem but something more sinister, humans have lost control.
(I must admit, I haven't thought this through very deeply!)
But it is just doing that as a mimic, or positive reinforcement machine. There is no conscious drive behind it, or end goal other than whatever it was originally designed to do (in this case try iterations of random poetry according to certain rules and check it against received consumer satisfaction levels).
That AI, nicking our jobs on 12:36 - May 3 by Guthrum
But it is just doing that as a mimic, or positive reinforcement machine. There is no conscious drive behind it, or end goal other than whatever it was originally designed to do (in this case try iterations of random poetry according to certain rules and check it against received consumer satisfaction levels).
I see what you're both saying now - it's not coming up with its own goals. What if the goal it is given is much broader and allows for the setting of sub-goals and spawning of new independent AIs to achieve those sub-goals? (I'm aware I'm pretty much just blurting stuff out here, to see what happens)
# WE ARE STEALING THE FUTURE FROM OUR CHILDREN --- WE MUST CHANGE COURSE #
.. we're doing a lot of research and work in this space, and using AI on a daily basis. A few quick observations:
- AI has already transformed our lives through the fairly simple algorithms used by social media, they've 'polarised' the planet somewhat, changing views and behaviours - Some of the more doomsday scenarios are a bit far fetched, given the lack of integration (including to the physical world) and the quality of data.. will they come to pass, depends on the safeguards put in place - Generative AI is the hot topic and becoming a game changer - 'A branch of artificial intelligence that uses machine learning to create new content or data that did not exist before, such as text, images, music, code, etc.​' - that's your ChatGPT... the speed of development is very fast due to competition - Some of the vision recognition and graphic/image creation development has gone exponential.. if you combine with knowledge mgmt systems... you can effectively replace engineers inspecting things. And of course you get the whole deep fake video and image direction. - think of AI as either assisting current jobs, replacing current jobs, or creating new jobs. The middle is indeed scary as jobs such as admin and back office roles - think finance, supply chain - really become redundant.. similarly even white collar roles such as solicitors become commoditised
That article is sensationalisation at best and fear-mongering at worst. AI is a long way off being anywhere near to human intelligence/taking our jobs (other than very basic jobs which are susceptible to automation anyway).
Some interesting viewing here:
Trust the process. Trust Phil.
0
That AI, nicking our jobs on 13:15 - May 3 with 1873 views
That AI, nicking our jobs on 13:00 - May 3 by NthQldITFC
I see what you're both saying now - it's not coming up with its own goals. What if the goal it is given is much broader and allows for the setting of sub-goals and spawning of new independent AIs to achieve those sub-goals? (I'm aware I'm pretty much just blurting stuff out here, to see what happens)
An AI can only come up with sub-goals if it's been told to and given the tools to perform that task. It can't suddenly decide it needs an assistant and clone itself, sci-fi style (in reality, would simply get slower and slower as it ran out of capacity).
After all, the mechanical AI which tells your kettle to switch off when its boiling hasn't suddenly decided it needs a toaster. You had to go out and buy one.
That AI, nicking our jobs on 13:00 - May 3 by NthQldITFC
I see what you're both saying now - it's not coming up with its own goals. What if the goal it is given is much broader and allows for the setting of sub-goals and spawning of new independent AIs to achieve those sub-goals? (I'm aware I'm pretty much just blurting stuff out here, to see what happens)
You've watched too much sci-fi fella!
Trust the process. Trust Phil.
1
That AI, nicking our jobs on 13:21 - May 3 with 1857 views
That AI, nicking our jobs on 12:16 - May 3 by NthQldITFC
On the bright side it could trip up the lumbering, planet-destroying behemoth of suicidal, unfettered, global capitalism in favour of a more caring, balanced society with evenly distributed happiness, health and actual wealth... but it won't unless it is significantly less stupid than its creators.
The problem is it's the unfettered global capitalists who are creating the AI. Why will they care about anything apart from the bottom line? It's an arms race with little to no consideration for society.
I must admit AI scares me shirtless, the progress in the past few years is mind boggling. We will either be redundant as a species or extinct before long at the rate we're going.
