Should ChatGPT bots be banned from the EEVblog Forum?

[tggzzz]

ChatGPT reminds me of autumns in 1990s, when freshmen students all over the world got access to Unix-type machines, and discovered newsgroups, filling them with their nonsense, including "Test" posts... 

That was so much a phenomenon that AOL's appearance was immortalised as "the year September never ended".

I'm also old enough to have used email for a couple of blissful years completely free of mass e-mail spam.

It wasn't all blissful. I remember UUCP "bang addressing" where the sender specified the sequence of every machine that the email passed through on the way to the recipient. I was, however, just too late to have seen the famous {decvax,philabs}!mcvax!moskvax!kremvax!chernenko post

Oh how I miss conversatlons like "Drat, it bounced at foomail" "Well try ucbvax; that's pretty well connected". Not.

In my opinion, it's just one more shitspout added to the world.  Moar content!  Quality does not matter!  Just make sure it looks nice and plausible!

Yes, it is going to be a real problem for search engines that rely on getting eyeballs to adverts. But that problem has been becoming worse for many a year

[Nominal Animal]

I'm also old enough to have used email for a couple of blissful years completely free of mass e-mail spam.

It wasn't all blissful.

No, but the early to mid nineties here in Finland, in this respect, was: we had Funet, with a central FTP repository nic.funet.fi, and everyone had Arpanet-style mailboxes (user@host.domain).  No software patents, and the neighboring Soviet Union had just dissolved: future looked bright.

[pcprogrammer]

ChatGPT reminds me of autumns in 1990s, when freshmen students all over the world got access to Unix-type machines, and discovered newsgroups, filling them with their nonsense, including "Test" posts...  I'm also old enough to have used email for a couple of blissful years completely free of mass e-mail spam.

In my opinion, it's just one more shitspout added to the world.  Moar content!  Quality does not matter!  Just make sure it looks nice and plausible!

Oh how we long for them good old days when everything was better and simpler. 

No software patents, and the neighboring Soviet Union had just dissolved: future looked bright.

Yes, and we were still young and eager to take on the world. 

But now we are grumpy old man complaining about the future. 

[xrunner]

But now we are grumpy old man complaining about the future. 

Exactly so. Sounds like a bunch of grumpy old men complaining about kids walking on their lawn, and they had to walk uphill both ways to school in the snow. 

May be time for a little Bob Dylan -

Come gather ’round people

Wherever you roam

And admit that the waters

Around you have grown

And accept it that soon

You’ll be drenched to the bone

If your time to you is worth savin’

Then you better start swimmin’ or you’ll sink like a stone

For the times they are a-changin’


Bob Dylan

[Neutrion]

But see guys, not all of are so grumpy and closed minded despite the age, some started to "swim":   

https://www.theguardian.com/world/2023/mar/02/romania-ion-ai-government-honorary-adviser-artificial-intelligence-pm-nicolae-ciuca

The home grown AI tells the PM what people think.

(I try to see the positive side, with a few script Hungary might be able to get back Transsylvania without a single shot    )

[Nominal Animal]

But now we are grumpy old man complaining about the future. 

Hey!    I am complaining about things that will make the future crappier; I'm not complaining about the future per se.

There is a distinction.  It may be a tiny one, but it is an important one to me! 

Sounds like a bunch of grumpy old men complaining about kids walking on their lawn, and they had to walk uphill both ways to school in the snow. 

Now that's just uncalled for!  I don't even have a lawn.

Besides, I described a time when things were easier, because there were fewer spouts of crap.  I had it easy when I started my Uni studies.  Today, the ones who want to learn for themselves, need to compete against those who just copy-paste and use ChatGPT and pay others to do their homework for them.

So, technically, I'm defending the youngsters, not complaining about them.

I'm telling you that any benefits you think you'll see from the use of ChatGPT will be vastly outnumbered by the downsides.

[xrunner]

I'm telling you that any benefits you think you'll see from the use of ChatGPT will be vastly outnumbered by the downsides.

