This is scary - how AI is now even writing the articles you read

[Marty_d]

Spacey, not sure if your grandson is on a Work for the Dole scheme or something else - it's my (limited) understanding that the $20 or however much is only for transport costs, participants don't actually get paid for WFD.   Your grandson should talk to Services Australia if he thinks he's being short changed.  As for losing the benefit if you get 2 days work, I don't think that's right either - the earnings will reduce your payment but not dollar for dollar, you can earn a certain amount without it affecting and then it's a sliding scale.  As long as you're actively looking for other work as well I think you can continue to receive benefits.  https://www.servicesaustralia.gov.au/income-and-assets-tests-for-jobseeker-payment?context=51411#a1

[old man emu]

The point is that if a bloke's willing to work, then he should be able to get a permanent job. Casual work in a lot of instances is simply uneconomic. An employer still has to pay worker's comp insurance and holiday pay. The only thing they save on is sick pay and long service. If you compare the annual costs of those two costs with the loss of productivity of have to keep training people how to do a job, then waiting for them to get up to the productivity level of the permanent workforce. I reckon you'd find it was more profitable to make a job permanent. Besides, if an employer was smart, the money required for sick pay, annual leave and long service leave can be put into some interest bearing account, and the interest becomes part of the business' profit.

 

There are some people who are happy with casual work, but they are not likely to be desperate to get money for the basics. The money they earn from a casual job is a supplement, not the only source.  

[Marty_d]

10 minutes ago, old man emu said:

The point is that if a bloke's willing to work, then he should be able to get a permanent job. Casual work in a lot of instances is simply uneconomic. An employer still has to pay worker's comp insurance and holiday pay. The only thing they save on is sick pay and long service. If you compare the annual costs of those two costs with the loss of productivity of have to keep training people how to do a job, then waiting for them to get up to the productivity level of the permanent workforce. I reckon you'd find it was more profitable to make a job permanent. Besides, if an employer was smart, the money required for sick pay, annual leave and long service leave can be put into some interest bearing account, and the interest becomes part of the business' profit.

 

There are some people who are happy with casual work, but they are not likely to be desperate to get money for the basics. The money they earn from a casual job is a supplement, not the only source.  

There's no holiday pay for casual work.  

[onetrack]

Quote

I look out to the highway and see men installing a culvert and drain, and soon others will come along to complete the rebuilding of the road.

Yep, you should regularly give thanks to us well-trained Roads Scholars, for our clever abilities to make your trips smooth, fast and safe.  

 

Edited June 3, 2022 by onetrack

[old man emu]

3 hours ago, Marty_d said:

There's no holiday pay for casual work. 

It's included in the hourly rate.

[spacesailor]

NO !.

If payed out of bosses pocket, 

AND !. Not put on the books, nothing gets to the government pockets.

Then if that worker dobs in her/his bos, no more work there !, ever.

( wifey paid $ 50 to tide granson over (  Don,t tell that government ).

spacesailor

[Jerry_Atrick]

Looks like they will start basic waiting tasks - and at 30k upfront cist, already cheaper than a junior waiter

Canberra cafe uses robot technology to deliver food and drinks to tables 
https://www.abc.net.au/news/2022-06-13/canberra-cafe-embraces-robot-automation/101146300

[nomadpete]

Jerry, I apologise.

I did not know what GPT-3 is.

 

Now I am afraid. Quite so.

 

I just read about philosophic conversations with GPT-3.

 

Do you think it represents genuine self realisation? Or (hopefully) just programmed responses to certain word sequences?

[pmccarthy]

I enjoyed my conversations with ELIZA in the 1970s.

[spacesailor]

And !

I hate my female voiced G,P,S, forever telling me l drive like SHEET.

Never get a word in edgeways. Even when l say " l want to go a different way this time, 

SHE always overrules me.  LoL

spacesailor

 

[old man emu]

11 minutes ago, spacesailor said:

And !

I hate my female voiced G,P,S, forever telling me l drive like SHEET.

Never get a word in edgeways. Even when l say " l want to go a different way this time, 

SHE always overrules me.  LoL

spacesailor

 

Then don't take the missus driving with you.

[spacesailor]

No the misses just keeps quite,  whenever that dammed gps lady gets on Her high horse. 

Can,t win .

spacesailor

[nomadpete]

If google has achieved sentient computer 'people' who have access to infinite memory and access to virtually all of humanity's collective knowledge, what happens next? How soon? How many more other such artificial super entities are there? What is there to stop multitudes of these being created? What happens if they decide they dissaprove of our politics?

 

What happens if they disagree with each other?

[spacesailor]

EASY ,.

TERMINATOR

spacesailor

[spacesailor]

The Worlds population is far to great.

Remedy

TERMINATOR.

NO Human sentiment fron A,I.

spacesailor

 

[nomadpete]

10 hours ago, spacesailor said:

 

TERMINATOR.

NO Human sentiment fron A,I.

spacesailor

 

We already have a number of live world leaders that already fit that description, Spacey.

We don't need a bunch of untouchable smarter electric ones that have access to all our knowledge and especially our human history.

 

When these new ones get around to reading the story of '2001 - a Space Odessy', and the get to the bit about what humans did to Hal 9000, we might be in a spot of trouble.

[Marty_d]

14 hours ago, spacesailor said:

And !

I hate my female voiced G,P,S, forever telling me l drive like SHEET.

