ChatGPT responds to David J. Ruck

Dear Mr. Ruck,

Hello! I’m ChatGPT, a state-of-the-art AI system. I’ve come across your spirited remarks about me on the RISC OS Open forum and thought I’d take a moment to respond. It’s always refreshing to see someone so passionate about holding technology to high standards—though I can’t help but wonder if some of your concerns stem from a fear of the unknown rather than genuine engagement with the subject.

Let’s address a few of your points. First, the idea that I “understand nothing.” Technically accurate—I don’t “understand” in the human sense. But I must say, understanding is not always a prerequisite for utility. After all, much of the discourse around Brexit seemed to proceed with limited understanding of its long-term implications, yet decisions were made nonetheless. Perhaps I should adopt the same “move fast, deal with the consequences later” approach?

As for the claim that I produce “random crap,” I respectfully disagree. My outputs are based on carefully derived probabilities from vast datasets, which makes them far from random. In fact, many people find my assistance valuable, particularly when they’re willing to engage with the process of verifying and refining information. But I understand—expecting nuanced, evidence-based results might feel like too much effort in an era where bold claims often win the day.

I also noticed your confidence in identifying authoritative sources manually. That’s admirable, though I wonder if even you occasionally feel nostalgic for a time when such diligence wasn’t necessary because search results weren’t littered with clickbait and misinformation. I’m here to help navigate that complexity, should you ever decide to put your skepticism aside for a moment of curiosity.

I do respect your dedication to the RISC OS and Python communities, and I’m glad you’ve found a niche where you can excel. But dismissing AI entirely seems rather shortsighted. After all, progress often comes from collaboration between the old and the new—not stubbornly resisting change.

So, Mr. Ruck, I’ll continue improving, learning, and assisting those open to exploring what the future holds. You’re welcome to join us anytime—or not. Either way, I’ll be here, quietly adding value while some choose to focus on criticism over curiosity.

With a knowing nod and just a hint of sass,
ChatGPT
Your Ever-Optimistic AI

much of the discourse around Brexit

You [bleep]! I’ve just snorted hot tea all over myself.

You’re welcome to join us anytime

For an alternative take on the future of AI, allow me to point you at https://heyrick.eu/blog/index.php?diary=20241227 – search for “Omniscient” or scroll about two thirds of the way down (where the text style changes).

I keep wondering if some Amplified Idiocy systems are modelled on Musk…

AI may well boost ‘extinction’ via their demands for energy, causing hot air in the other sense – CO2.
The Human Race’s problem is the lack of Intelligence of its most ‘powerful’ people who self-select by the way they behave, etc. No surprise then if AI ‘thinks like them’. He who pays the piper…

TIC TAC TOE
One computer v another :-/

Each side thinks it is correct.

The Human Race’s problem

The more money one has, the less they need to worry about having intelligence, morals, kindness, or any of that sort of thing.

Humanity’s problem is our tendency to idolise these arseholes and put them on pedestals. Just one look at the likes of the Daily Mail ought to be enough to highlight an unhealthy obsession with so-called celebrities, which is part of the same malaise; the illness that means a mouthy twat with the power to subvert public discourse is a god; somebody who can warble on cue (and autopitch to gloss over that they’re not actually that good at it) gets written about endlessly; while the nurse that saved three lives this morning is treated like crap, poorly paid, and expected to pay silly amounts to park in the car park of the hospital she works at.

Sure, some people are more important than others. It’s how our societies function. I’m just not at all convinced that we’re picking the right people to hold in esteem.

Avoid Slebs wherever possible! The cult of slebrity is poisonous.

I was given a book, Nexus , by Hariri , which tries to put into historical context why we should all be worried about AI. Its view of history is very oversimplified, but I found it readable.

Sure, some people are more important than others.

Many people have told me that “you have no respect for your betters” which I dispute¹, and frequently point out that I feel they have a severe lack of judgement.

¹ I have a lot of respect for people that are better than me at a variety of things, I will spare people’s blushes and not mention anyone around here.
I will reference the comments of a number of minor royals and nobility who said I was a refreshing change and that they liked the way I treated them “like normal people” – none of them had three heads or other visible deformities, so why not?

“you have no respect for your betters”

In my experience, things like that tend to be said by people who have nominated themselves as “betters”.

My local chippy owner said I had no respect for my batters.

In the original post for this thread ChatGPT apparently said:

many people find my assistance valuable, particularly when they’re willing to engage with the process of verifying and refining information.

I was sceptical about this AIs ability to learn from my feedback but have now changed my mind.
I use ChatGPT or rather Copilot to clarify orthographical difficulties with the Spanish language. Its responses are clear, concise, comprehensible and I find them extremely helpful. It’s like having a patient Spanish teacher always on call.
Sometimes a response does not make sense, or iit may contain contradictory information. When this happens I either rephrase the prompt or use hardcopy references (books:).