What would you like me to do about it? 

[pcprogrammer]

May be time for a little Bob Dylan -

Talking about an old man    Don't know if he is grumpy though. 

Sounds like a bunch of grumpy old men complaining about kids walking on their lawn, and they had to walk uphill both ways to school in the snow. 

For walking uphill both ways you also have to go down hill, so something is wrong here. 

I do have a lawn but only had unwanted cows walk on it, twice. 

P. S. don't mind cows, they are nice, but they have to stay in their own fields.

[PlainName]

But see guys, not all of are so grumpy and closed minded despite the age, some started to "swim":   

https://www.theguardian.com/world/2023/mar/02/romania-ion-ai-government-honorary-adviser-artificial-intelligence-pm-nicolae-ciuca

The home grown AI tells the PM what people think.

(I try to see the positive side, with a few script Hungary might be able to get back Transsylvania without a single shot    )

Hmmm.

Developed by Romanian researchers, Ion’s main task will be to scan social networks to inform the government “in real time of Romanians’ proposals and wishes”, Ciuca said on Wednesday.

Cancel culture magnification set to 11. Whatever could go wrong?

[Nominal Animal]

I'm telling you that any benefits you think you'll see from the use of ChatGPT will be vastly outnumbered by the downsides.

What would you like me to do about it? 

Not call me a grumpy old man complaining about kids walking on their lawn, having had to walk uphill both ways to school in the snow.

I am a grumpy old man, but I like todays' kids.  I didn't have to walk uphill both ways to school in the snow; it was actually a rather nice walking/biking/skiing distance for the first six years, after that I took the bus.

All I said was ChatGPT was just another shitspout, making future kids' lives harder.  It's exactly as cheerful a thing as microplastics in the environment.

P. S. don't mind cows, they are nice, but they have to stay in their own fields.

Cleaning after the patties isn't fun, but cows themselves are quite funny.  Except Highland cows, they're scary.

[xrunner]

Sounds like a bunch of grumpy old men complaining about kids walking on their lawn, and they had to walk uphill both ways to school in the snow. 

For walking uphill both ways you also have to go down hill, so something is wrong here. 

The grumpy old man would tell you "Stop over-analyzing what I'm saying! You kids!" 

[MT]

The whole money making business of advertising on the net is very annoying and drives me away from for instance youtube. The more adverts show up the less I will use it, and I'm not going to pay for it, because the ratio crap to usefulness is to bad.

"uBlock Origin" if you want zero ads wherever you go.

[RoGeorge]

I'm telling you that any benefits you think you'll see from the use of ChatGPT will be vastly outnumbered by the downsides.

What would you like me to do about it? 

- Ask for AI regulation, ask to ban AI at least from:

• autonomous AI driven weapons (in military)
• propaganda bots (in politics)
• over-representing and/or public human opinion-flooding by AI bots (in social media)

- Don't get the victim mindset

• “Defeat is a state of mind; no one is ever defeated until defeat has been accepted as a reality.” (in life)

- Don't use ChatGPT when not needed

• don't add an unnecessary ChatGPT bot/service (in EEVblog/forum)

First 3 actions that came to mind.  Pretty sure there are plenty more.


This all ChatGPT frenzy looks to me like preaching the benefits of radioactive cosmetics from the 50s. 
https://www.orau.org/health-physics-museum/collection/radioactive-quack-cures/pills-potions-and-other-miscellany/tho-radia-items.html

https://awesci.com/radioactive-toothpaste-shocking-ad-50s/

[xrunner]

Now that's a rational answer! 

[MT]

https://fortune.com/2023/01/06/schools-chatgpt-ai-student-cheating-ban/

Can it be used for writing school papers?
This is what ChatGPT said when The Associated Press asked it to answer that question in all caps from the perspective of a principal shouting a brief message through a school’s PA system:

“DO NOT USE CHATGPT OR ANY OTHER AUTOMATED WRITING TOOL FOR SCHOOL PAPERS. THIS IS
CHEATING AND WILL NOT BE TOLERATED. IF YOU ARE CAUGHT USING CHATGPT OR ANY OTHER
AUTOMATED WRITING TOOL FOR SCHOOL PAPERS, THERE WILL BE SERIOUS CONSEQUENCES.”