1
That AI, nicking our jobs on 13:30 - May 3 with 1835 views
That AI, nicking our jobs on 11:39 - May 3 by Cotty
[quote] It can just think faster and in greater depth than humans. [quote]
One of the big problems around AI is the misuse of language regarding its description. AI software do not actually "learn", "think", or any of the other things that we attribute to them because of our natural drive to personify the things around us, particularly those things which are designed to mimic cognisance. But they are not conscience beings, they are large statistical models. They cannot think at all, they are confined by the limitations and biasses of their training data and the complexity of the statistical model that allows them to interpolate between data points. This is why chatGPT is actually so bad at doing creative (i.e. artistic) tasks. Ask it to write a 3 stanza Limerick about cheese and it will not do a convincing job. Much to my ire.
Far beit for me to question a stats man, but if that database is big enough and on a powerful enough machine, surely it can do anything:
That AI, nicking our jobs on 12:26 - May 3 by NthQldITFC
Why? Using your poem example as a simple analogy, what's to stop an AI creating a poem, publishing it on youtube, monitoring the sections which are played least frequently, updating its dataset/algorithms on that basis, rewriting those sections, redeploying and testing until they are more popular than the original?
If we're not talking about a poem but something more sinister, humans have lost control.
(I must admit, I haven't thought this through very deeply!)
I'm by no means an expert (this is more along Cotty's area of expertise I believe) but for your very specific example, it may be possible.
But that's the thing, it's a very, very specific example. It may get very good not at writing poems, but at writing something that the YouTube algorithm likes. By the end, it may not even represent a poem.
As Cotty mentioned, these things don't really reason or "learn" as most people understand it. It has to have really good data for it to give you things that are even vaguely meaningful.
Like many tools, they could be used for a lot of harm and used for a lot of good too. And it's all just maths, which you can't really ban (see also encryption).
However, producing a song in the style of a particular artist is not a difficult task. Recent court cases involving performers from Ed Sheeran to Led Zeppelin show that popular music tends to follow certain well-worn tropes - likely those most pleasing to the human ear. Lyrics are often very generic.
Assuming a modicum of musical knowlege and some software to mimic the singing voice (which are sometimes vocoded anyway), you could probably do it yourself.
The transgression happened when some human posted it on a streaming service, pretending (or implying) it came from the artists named. That wasn't the AI (altho other AIs may have been responsible for spreading the song's distribution, but only because it already showed signs of popularity).
That AI, nicking our jobs on 13:15 - May 3 by Guthrum
An AI can only come up with sub-goals if it's been told to and given the tools to perform that task. It can't suddenly decide it needs an assistant and clone itself, sci-fi style (in reality, would simply get slower and slower as it ran out of capacity).
After all, the mechanical AI which tells your kettle to switch off when its boiling hasn't suddenly decided it needs a toaster. You had to go out and buy one.
But surely createChild($objSubgoal, $doNoHarm=0, $independent=1, $allowProgeny=1) could be called on some condition being met, and deployed to a cloud server account with more-or-less limitless data access and processing capacity, could it not?
# WE ARE STEALING THE FUTURE FROM OUR CHILDREN --- WE MUST CHANGE COURSE #
That AI, nicking our jobs on 13:43 - May 3 by Guthrum
However, producing a song in the style of a particular artist is not a difficult task. Recent court cases involving performers from Ed Sheeran to Led Zeppelin show that popular music tends to follow certain well-worn tropes - likely those most pleasing to the human ear. Lyrics are often very generic.
Assuming a modicum of musical knowlege and some software to mimic the singing voice (which are sometimes vocoded anyway), you could probably do it yourself.
The transgression happened when some human posted it on a streaming service, pretending (or implying) it came from the artists named. That wasn't the AI (altho other AIs may have been responsible for spreading the song's distribution, but only because it already showed signs of popularity).
I take what you're saying and you're right that about music following well-established patterns and rules, but isn't that true with most creative or thinking industries? Be it music, graphic design, writing, financial trading, data analysis, coding etc...
Has anyone ever looked at their own postings for last day or so? Oh my... so sorry. Was Ullaa
By continuing to use the site, you agree to our use of cookies and to abide by our Terms and Conditions.
