Post Reply 
ChatGPT for programming
03-29-2023, 12:27 AM
Post: #1
ChatGPT for programming
Hello friend, I was exploring this chatGPT and if you can teach it to program in user, system and ml and PPL is a tool with which I create, you can polish and improve various programs
Find all posts by this user
Quote this message in a reply
03-29-2023, 02:56 AM
Post: #2
RE: ChatGPT for programming
(03-29-2023 12:27 AM)hpmin48gx Wrote:  Hello friend, I was exploring this chatGPT and if you can teach it to program in user, system and ml and PPL is a tool with which I create, you can polish and improve various programs

Lot of members of this forum already tried to use this AI tool to elaborate some easy programme and the results are disappointing.

ChatGTP is essentially an expensive parrot which shoots columns of "RLC 0" and "GTO a" that are meaningless.

If you like, you can try yourself and see.

Good luck!



Put a calculator into your life!
Visit this user's website Find all posts by this user
Quote this message in a reply
03-29-2023, 02:25 PM
Post: #3
RE: ChatGPT for programming
My experience using ChatGPT as an RPN and RPL programming assistant has been positive.

Some relevant questions for those who have had disappointing results using ChatGPT as a programming aid are:

1) Did you use ChatGPT3.5 or ChatGPT4? ChatGPT4 is vastly improved for writing code.
2) Did you try varying an unsuccessful initial prompt?
3) As important as initial prompt wording, is whether you continued the conversation with ChatGPT to inform it of errors. ChatGPT responds to being told about errors by trying to correct them (and by apologizing Smile ) and often (but not always) succeeds. Providing ChatGPT with some context or guidance about the type of error can help.

While ChatGPT, including ChatGPT4, still can make a lot of programming errors, what it does correctly right now is jaw-dropping. And at the astounding rate that it's improving, I wouldn't want to bet against it.
Find all posts by this user
Quote this message in a reply
03-29-2023, 03:24 PM (This post was last modified: 03-29-2023 03:44 PM by Roberto Volpi.)
Post: #4
RE: ChatGPT for programming
(03-29-2023 02:25 PM)carey Wrote:  My experience using ChatGPT as an RPN and RPL programming assistant has been positive.

Some relevant questions for those who have had disappointing results using ChatGPT as a programming aid are:

1) Did you use ChatGPT3.5 or ChatGPT4? ChatGPT4 is vastly improved for writing code.
2) Did you try varying an unsuccessful initial prompt?
3) As important as initial prompt wording, is whether you continued the conversation with ChatGPT to inform it of errors. ChatGPT responds to being told about errors by trying to correct them (and by apologizing Smile ) and often (but not always) succeeds. Providing ChatGPT with some context or guidance about the type of error can help.

While ChatGPT, including ChatGPT4, still can make a lot of programming errors, what it does correctly right now is jaw-dropping. And at the astounding rate that it's improving, I wouldn't want to bet against it.

Well, there are threads already about it:

https://www.hpmuseum.org/forum/thread-19351.html

https://www.hpmuseum.org/forum/thread-19446.html

People far more experts than me have already given their opinions.

Furthermore, it is important to specify exactly what "positive experience" really means.

If negotiating with a PC is your scope, you can have a very good time, as ChatGTP has a spiffing sense of humor in its answers, but I question its ability to compile decent code for a HP calculator.

But maybe I am wrong, so we are going to see soon tons of programmes submitted by you and other users in this forum.

Peace.



Put a calculator into your life!
Visit this user's website Find all posts by this user
Quote this message in a reply
03-29-2023, 03:58 PM (This post was last modified: 03-29-2023 04:01 PM by carey.)
Post: #5
RE: ChatGPT for programming
(03-29-2023 03:24 PM)Roberto Volpi Wrote:  
(03-29-2023 02:25 PM)carey Wrote:  My experience using ChatGPT as an RPN and RPL programming assistant has been positive.

Some relevant questions for those who have had disappointing results using ChatGPT as a programming aid are:

1) Did you use ChatGPT3.5 or ChatGPT4? ChatGPT4 is vastly improved for writing code.
2) Did you try varying an unsuccessful initial prompt?
3) As important as initial prompt wording, is whether you continued the conversation with ChatGPT to inform it of errors. ChatGPT responds to being told about errors by trying to correct them (and by apologizing Smile ) and often (but not always) succeeds. Providing ChatGPT with some context or guidance about the type of error can help.