But when asked to answer the same question on its own, ChatGPT offered this more measured warning:
“As a general rule, it is not appropriate to use ChatGPT or any other automated writing tool for school
papers, as it is considered cheating and does not benefit the student in the long run.”

[rstofer]

But will it change our lives for the better? The jury is still out, but I'll bet it won't.

Entirely beside the point about whether or not to have a discussion section about it here. I think it's very worthwhile and I'll add it. if no one uses it, no big deal.

ChatGPT is a lot like Windows 1.0 - a decent first cut.  Over the next 10 years we can expect a massive change in how things get done and we'll all be looking back at where it started.  It would be nice to see the evolution and keeping track with a section about AI in general or specifically ChatGTP derivatives seems important.

I'm also interested in classification and recommender projects (I want to crack Wall Street - as does every other person playing with AI) and I'm sure it's just a matter of pulling in enough features.  More memory, larger NVIDIA board, networked computers, etc.  Math, meet Computer Science

I have been playing with the Digit Recognition code in various forms and it's pretty interesting.  It's the "Hello World" of AI.  Who would have contemplated a 784x784x1 matrix back in college circa '73?  And that's just for a toy neural network.  Heck, 4x4 was hard enough to cause a mad rush to Fortran.

ChatGTP is just one small piece of AI and there's a lot of interesting work going on in other arenas.

[rstofer]

https://fortune.com/2023/01/06/schools-chatgpt-ai-student-cheating-ban/

Can it be used for writing school papers?
This is what ChatGPT said when The Associated Press asked it to answer that question in all caps from the perspective of a principal shouting a brief message through a school’s PA system:

“DO NOT USE CHATGPT OR ANY OTHER AUTOMATED WRITING TOOL FOR SCHOOL PAPERS. THIS IS
CHEATING AND WILL NOT BE TOLERATED. IF YOU ARE CAUGHT USING CHATGPT OR ANY OTHER
AUTOMATED WRITING TOOL FOR SCHOOL PAPERS, THERE WILL BE SERIOUS CONSEQUENCES.”

But when asked to answer the same question on its own, ChatGPT offered this more measured warning:
“As a general rule, it is not appropriate to use ChatGPT or any other automated writing tool for school
papers, as it is considered cheating and does not benefit the student in the long run.”

ChatGPT probably realizes that not every institution has banned it.  Google for different approaches being taken.

https://www.theguardian.com/technology/2023/feb/27/chatgpt-allowed-international-baccalaureate-essays-chatbot

Teaching will have to adapt and thesis writers will have to get a real job.

[MT]

https://fortune.com/2023/01/06/schools-chatgpt-ai-student-cheating-ban/

Can it be used for writing school papers?
This is what ChatGPT said when The Associated Press asked it to answer that question in all caps from the perspective of a principal shouting a brief message through a school’s PA system:

“DO NOT USE CHATGPT OR ANY OTHER AUTOMATED WRITING TOOL FOR SCHOOL PAPERS. THIS IS
CHEATING AND WILL NOT BE TOLERATED. IF YOU ARE CAUGHT USING CHATGPT OR ANY OTHER
AUTOMATED WRITING TOOL FOR SCHOOL PAPERS, THERE WILL BE SERIOUS CONSEQUENCES.”

But when asked to answer the same question on its own, ChatGPT offered this more measured warning:
“As a general rule, it is not appropriate to use ChatGPT or any other automated writing tool for school
papers, as it is considered cheating and does not benefit the student in the long run.”

ChatGPT probably realizes that not every institution has banned it.  Google for different approaches being taken.

https://www.theguardian.com/technology/2023/feb/27/chatgpt-allowed-international-baccalaureate-essays-chatbot

Teaching will have to adapt and thesis writers will have to get a real job.

