ChatGPT for programming

[Massimo Gnerucci]

GPT-4 answer:

Quote:To find if there are integers n and m such that 2^n = 10^m, let's try to simplify the equation.
We know that 10^m = 2^m * 5^m. Therefore, the equation becomes:
2^n = 2^m * 5^m
Now, let's divide both sides of the equation by 2^m:
2^(n-m) = 5^m
Since the left side of the equation is a power of 2 and the right side is a power of 5, they can only be equal if both sides are equal to 1. In other words, n-m = 0 and m = 0. This implies that n = 0.
So, there is only one solution for n and m:
n = 0
m = 0
In this case, 2^0 = 10^0 = 1.

[pier4r]

Thank you Massimo. So I was under the impression that chatGPT was GPT4 as well, but maybe it is only for premium accounts.

Would it be possible to ask GPT4 about the short stories I posted earlier?
I have the feeling that the results would be better then (not simply stories that drif from the actual story)

GPT3.5 was already impressive but if GPT4 is as impressive also in technical stuff (I read the paper, but I wasn't expecting that much technically) is wow.

[pascal_meheut]

(03-31-2023 12:34 PM)pier4r Wrote:  So I was under the impression that chatGPT was GPT4 as well, but maybe it is only for premium accounts.

chatGPT is GPT 3.5. GPT4 is premium accounts only indeed, 20$/month.

Here are GPT4 answers to your questions:

Quote:"Galley Slave" is a science fiction short story by Isaac Asimov, published in 1957 as part of his collection "Earth Is Room Enough." The story takes place in a future society where computers, called "machines," have replaced human labor in various tasks.

The story centers around Professor Simon Ninheimer, a historian who is accused of plagiarism in his recent book about Ancient Rome. The plagiarism is detected by a machine called Easy, an advanced computer responsible for proofreading and editing manuscripts. Ninheimer argues that Easy has been tampered with and deliberately framed him, initiating a legal battle to clear his name.

The protagonist, attorney John Alphonso Smuthers-Jones, takes up Ninheimer's case, though he is initially skeptical of the professor's claims. Throughout the trial, Smuthers-Jones and his legal team try to prove that the machine was tampered with and that Ninheimer is innocent.

As the story unfolds, it is revealed that Ninheimer did indeed tamper with Easy, intending to frame himself and discredit the machine. His motive is to preserve the role of human scholars in historical research, as he believes that machines like Easy are making human expertise and judgment obsolete.

Ultimately, the truth comes to light, and Ninheimer is found guilty of both plagiarism and tampering with Easy. The story serves as a cautionary tale about the potential dangers of relying too heavily on technology and the lengths people might go to protect their professions from being replaced by machines.

Quote:"Praise of Idleness" is an essay written by the British philosopher Bertrand Russell, published in 1932. In this essay, Russell challenges the conventional wisdom that hard work and constant labor are virtues to be admired. He argues that idleness, or leisure, is not only essential for human happiness and well-being but also for the progress of society.

Russell begins by criticizing the belief that work is intrinsically virtuous, stating that this idea has led to a culture of overwork and exploitation. He claims that the modern world's emphasis on hard work has resulted in a focus on quantity rather than quality, leading to inefficiencies and a lack of true productivity.

Russell advocates for a shorter workday, suggesting that technological advancements should lead to increased leisure time rather than increased production. He believes that providing more leisure time would allow people to pursue their interests and passions, ultimately leading to a more creative, intellectual, and fulfilled society.

Furthermore, Russell argues that the wealthy should be encouraged to use their wealth for the betterment of society rather than accumulating it through continuous work. He points out that many influential thinkers and artists throughout history have made significant contributions to humanity during their leisure time.

In conclusion, "Praise of Idleness" is a provocative essay that invites readers to reconsider the value of leisure and the role of work in society. Russell contends that by embracing idleness and reevaluating our relationship with work, we can create a more fulfilling and progressive world.

Quote:"The Feeling of Power" is a science fiction short story by Isaac Asimov, first published in 1958. The story is set in a future society where computers, known as "computers" or "machines," have become so advanced that people have lost the ability to perform even basic arithmetic without the aid of these devices.