While ChatGPT, including ChatGPT4, still can make a lot of programming errors, what it does correctly right now is jaw-dropping. And at the astounding rate that it's improving, I wouldn't want to bet against it.

Well, there are threads already about it:

https://www.hpmuseum.org/forum/thread-19351.html

https://www.hpmuseum.org/forum/thread-19446.html

People far more experts than me and you have already given their opinions.

Furthermore, it is important to specify exactly what "positive experience" really means.

If negotiating with a PC is your scope, you can have a very good time, as ChatGTP has a spiffing sense of humor in its answers, but I question its ability to compile decent code for a HP calculator.

But maybe I am wrong, so we are going to see soon tons of programmes submitted by you and other users in this forum.

Peace.
Thank you for your reply and threads links! Note however that the threads are to comments posted in 2022 and before ChatGPT4. Anyone can just try ChatGPT4 and not rely on the opinion of others (including my own!). Importantly, it's not restricted to experts!

Programming is a fun activity and doesn't need ChatGPT, but for programs that go beyond recreational math, where the output might actually matter, it can be helpful in many ways as long as it is not relied on entirely (at least at present! Smile ).
Find all posts by this user
Quote this message in a reply
03-29-2023, 08:28 PM (This post was last modified: 03-29-2023 08:29 PM by pier4r.)
Post: #6
RE: ChatGPT for programming
The problem with LLM is that they need to be trained on something. The best training data is natural language. For code, the more niche it becomes, the harder it gets (be it GPT 3,5 or GPT4).

Let's be honest there isn't the same amount of code in RPN as there is in python on the internet.

Anyway I find that the biggest help would be using it for languages. Like "summarize this report", "provide me an outline for an article", "expand those ideas", "correct this paragraph", "describe this situation", "provide alternatives to this text" and so on. Even better with recommendations on (again) popular topics.

Code of course is ok too, but mostly in common languages, otherwise one stays there fixing things a lot. It is great to get a start.

Further one can ask about facts, like summaries of articles or books (a sort of review, before reading them).

Anyway one still has to be educated in the field. Examples (just asked, as of 2023-03-29).

Quote:could you provide me a summary of Galley Slave from Asimov?

"Galley Slave" is a short story written by Isaac Asimov, first published in 1957. The story is set in a future where space travel is common, and interstellar ships are powered by "gravitic" engines, which allow them to travel faster than light.

The protagonist, Andrew Martin, is a linguist who has been hired to work on a new spaceship that is being prepared for its maiden voyage. As part of his job, Andrew must learn the language of the aliens that the ship will encounter on its journey.

However, when the ship's captain dies unexpectedly, Andrew is forced to take command of the vessel. The crew, made up of prisoners serving as galley slaves, resents Andrew and refuses to obey him. Andrew realizes that the only way to gain their respect is to earn it through his actions.

Despite the challenges he faces, Andrew successfully leads the ship through several dangerous encounters with hostile aliens. Along the way, he earns the respect of the crew and proves his worth as a leader.

The story ends with Andrew being hailed as a hero upon the ship's return to Earth. Andrew is left wondering if he will continue his career as a linguist or embrace his newfound role as a leader.
That's wrong (and I recommend the short story!)

Quote:could you provide me a summary of "praises of idleness" of B. Russell?

"The Praise of Idleness" is an essay written by philosopher Bertrand Russell, first published in 1932. The essay argues that society should work towards reducing the amount of time people spend working, in order to increase leisure time and improve the overall quality of life.

Russell begins the essay by stating that, in modern society, there is a glorification of work and a belief that leisure time is unproductive and wasteful. He argues that this mindset is harmful, as it leads to overwork and stress, which in turn can lead to health problems and decreased productivity.

Russell then argues that reducing the amount of time people spend working would have several benefits. It would allow people to pursue leisure activities, such as art, music, and sports, which would enrich their lives and increase their happiness. It would also lead to a more equal distribution of work, as fewer people would be required to work long hours, and there would be more opportunities for people to pursue meaningful careers.

Furthermore, Russell argues that reducing the amount of work would lead to increased creativity and innovation. When people have more leisure time, they are able to think more deeply and engage in more creative pursuits, which can lead to new ideas and discoveries.

Finally, Russell argues that reducing the amount of work would lead to a more peaceful and harmonious society. When people have more leisure time, they are less likely to be stressed and unhappy, which can lead to a decrease in crime and social unrest.