Current ChatGPT model/algorithm have no ability to realize things.

[Bud]

This thing us just waiting for a couple SJWs to claim it is Good for <insert your favourite gender or social or other group here> and the poor principal will be cancelled.

[MK14]

Current ChatGPT model/algorithm have no ability to realize things.

Is that the right and best measuring stick for it?

Such lack of realization, doesn't seem to stop powerful chess playing computers, from beating, even the worlds best human players.  Or AI self driving cars, (when enabled/allowed, which sometimes happens, already, in some places, some of the time), driving me to where I need to go.

Although ChatGPT is NOT really intelligent, as such.  It does quite an impressive range of 'intelligence' like tasks, which even 10 or 20 years ago, seemed like a very distant dream.

It certainly seems to be able to read random texts, that it is given and process the text, and provide reasonably intelligent (like) responses to the text.  Showing a reasonable or better measure of 'understanding'.  Even though it doesn't really understand it (properly, yet!).

[rstofer]

https://fortune.com/2023/01/06/schools-chatgpt-ai-student-cheating-ban/

Can it be used for writing school papers?
This is what ChatGPT said when The Associated Press asked it to answer that question in all caps from the perspective of a principal shouting a brief message through a school’s PA system:

“DO NOT USE CHATGPT OR ANY OTHER AUTOMATED WRITING TOOL FOR SCHOOL PAPERS. THIS IS
CHEATING AND WILL NOT BE TOLERATED. IF YOU ARE CAUGHT USING CHATGPT OR ANY OTHER
AUTOMATED WRITING TOOL FOR SCHOOL PAPERS, THERE WILL BE SERIOUS CONSEQUENCES.”

But when asked to answer the same question on its own, ChatGPT offered this more measured warning:
“As a general rule, it is not appropriate to use ChatGPT or any other automated writing tool for school
papers, as it is considered cheating and does not benefit the student in the long run.”

ChatGPT probably realizes that not every institution has banned it.  Google for different approaches being taken.

https://www.theguardian.com/technology/2023/feb/27/chatgpt-allowed-international-baccalaureate-essays-chatbot

Teaching will have to adapt and thesis writers will have to get a real job.

Current ChatGPT model/algorithm have no ability to realize things.

All it has to do is find institutions that tolerate ChatGPT.  That's all I did, just find one counter example.

[MK14]

As cheap, small and good, electronic handheld calculators, came onto the market.  With green florescent displays, then (typically) red LED, and later LCD screens.

I remember big arguments in education, as to how they would handle the situation.  So, for a number of years (if I remember correctly, in the UK), calculators were totally banned from exam rooms, at least in schools.

But they (education establishments), eventually relented, and allowed calculators into exams, even maths exams (if I remember, correctly).

I suspect ChatGPT will or may, go through similar approval cycles.

Because at the end of the day.  The skills of using ChatGPT well and effectively, are probably going to be an important part of modern day, future activities.  In the same way, even using a calculator (or computer), well and effectively, also involves some skills.

[tggzzz]

As cheap, small and good, electronic handheld calculators, came onto the market.  With green florescent displays, then (typically) red LED, and later LCD screens.

I remember big arguments in education, as to how they would handle the situation.  So, for a number of years (if I remember correctly, in the UK), calculators were totally banned from exam rooms, at least in schools.

Such issues are older than calculators; I remember using this to annoy a teacher railing against calculators... In Phaedrus (~370BC), Plato recorded Socrates's discussion of the Egyptian myth of the creation of writing. In the process Socrates faulted writing for weakening the necessity and power of memory, and for allowing the pretense of understanding, rather than true understanding.

But they (education establishments), eventually relented, and allowed calculators into exams, even maths exams (if I remember, correctly).

Eventually yes, but they still disliked programmable calculators that could "invisibly store answers".