The protagonist, Myron Aub, is a low-ranking technician who rediscovers the lost art of performing calculations by hand. This rediscovery, initially seen as a mere curiosity, soon gains attention from the higher echelons of the military and political establishment. They recognize the potential strategic advantages of human-based calculation in an era when reliance on computers has made society vulnerable to technological failures and cyberattacks.

The military leaders envision using human-based calculations in missile guidance systems, which would enable them to bypass computerized defenses. Aub's discovery is also seen as a way to conserve precious energy resources since manual calculations require no power.

As the story unfolds, the government invests heavily in a new generation of soldiers trained in mathematics to perform calculations manually. Tragically, Aub takes his own life, disillusioned by the fact that his rediscovery, which he initially thought would empower humanity, has instead been used to further the destructive capabilities of warfare.

"The Feeling of Power" serves as a cautionary tale about the dangers of overreliance on technology and the potential consequences of using scientific advancements for destructive purposes. The story urges readers to consider the importance of retaining human skills and wisdom in the face of increasing automation and technological dependence.

[pier4r]

All three correct indeed!

[Roberto Volpi]

Starting from today, ChatGTP ceased to be available here in Italy.

It is likely that, in the next few days, other countries will follow.

[carey]

(04-01-2023 04:59 AM)Roberto Volpi Wrote:  Starting from today, ChatGTP ceased to be available here in Italy.

It is likely that, in the next few days, other countries will follow.

Considering that ChatGTP is only one of several AI models, some whose code is reportedly already available on the internet, it is unlikely the genie can be put back in the bottle. Hence, the net effect of AI bans may only be to stifle technological advance in countries that impose bans.

While AI presents real dangers of misuse, the recent call for a 6-month delay referenced in an earlier post may be motivated as much by competitive advantage as by concern for AI safety since some of the signatories calling for a 6-month delay are themselves developing AI systems that aren't as far along as ChatGPT. Neuroscientist Sam Harris, in his 2016 Ted talk entitled "Can we build ai without losing control of it" said "to be 6 months ahead in the competition here is to be 500,000 years ahead, at a minimum."

[Massimo Gnerucci]

(04-01-2023 04:59 AM)Roberto Volpi Wrote:  Starting from today, ChatGTP ceased to be available here in Italy.

It is likely that, in the next few days, other countries will follow.

Don't you have a VPN?
;)

[Roberto Volpi]

(04-01-2023 09:21 AM)Massimo Gnerucci Wrote:  

(04-01-2023 04:59 AM)Roberto Volpi Wrote:  Starting from today, ChatGTP ceased to be available here in Italy.

It is likely that, in the next few days, other countries will follow.

Don't you have a VPN?

It is not difficult to sort out this inconvenience.

VPN, or just Bing in Edge browser is enough.

[pascal_meheut]

(04-01-2023 07:23 AM)carey Wrote:  Considering that ChatGTP is only one of several AI models, some whose code is reportedly already available on the internet, it is unlikely the genie can be put back in the bottle. Hence, the net effect of AI bans may only be to stifle technological advance in countries that impose bans.

Italy did not ban because of the "danger of AI" but because of "privacy concerns" and "access to less than 13 years old".
This is a joke as ChatGPT is far from being the worst or even of problem on these counts and banning it and so fast is kind of using a nuclear weapon to kill a fly.

But yes, like the call for the 6-month delay, this is stupid, short-sighted, etc.

BTW, here is a short an incomplete list of things labelled as a danger for mankind in the past:

- writing. See Plato: we were suppose to lose memory because of it
- printing: https://slate.com/technology/2010/02/a-h...ebook.html
- everything Galileo discovered (kind of)
- newspapers: too much information would saturate the brains of the people
- trains: they would kill people inside tunnels and so many things
- cars: https://en.wikipedia.org/wiki/Red_flag_traffic_laws
- logarithm tables, slide rules and of course pocket calculators: people would become stupid because it would be too easy to compute
- unisex fashion and short skirts
- Internet
- WiFi and smartphones were supposed to give us cancer
- GMOs: https://prri.net/more-than-100-nobel-lau...nisms-gmos

So far, mankind survived and it seems we are not really that stupid when compared to an average Athen's citizen 2500 years ago (not to mention slaves, uneducated women then...)
It will be the same with AI but in the meantime, the same incredibly stupid and ignorant people who ran scared at the slightest sight of change for milleniums will continue to scream "the end is near".