Overall, "The Praise of Idleness" is a persuasive argument for the value of leisure time and the need to reduce the amount of time people spend working in modern society.
That is mostly correct, and I suggest you to read the original article. https://harpers.org/archive/1932/10/in-p...-idleness/

Quote:Could you provide a summary of "Feeling of power" from Asimov ?

"The Feeling of Power" is a science fiction short story written by Isaac Asimov, first published in 1958. The story is set in a future where humans have become heavily reliant on advanced technology, to the point where they have forgotten basic mathematical skills and rely on computers to do all calculations.

The protagonist, Myron Aub, is a low-level technician working for a government agency that oversees the operation of the world's computers. One day, Myron discovers a mathematical formula that allows him to perform calculations without the use of a computer. He realizes that this knowledge has been lost to society due to their reliance on technology.

Myron is excited by his discovery and begins to experiment with performing more and more complex calculations. However, he soon realizes that his newfound knowledge is dangerous. If the public were to discover that they no longer need computers to perform calculations, it could destabilize society and lead to chaos.

Myron becomes conflicted about what to do with his discovery. He ultimately decides to share it with his superiors, who are initially skeptical but eventually realize the potential of Myron's formula. They decide to keep it a secret and use it to gain power and influence over other nations.

The story ends with Myron realizing the terrible consequences of his discovery. He realizes that the power of his formula could be used for good, but instead, it has been co-opted by those in power for their own gain.

Overall, "The Feeling of Power" is a cautionary tale about the dangers of becoming too reliant on technology and the importance of preserving basic knowledge and skills. It also explores the themes of power, knowledge, and responsibility.
that's wrong again (and again I recommend the short story).

This to say, GPT is jaw dropping for the fact that is able to form sentences that are much better than mine. That is the command of the language is incredible for an algorithm, even if the system doesn't know the meaning of things. Nonetheless when asked about facts, reasoning and so on, it may contain errors that are not easy to spot if one is not good already. Thus one needs to be good or prepared.

How do I know if the summary is wrong, if I didn't read the original thing? How do I know that the code is wrong if I cannot code well? And so on.

Wikis are great, Contribute :)
Find all posts by this user
Quote this message in a reply
03-29-2023, 08:42 PM
Post: #7
RE: ChatGPT for programming
Additionally this is maybe of interest

https://arxiv.org/abs/2303.12712 (155 pages paper warning)

Wikis are great, Contribute :)
Find all posts by this user
Quote this message in a reply
03-29-2023, 10:44 PM
Post: #8
RE: ChatGPT for programming
.
Mentioned in The New York Times today:

Open Letter signed by Elon Musk & 1,000+ other people

V.

  
All My Articles & other Materials here:  Valentin Albillo's HP Collection
 
Visit this user's website Find all posts by this user
Quote this message in a reply
03-30-2023, 12:37 AM (This post was last modified: 03-30-2023 01:40 AM by TallKey.)
Post: #9
RE: ChatGPT for programming
(03-29-2023 10:44 PM)Valentin Albillo Wrote:  .
Mentioned in The New York Times today:

Open Letter signed by Elon Musk & 1,000+ other people

V.

Rewind your clocks to 2003/2004...

My rewrite of paragraph 1:
Social Media can pose profound risks to society and humanity, as shown by extensive research and acknowledged by society. Social Media could represent a profound change in the history of life on Earth, and should be planned for and managed with commensurate care and resources. Unfortunately, this level of planning and management is not happening, even though last decades have seen these companies locked in an out-of-control race to develop and deploy ever more powerful echo chambers – not even their creators – can understand, predict, or reliably control.


*2003/2004 are the launches of Myspace and FB.
Find all posts by this user
Quote this message in a reply
03-30-2023, 01:19 AM
Post: #10
RE: ChatGPT for programming
Brings to mind the Fail Safe movie/book (Henry Fonda as U.S. president in the movie, JFK implied in the book) from 1964 (less than 2 years after the Cuban Missile Crisis) where the Soviet computer came up with the conclusion that it was an appropriate time to interfere with the Fail Safe communication hardware/system. The increasingly complex, multi-layered system was looked upon as having the inevitable failure. Scary then, scary now…
Find all posts by this user
Quote this message in a reply
03-30-2023, 02:49 AM
Post: #11
RE: ChatGPT for programming
.
This interesting 51-page PDF document is also quite on topic:

ChatGPT for programming numerical methods
    "ChatGPT is a large language model recently released by the OpenAI company. In this technical report, we explore for the first time the capability of ChatGPT for programming numerical algorithms. Specifically, we examine the capability of GhatGPT for generating codes for numerical algorithms in different programming languages, for debugging and improving written codes by users, for completing missed parts of numerical codes, rewriting available codes in other programming languages, and for parallelizing serial codes.