I also remember in my first year uni exams there was a question that could only reasonably be answered by using a calculator - but they didn't warn the students that was the case. I recognised the issue and got good marks; those that didn't got poor marks. Quite Right Too. Eric Laithwaite would have been proud

(Question was to indicate the voltages at the nodes of a non-inverting opamp circuit with an overall gain of 10)

I suspect ChatGPT will or may, go through similar approval cycles.

Because at the end of the day.  The skills of using ChatGPT well and effectively, are probably going to be an important part of modern day, future activities.  In the same way, even using a calculator (or computer), well and effectively, also involves some skills.

True. But, unlike with calculators, chatbots have the potential to spew across all aspects of our life. (Counterpoint: what proportion of spreadsheets contain significant errors?)

[tom66]

As cheap, small and good, electronic handheld calculators, came onto the market.  With green florescent displays, then (typically) red LED, and later LCD screens.

I remember big arguments in education, as to how they would handle the situation.  So, for a number of years (if I remember correctly, in the UK), calculators were totally banned from exam rooms, at least in schools.

But they (education establishments), eventually relented, and allowed calculators into exams, even maths exams (if I remember, correctly).

I suspect ChatGPT will or may, go through similar approval cycles.

Because at the end of the day.  The skills of using ChatGPT well and effectively, are probably going to be an important part of modern day, future activities.  In the same way, even using a calculator (or computer), well and effectively, also involves some skills.

I remember my maths teacher back in 2006-2007 telling me, "You're never going to have a calculator on your person all the time, so you'll have to learn how to do this all by rote."

Turned out that a few years later Apple launched the iPhone and a few years after that, smartphones were ubiquitous.  So, he was right, I did not always have a calculator,  I had a miniature computer with access to all the world's information at my fingertips, in my pocket at all times. 

The same thing will happen with large language models.  Teachers will have to adapt how they assess their students, but not including AI in the teaching syllabus now is (IMO) a terrible mistake.

[MT]

Is that the right and best measuring stick for it?
Such lack of realization, doesn't seem to stop powerful chess playing computers, from beating, even the worlds best human players.
Or AI self driving cars, (when enabled/allowed, which sometimes happens, already, in some places, some of the time), driving me to where I need to go.

My post was about "education". But ok lets switch topics, there is no real natural "intelligent" cars driving around. Chess is not a measurement tool of human intelligence , nothing intelligent in remembering vast number of sequential pattern combinations dictated under certain rules. Go watch interviews with Garry Kasparov talking geo/politics the dude is a complete moronic fool.

Although ChatGPT is NOT really intelligent, as such.  It does quite an impressive range of 'intelligence' like tasks, which even 10 or 20 years ago, seemed like a very distant dream.
It certainly seems to be able to read random texts, that it is given and process the text, and provide reasonably intelligent (like) responses to the text.
Showing a reasonable or better measure of 'understanding'.  Even though it doesn't really understand it (properly, yet!).

There you go, you answered all your own arguments. Notice your own reasoning/wording? "seams", "able to", "like" etc. The basic course in "philosophy" gives that "real" intelligence requires
a bare minimum of self awareness, self consciousness , etc. This area of science is enormously vast.

First thing in such a philosophy course is the "reality test" (if the course is of decent standards):
1: Take your finger.
2: Look straight forward.
3: Gently press that finger tip on the side of "one" of your eye globes.

What do you see? Well you see that "reality" is in the eye of the beholder.  Now inhale/eat a magic mushroom, what is your reality now?
Not that ChatGTP ever will do any of these things "out of the blue" (another self awareness/ self consciousness thing) based on its level of human created so called artificial intelligence.

If ChatGPT is so terrible as this thread and voting portray it to be why not just let it replace all teachers on all levels of education and we will see how that will turn out.
Which is what WEF, nazi Klaus Schwab have declared they want to do. Go watch interviews whit him, read his book The Great Reset, its full of AI ChatGTP's dictating "your" life.

What about a depressed ChatGPT and Tesla self driving algorithm and banking system?