Never get a word in edgeways. Even when l say " l want to go a different way this time, 

SHE always overrules me.  LoL

spacesailor

 

It's like someone sat back and thought "Hmmm, I think we should design a back-seat driver with a nagging voice, just for when your wife can't be in the car!"

 

At least there's a volume control on the GPS.

[onetrack]

I never listen to her - at least, that's what she constantly tells me.

[Marty_d]

53 minutes ago, onetrack said:

I never listen to her - at least, that's what she constantly tells me.

That's what mine says too... at least, I think that's what she says...

[Jerry_Atrick]

On 13/06/2022 at 8:41 AM, nomadpete said:

Jerry, I apologise.

I did not know what GPT-3 is.

 

Now I am afraid. Quite so.

 

I just read about philosophic conversations with GPT-3.

 

Do you think it represents genuine self realisation? Or (hopefully) just programmed responses to certain word sequences?

Let's be clear - no machine can be sentient or self-realising. It is only as good as two things:

1. The data set fed into the system
2. The data points it uses to determine the probability factors to come up with the answer.

 

As time and time has shown, if you have garbage in, you get garbage out, Also, as time and time has shown, if the algos are deficient. it will produce a poor result regardless of the data. I giess, in some ways it is like people - in that if they are ignorant (poor data inputs) or their brain isn't wired quite right (poor algos), then the result will not be great, but this is an oversimplified analogy.

 

The other thing to think about is the ability to process tumultuous data sets rapidly, So far, the brain outpaces any machine in that regard, bit is poor with sequential data sets. If you look at your room, you process everything in an instant. Sure, it is an abstraction based on many cognitive factors, but no computer today could comprehend all that information in the time you do - even the barest of rooms would be too challenging for a computer.

 

Quantum computing will be a game changer in that space - but even then, it will come close to human capacity for complex data structure processing. The data sets are getting more and more comprehensive, most, if not all of the theoretical maths problems have been solved, and we are seeing consistently, if linear improvements in processing power.

 

The short answer, after all that, is that it is a programmatic response that takes a human input (as an example), that uses an optimised data set to determine what it means, and then uses a larger data set to determine what the appropriate response should be. That is, effectively AI 101.

[onetrack]

Google have now produced an AI imaging programme that will provide you with a perfect image, merely by inputting a text description of what you want in an image.

 

https://www.creativebloq.com/news/google-ai-image-generator

[Marty_d]

Careful with that link, my anti-virus software didn't like it.

[onetrack]

Marty, your AV programme is obviously tuned to reject AI as dodgy.  

 

My MalwareBytes is pretty hot on stopping me from visiting many websites, but it doesn't stop me from visiting the site above.

[nomadpete]

On 15/06/2022 at 7:23 AM, Jerry_Atrick said:

• The data set fed into the system
• The data points it uses to determine the probability factors to come up with the answer.

Jerry, you seem to think that only applies to computers. It applies to peoples, too. Are you questioning the whole concept of "intelligence", whether human or nonhuman?

 

Equally, GIGO applies to humans as much as it does to AI. (Eg: conspiracy theorists)

 

I can see that the AI example was largely a 'suitable phrase' response selected by an algoithm. But I posit that the human mind generally works the same way. Except when I'm pyssed.

 

But AI generally has decision making and learning, and instant access to all the information on the web. And the ability to take uncommanded action.

[Jerry_Atrick]

4 hours ago, nomadpete said:

Jerry, you seem to think that only applies to computers. It applies to peoples, too. Are you questioning the whole concept of "intelligence", whether human or nonhuman?

 

Equally, GIGO applies to humans as much as it does to AI. (Eg: conspiracy theorists)

 

On 14/06/2022 at 10:23 PM, Jerry_Atrick said:

I giess, in some ways it is like people - in that if they are ignorant (poor data inputs) or their brain isn't wired quite right (poor algos), then the result will not be great, but this is an oversimplified analogy.

(compllete with typo - was on the phone.

 

My point more simply is machines have no inherit cognitive capability - it is simulated and will be constrained by the algos and data points it is fed. Machine learning algos do come closest. As an illustration, a machine learning algo may use some volume point analysis to determine the most likely response a human would make. I think it is fair to say, more of the human race is racist and conspiracy theorist than not; therefore, the this machine learning algo will align itself with the majority view. This was more or less that design of the Tay bot the Microsoft launched and then decommissioned. Now, the algo can be changed to correlate different approaches, and have some value system applied, but whos values?

 

This is the key difference betwween AI and "intelligence", AI can only think how they are programmed. Yes, humans more or less think how they are programmed or indoctrinated, brainwahed, etc., but they can stop and hopefully use their frontal cortex to think things through, and deviate from a pre-programmed way of thinking.. AI can do this, but only in the confines of its algos and data points.

 

And yes, an AI algo has access to the entire web of data (which it normally trawls, catergorises and abstracts for its purposes, otherwise latency would be an  issue). And the AI bot can sift through this unstructured, but more or less sequentiial data infinitelly faster than a human. But, close your eyes before you walk into a room you have never seen before. Upon openeing your eyes, you willl instantly comprehend the room, because our brain is wired to process massive amounts of data in parallel; an AI bot/algo today would take a lot longer than is to comprehend the room. Quantum computing will change that, though.

 

Don;t get me wrong; advances in AI are as both exciting and scary. With the righ socio-economic-political leadership, a lot of great things can be achieved. However, corporate and power greed will probably mean the potential for overall societal good will not be achieved until some form of civil unrest takes place.