    Additionally, we assess if ChatGPT can recognize if given codes are written by humans or machines. To reach this goal, we consider a variety of mathematical problems such as the Poisson equation, the diffusion equation, the incompressible Navier-Stokes equations, compressible inviscid flow, eigenvalue problems, solving linear systems of equations, storing sparse matrices, etc. Furthermore, we exemplify scientific machine learning such as physics-informed neural networks and convolutional neural networks with applications to computational physics. Through these examples, we investigate the successes, failures, and challenges of ChatGPT.

    Examples of failures are producing singular matrices, operations on arrays with incompatible sizes, programming interruption for relatively long codes, etc. Our outcomes suggest that ChatGPT can successfully program numerical algorithms in different programming languages, but certain limitations and challenges exist that require further improvement of this machine learning model."
I found quite interesting point 9.3 Including unknown or unwanted libraries, where the A.I. "hallucinates" nonexistent (or at least unknown to exist) libraries and includes/references them in the code generated to solve some given task. The resulting code doesn't compile/run, of course.

V.

  
All My Articles & other Materials here:  Valentin Albillo's HP Collection
 
Visit this user's website Find all posts by this user
Quote this message in a reply
03-30-2023, 09:09 AM (This post was last modified: 03-30-2023 09:10 AM by pier4r.)
Post: #12
RE: ChatGPT for programming
Thank you valentin! Nice point TallKey.

I also asked a little questions.

Quote:I want to make a function - pseudocode - that picks randomly between the numbers 0 to 7 and 20 to 23. How would you do it?

Here's an example pseudocode function that randomly picks a number between 0 to 7 and 20 to 23:

Code:
function random_number():
    // Generate a random integer between 0 and 1
    random_bit = random_integer(0, 1)
    
    // If the random bit is 0, pick a random integer between 0 and 7
    if random_bit == 0:
        return random_integer(0, 7)
    
    // If the random bit is 1, pick a random integer between 20 and 23
    else:
        return random_integer(20, 23)

...further explanations...


I'd like an uniform distribution and that what you suggest, although working, is not uniform.

I apologize for the confusion. You're right that the previous approach I suggested doesn't result in a uniform distribution, as it will pick numbers from the two ranges with equal probability but not uniformly within each range.

Here's a revised pseudocode function that should pick a random number uniformly from the given ranges:

Code:
function random_number():
    // Pick a random number from one of the two ranges with equal probability
    range_bit = random_integer(0, 1)
    
    // If the range_bit is 0, pick a random number between 0 and 7 inclusive
    if range_bit == 0:
        return random_integer(0, 7)
    
    // Otherwise, pick a random number between 20 and 23 inclusive
    else:
        return random_integer(20, 23)

Then it continues even if I say that the chat identified the problem but didn't solve it in the implementation. The interesting part is that the Issue is indeed recognized.

Wikis are great, Contribute :)
Find all posts by this user
Quote this message in a reply
03-30-2023, 09:23 AM (This post was last modified: 03-30-2023 09:25 AM by Roberto Volpi.)
Post: #13
RE: ChatGPT for programming
(03-30-2023 09:09 AM)pier4r Wrote:  Thank you valentin! Nice point TallKey.

I also asked a little questions.

Quote:I want to make a function - pseudocode - that picks randomly between the numbers 0 to 7 and 20 to 23. How would you do it?

Here's an example pseudocode function that randomly picks a number between 0 to 7 and 20 to 23:

Code:
function random_number():
    // Generate a random integer between 0 and 1
    random_bit = random_integer(0, 1)
    
    // If the random bit is 0, pick a random integer between 0 and 7
    if random_bit == 0:
        return random_integer(0, 7)
    
    // If the random bit is 1, pick a random integer between 20 and 23
    else:
        return random_integer(20, 23)

...further explanations...


I'd like an uniform distribution and that what you suggest, although working, is not uniform.

