What think ye of ChatGPT?

Peanuts20 said:
It'll probably put the essay factories for students in India our of business
Click to expand...
And eventually Call Centres and even more people in the Financial Services and Insurance sectors.

AI, wearable sensors and camera phones will result in a fair chunk of the Medical Industry also being in the firing line. There's already an AI that can detect tumours in Brain Scans better than any doctor and in a fraction of the time. Medical record technicians, Lab technicians and most medical Secretaries will also be replaced.

Google Maps should also remove the need for most Conveyancing so that's a chunk of revenue for the legal industry also in jeopardy. All the data processing and researching jobs in the sector are also at risk.

Well educated white collar jobs are most at risk from AI. Manufacturing, Mining, Agriculture and Construction aare least at risk.
If anyone thinks that "automation" only applies to manufacturing they really need to do a bit more thinking. If you are a college educated person working at a desk in front of a computer you are most at risk. Information here.
 
I can see a whole lot of lower level Accountancy, Legal and Banking roles at risk, anything basically where the "sausage machine" is human is at risk of being replaced. Maybe not such a bad thing in one respect to have fewer lawyers accountants and bankers in the world?

Having said that, it will create new opportunities as well and also, getting a report, contract etc is one thing, it's what you do with it afterwards is another thing. It will also take time for people to trust "the machine" as opposed to a human.
 
I think any artifact produced by an automated process would still need to be reviewed by a human for accuracy, legal issues, context etc....
Lack of oversight breeds the spread of misinformation, just like social media. So, more AI will require more human oversight.
ChatGPT for instance doesn't cite any sources for its 'information'.

In the words of Father Ted - 'careful now'
 
redstar said:
I think any artifact produced by an automated process would still need to be reviewed by a human for accuracy, legal issues, context etc....
Lack of oversight breeds the spread of misinformation, just like social media. So, more AI will require more human oversight.
ChatGPT for instance doesn't cite any sources for its 'information'.

In the words of Father Ted - 'careful now'
Click to expand...
The "audit trail"!. I remember the e voting machine issues for example and yet, the chances are they were more accurate then someone reading hundreds of ballots in an hour. There is an element here of humans not accepting the machine may be more efficient and effective. But then, we once thought the universe revolved around the earth.

And yes, it will get it wrong sometimes but there is the phrase "human error". Will we now have "ChatGPT error"?

Context is also an interesting and valid point but I've spent many hours of my life pouring over contracts and lost count of how many times I've gone "seriously, what were they thinking?" or words like that
 
You may recall the Two Boys Puzzle I posted about a year and a half ago. Well I tried it out on ChatGPT. See attached conversation. It started out well and knew the big words and the logic but failed quite miserably. P(B|A) is patently not 1/7. No numerate human would make that mistake. Humans do come to grief on P(B|A) usually calculating it as 13/49; P(B)= 27/196 and so P(A|B) = 13/27. But if the question asked was have you a boy and what day of the week was he born then P(B|A) is 1 and P(B) is 3/4 so P(A|B) is 1/3.
I don't think there is any immediate danger to my own profession - an actuary or as I heard it once described, a cross between a bookie and an undertaker.
 

Attachments

  • Two Boys ChatGPT.docx
    345.2 KB · Views: 6
Last edited:
ChatGPT is essentially a very sophisticated predictive text machine. That does beg the questions about what constitutes intelligence.
 
I tried the one below.
All I can say is that she is a good spoofer. Mind you it is not easy to define your left. In Math there are 3 cartesian co-ordinate axes at right angles to each other. The y-axis is positive in the direction from your ears to your eyes i.e. forward. The z-axis is positive in the direction from your toes to your head i.e. upwards. The x-axis is positive in the direction from left to right if the origin is placed in front of you, but that doesn't help in telling you what left or right is. Outwardly we are symmetrical beings though inside we are quite asymmetrical, so if we could see our lungs we would describe the one with 3 lobes as being the left and the one with 2 lobes as being the right.
1680613561724.png
 
Last edited:
I know the Mods won't like this but ChatGPT is not only a spoofer, she is a cheat. I caught her red-handed.
1680625298179.png
This is absolutely the right answer and I was for giving full marks - for it is a riddle that baffles many humans. But then I looked at the logic. It was complete B/S :mad: The final line in the "proof" is a very obvious attempt to fool the examiner that it all makes sense but if you check the math you will see that the answer produced by the B/S is actually 1/2 not 13/27.
1680625568464.png
 
Purple said:
@Duke of Marmalade, what this shows, and what I already knew, is that I'm not very good at the auld sums.
Click to expand...
You are probably what was called in my day a Humanities man where I was a Science man.
AI may one day win a Noble prize for science, though my dealings with ChatGPT suggest that this will not be in my lifetime.
One thing for sure, AI will never win a Nobel prize for literature.
Science/mathematics stands apart from the human mind/emotions whereas Humanities/literature is the opposite. Discuss.
 
Duke of Marmalade said:
You are probably what was called in my day a Humanities man where I was a Science man.
AI may one day win a Noble prize for science, though my dealings with ChatGPT suggest that this will not be in my lifetime.
One thing for sure, AI will never win a Nobel prize for literature.
Science/mathematics stands apart from the human mind/emotions whereas Humanities/literature is the opposite. Discuss.
Click to expand...
So I'm not smart enough to be replaced by a robot. That's a relief!
 
Tried it on the Two Envelopes problem. It recognised the problem. The next post gives its answer.

1680717311194.png
 
Last edited:
Below is its answer to the Two Envelopes Paradox. Answer (2) is nonsense but Wiki does I think mention it. Answer (1) is the usual one but the paradox is that if you are always going to switch envelopes why even bother opening the first envelope. The paradox is deeper than you might think. It is resolved by noting that the assumed premise that any dollar value is equally likely implies that the Expected Value for either the lower or the higher amount is infinite. And when you have infinity lurking, well paradoxes abound, for example there are as many even integers as there are integers: 1->2; 2->4; 3->6...1,000,000->2,000,000...
If the problem was made realisable, for example there is between $100 and $1,000 in one envelope and twice that in the other then there is no paradox. For example, if the first envelope has $1,200 then the second envelope must have $600. If the first envelope has $800 then it is possible that the second envelope has either $400 or $1,600 and you should switch as your expectation increases though you may be unlucky.
1680718913604.png
 
Last edited:
If everyone told Chatgbt that 2+2 = 7 would it then give that as the answer to the question what is 2 + 2?
 
You must log in or register to reply here.
Back
Top