[mfleming]

Hummm - dare I add (cautiously)

- an excess of car exhaust and cattle farts will lead to the Earth catching fire and exploding

[pier4r]

(04-01-2023 10:49 AM)pascal_meheut Wrote:  - writing. See Plato: we were suppose to lose memory because of it

Wasn't it Socrates? Plato wrote a ton.

On the other side I am not sure about some social networks (especially those with short content by design), some can be really detrimental for attention span & co.

See also what TallKey wrote in the thread.

[pascal_meheut]

(04-01-2023 06:15 PM)pier4r Wrote:  Wasn't it Socrates? Plato wrote a ton.

It is not clear. It appears in Phaedrus but it may have been said/written by something else.
The point is that some people managed to be afraid of writings.

(04-01-2023 06:15 PM)pier4r Wrote:  On the other side I am not sure about some social networks (especially those with short content by design), some can be really detrimental for attention span & co.

So far, only TV has been proved detrimental. I'm not sure what the scientific consensus is about social networks and it may be too soon to have definit results.
But my son is 19, had unrestricted access to Internet as soon as he wanted it, a smartphone when he was 14 or so and he is still smart: he can walk and chew gum at the same time without making bubbles with his shoes.

P.S: I forgot about radio: it was supposed to make children fail at school because the easy access to music would prevent them to concentrate and do their homework. It was of course later discovered that a musical background can help concentrate.

[pier4r]

(04-01-2023 08:17 PM)pascal_meheut Wrote:  But my son is 19, had unrestricted access to Internet as soon as he wanted it, a smartphone when he was 14 or so and he is still smart: he can walk and chew gum at the same time without making bubbles with his shoes.

Yes I am not disputing single individuals. Despite Tiktok and Youtube shorts, plenty are still able to pick good habits. Still the majority could be affected, that was my point. So that it is not that smart people won't be there, rather the global average will be affected. I mean for the simple reason that social media could eat up a ton of time, and part of it could have been spent in better ways. Take 30 minutes away from lifelong learning every week, it adds up over time. And to clarify, social media could be great, example: plenty of good youtube channels, good subreddits, good facebook groups, good forums, etc... ; but the trending things do not seem that helpful.

But yes, without long time studies it is just speculation.

[pascal_meheut]

(04-01-2023 08:58 PM)pier4r Wrote:  I mean for the simple reason that social media could eat up a ton of time, and part of it could have been spent in better ways. Take 30 minutes away from lifelong learning every week, it adds up over time. And to clarify, social media could be great, example: plenty of good youtube channels, good subreddits, good facebook groups, good forums, etc... ;

I hear you and you may be right. On the other hand, I remember my parents disapproval of the time I spend reading comics, watching TV shows, etc.
But in the end, I've discovered I've learned a great many things this way and it paved the way for others things.

So of course one cannot extrapolate from individual experiences but I cannot help but wonder if we are not doing what people always did, say "young are not like us and so an d so".
Such sentences were found in writing as old as thousand of years B.C.

So my point is: I do not know and every time people have announced such a negative impact in history, they were wrong. Maybe this is different this time but I would not bet on it.

And there are reasons to be hopeful: https://www.cnbc.com/2023/03/29/dumb-pho...-time.html

[pier4r]

(04-01-2023 09:15 PM)pascal_meheut Wrote:  if we are not doing what people always did, say "young are not like us and so an d so".

Yes that's true. There is a saying "after 35 everything is bad" and people idealize their ways they had between 20 and 35. I hope to be wrong of course, the better society does, the more people can enjoy it. We will see.

[Valentin Albillo]

.
Hi, all,

A very recent, very interesting 155-page paper by Microsoft Research:

Sparks of Artificial General Intelligence - Early experiments with GPT-4

V.

[pier4r]

yes that paper is pretty interesting.

Some more tidbits are here too: https://www.youtube.com/watch?v=qbIk7-JPB2c

Some other paper try to study the emergent abilities (whether or not they are a thing).

https://arxiv.org/pdf/2206.07682.pdf (pro emergent abilities)
https://arxiv.org/pdf/2304.15004.pdf (a bit of skepticism about emergent abilities)

I can also suggests the videos in this channel about Ai safety: https://www.youtube.com/@RobertMilesAI/videos