I apologize for the confusion. You're right that the previous approach I suggested doesn't result in a uniform distribution, as it will pick numbers from the two ranges with equal probability but not uniformly within each range.

Here's a revised pseudocode function that should pick a random number uniformly from the given ranges:

Code:
function random_number():
    // Pick a random number from one of the two ranges with equal probability
    range_bit = random_integer(0, 1)
    
    // If the range_bit is 0, pick a random number between 0 and 7 inclusive
    if range_bit == 0:
        return random_integer(0, 7)
    
    // Otherwise, pick a random number between 20 and 23 inclusive
    else:
        return random_integer(20, 23)

Then it continues even if I say that the chat identified the problem but didn't solve it in the implementation. The interesting part is that the Issue is indeed recognized.

Identifying a problem (whatever it means) is not a great achievement.

It is important to take into account that ChatGTP is, in a nutshell, a probabilist model to chain words, which meaning is given by the reader.

It remember me of a monkey that has been taught with signs language: the monkey can learn sooner or later to master those signs meaning “give me a banana” or similar, but that’s it.



Put a calculator into your life!
Visit this user's website Find all posts by this user
Quote this message in a reply
03-30-2023, 09:32 AM
Post: #14
RE: ChatGPT for programming
(03-30-2023 09:23 AM)Roberto Volpi Wrote:  It is important to take into account that ChatGTP is, in a nutshell, a probabilist model to chain words, which meaning is given by the reader.

It remember me of a monkey that has been taught with signs language: the monkey can learn sooner or later to master those signs meaning “give me a banana” or similar, but that’s it.

yes this is known, see for example: https://en.wikipedia.org/wiki/Chinese_room

Anyway I find what you describe a bit reductive. Imagine a database with all the answers that you need (in chess for example, a tablebase for endgames). They aren't intelligent but they provide answers anyway.

My point was not on the "uh they are intelligent" (they may be according to some definition), rather "they identify points" (based on statistics or whatever other approach) and provide useful answers.

Like calculators, they aren't smart, but they can solve problems. If the number of problem solved grows and grows, then the utility of a solution can skyrocket. The very fact that the language is well understood (despite my broken english) is incredible.

For me GPT is incredible in languages (I use it to correct me!), less so in more technical stuff.

Wikis are great, Contribute :)
Find all posts by this user
Quote this message in a reply
03-31-2023, 01:26 AM (This post was last modified: 03-31-2023 02:34 AM by pascal_meheut.)
Post: #15
RE: ChatGPT for programming
(03-30-2023 09:23 AM)Roberto Volpi Wrote:  Identifying a problem (whatever it means) is not a great achievement.

It is important to take into account that ChatGTP is, in a nutshell, a probabilist model to chain words, which meaning is given by the reader.

It remember me of a monkey that has been taught with signs language: the monkey can learn sooner or later to master those signs meaning “give me a banana” or similar, but that’s it.

It seems to me you are wrong and misinformed.

Such IAs have already been evaluated as tools to improve professional programmers productivity and the results are beyond impressive and never seen before.
As most of the critics following you line of thought, you never define the very generic words you use to despise IA performance.

Here is what a monkey cannot do:
- help me code something I do not know when using a new API. ChatGPT did it last week and btw, I have been coding for the last 40 years, still do it professionally every day. And maybe you used some of my code if you ever played with the WP-34c emulators

- solve "is there n and m integers where 2^n = 10^m" and give the exact answer

- answer IT questions asked on a forum (tried that week too to help a newbie who had problem with multiple Python versions on Linux)

- write the history of MRI and when asked, give more details about a specific period

- behave like an interviewer to help someone train when applying to a job

- pass misc exams such as LSAT, SAT math, AP Biology, AP Calculus BC, Uniform Bar Exam, Graduate Record Examination. Can you do so?

And so many other things including "speak many, many languages, both human and programming".

But you can still think that you are very smart and despise IA. You would be very disappointed when you discover that except for writing RPN code, their performance is better than yours in so many fields.
Find all posts by this user
Quote this message in a reply
03-31-2023, 03:36 AM
Post: #16
RE: ChatGPT for programming
(03-31-2023 01:26 AM)pascal_meheut Wrote:  
(03-30-2023 09:23 AM)Roberto Volpi Wrote:  Identifying a problem (whatever it means) is not a great achievement.

It is important to take into account that ChatGTP is, in a nutshell, a probabilist model to chain words, which meaning is given by the reader.

It remember me of a monkey that has been taught with signs language: the monkey can learn sooner or later to master those signs meaning “give me a banana” or similar, but that’s it.

It seems to me you are wrong and misinformed.

Such IAs have already been evaluated as tools to improve professional programmers productivity and the results are beyond impressive and never seen before.
As most of the critics following you line of thought, you never define the very generic words you use to despise IA performance.

Here is what a monkey cannot do:
- help me code something I do not know when using a new API. ChatGPT did it last week and btw, I have been coding for the last 40 years, still do it professionally every day. And maybe you used some of my code if you ever played with the WP-34c emulators

- solve "is there n and m integers where 2^n = 10^m" and give the exact answer

- answer IT questions asked on a forum (tried that week too to help a newbie who had problem with multiple Python versions on Linux)

- write the history of MRI and when asked, give more details about a specific period

- behave like an interviewer to help someone train when applying to a job

- pass misc exams such as LSAT, SAT math, AP Biology, AP Calculus BC, Uniform Bar Exam, Graduate Record Examination. Can you do so?

And so many other things including "speak many, many languages, both human and programming".

But you can still think that you are very smart and despise IA. You would be very disappointed when you discover that except for writing RPN code, their performance is better than yours in so many fields.

Well, in this case I can't wait to see those new prodigious programmes compiled by ChatGTP for HP35S in this forum.



Put a calculator into your life!
Visit this user's website Find all posts by this user
Quote this message in a reply
03-31-2023, 03:49 AM
Post: #17
RE: ChatGPT for programming
I do wonder how well it would recreate the HP 35 firmware.
All the data is in the model, so nothing novel is required.


Pauli
Find all posts by this user
Quote this message in a reply
03-31-2023, 05:54 AM
Post: #18
RE: ChatGPT for programming
(03-31-2023 03:49 AM)Paul Dale Wrote:  I do wonder how well it would recreate the HP 35 firmware.
All the data is in the model, so nothing novel is required.

It would not. It can do impressive things and be a very useful tool to increase our productivity.
But it cannot do everything of course.

This is a new tool and the smart thing to do is to learn what it can do, what it cannot, how to use it best.
It takes some time to do so, an open mind, to read about it, etc.

One can also reject it because it is new and explain why this is either useless or dangerous like people in the past did with writings, the Gutenberg printer, trains, cars, pencils, pocket calculators, computers, Internet, smartphone and many other things.
Find all posts by this user
Quote this message in a reply
03-31-2023, 09:49 AM
Post: #19
RE: ChatGPT for programming
(03-31-2023 01:26 AM)pascal_meheut Wrote:  - solve "is there n and m integers where 2^n = 10^m" and give the exact answer

I was intrigued by this and asked, likely my version is a bit more trolling that yours (I use the free account, I have the playground account of the API but I don't want to go there every time)

Quote:is there n and m integers where 2^n = 10^m ?

No, there are no integers n and m that satisfy the equation 2^n = 10^m.

To see why, note that 10 is not divisible by 2. In other words, there is no integer k such that 2k = 10. Therefore, we cannot express 10^m as a product of only 2's. However, 2^n is a product of only 2's, so it is impossible for 2^n to be equal to 10^m.

In fact, we can prove that the only common factors of 2^n and 10^m are 1 and 2. To see this, note that any factor of 10^m must contain at least one factor of 5, while any factor of 2^n must contain only factors of 2. Since 2 and 5 are relatively prime, the only common factors of 2^n and 10^m are 1 and 2.

Therefore, there is no integer n and m that satisfies the equation 2^n = 10^m.

I agree with the rest of the point that one should see the strong points of the solution and use those. For example natural language as I already said.
The Interview mockup is also interesting.

Wikis are great, Contribute :)
Find all posts by this user
Quote this message in a reply
03-31-2023, 09:55 AM
Post: #20
RE: ChatGPT for programming
[quote='pier4r' pid='170800' dateline='1680256184']
I was intrigued by this and asked, likely my version is a bit more trolling that yours (I use the free account, I have the playground account of the API but I don't want to go there every time)
[quote]
ChatGPT is not very good with such problems.
GPT-4 is much better and it is the one I quoted.
Find all posts by this user
Quote this message in a reply
Post Reply 




User(s) browsing this thread: 1 Guest(s)