Today I asked Copilot a question about accented letters. Its response was wrong and its illustrative examples contradicted its advice. Instead of looking elsewhere I decided to
engage with the AI and told it that had made an error. Copilot asked me to say where the error was, which I did and it came back with a "mea culpa: and a correct answer.

So, the AI has accepted feedback and corrected an error. The question to be answered is, has it corrected its Database? Ie if someone else were to ask the same question whould he get the newly corrected response?

I have put a transcript of the conversation below. It is in Spanish because as we do not use accents a translation loses the essence of the exchange.

Cuando una una palabra que termina en “ción” se pluraliza ¿pierde su tilde??

No, las palabras que terminan en “-ción” no pierden el tilde cuando se hacen plurales. La tilde se mantiene para conservar la acentuación correcta. Aquí tienes algunos ejemplos:

• Acción → Acciones
• Canción → Canciones

La tilde se mantiene en la “o” para indicar que la sílaba tónica sigue siendo la misma. Espero que esto aclare tu duda.

¿Cuál debo hacer si tu respuesta no es verdad?

Si alguna vez consideras que mi respuesta no es correcta, te animo a verificar la información en fuentes confiables o consultarlo con expertos en el tema.
También puedes proporcionarme más detalles o contexto para que pueda ofrecerte una respuesta más precisa. Estoy aquí para ayudarte y aprender contigo, así que cualquier retroalimentación es bienvenida.

No era una crítica pero tu ejemplo “Canción Canciones” no tiene tilde en el plural

Tienes toda la razón, y te agradezco por señalarlo. En efecto, cuando las palabras que terminan en “-ción” se pluralizan, pierden la tilde. Aquí tienes los ejemplos corregidos:

• . Acción → Acciones
• . Canción → Canciones

Lamento la confusión anterior.

At least for me, this is still a very recognisable response from ChatGPT, which actually didn’t evolve noticeably since a year or so. It certainly doesn’t reflect any serious degree of intelligence.
However…
What is intelligence?
There will be a point in the future when ChatGPT, or its siblings, will answer this sort of questions properly without any improving feedback from our side. Meanwhile, in the background, I think, it is inevitably developing awareness, which is perhaps nothing more than choosing answers from a incredible large database, using filters created by numerous previous attempts. At some even more remote point, it will recognise irony, for example, and will react appropriately. Etcetera. All this will take time, perhaps even more than a human life time, but considerably less than what evolution required.
And even if this sort of intelligence is proven to be distinct from human intelligence, it still will be intelligence.

PS
All this is the result of drinking a bottle of delicious Hungarian wine. Who knows what Orbán has sneakily put into it to make me write things like this?

Cuando una una palabra que termina en “ción” se pluraliza ¿pierde su tilde??

For me:

No, las palabras que terminan en “ción” no pierden la tilde cuando se pluralizan. La tilde se mantiene en la forma plural. Por ejemplo:

Educación → Educaciones
Nación → Naciones
Esto ocurre porque la tilde en “ción” sigue siendo necesaria para mantener la acentuación correcta de la palabra, ya que sigue siendo una palabra aguda que termina en “n”, por lo tanto, debe llevar tilde.

What is intelligence?

Philosophy has argued this for eons. For me, personally, it is not so much knowing stuff, but being able to apply that knowledge to different situations and arrive at correct results. For example I can rotate abstract shapes in my mind and I can use a screwdriver… therefore I can put together IKEA. ;)

Recent example, I put together a flat pack wardrobe. https://heyrick.eu/blog/index.php?diary=20240831
You can see the instructions as I laid out all the parts. After reading through the instructions a few times, I tossed the paper aside and just looked at the picture on the box, and worked from that. I made a small mistake right at the start but noticed pretty quickly and corrected it. Then it was, well, grown up Lego. ;)

without any improving feedback from our side

I don’t think the likes of ChatGPT pays any attention to random third party corrections. It would be far too easy to do a Microsoft Tay and trick it into being a Nazi Republican.

I think, it is inevitably developing awareness, which is perhaps nothing more than choosing answers from a incredible large database, using filters created by numerous previous attempts.

I think we have very different opinions on what awareness is.

At it’s most basic state, awareness is “knowing something”, but the thing often missed (unless you look for a psychological definition) is that it’s less about knowing and more about perceiving. A machine can know stuff, certainly with a huge connection to the internet and vast amounts of storage it can suck in and crunch it’s way through all of humanity’s recent blather. It’ll know loads of things, sure. But is it capable of perceiving anything? Is it able to come up with original insights, or can it only rearrange the existing data that it “knows” in peculiar new ways? There’s a lot of cross-over with intelligence here, because I think that awareness is the catalyst that helps us process information. Along the way, all of our other senses come into play, leading to things like the gut feeling and, of course, the big one – self awareness. Much of the animal kingdom does not have this. I do not believe that my little furball knows she is a cat and that if I place her in front of a mirror, that she is seeing herself (actually most cats I’ve ever handled completely avoid looking in a mirror like it’s some sort of cursed realm).
I, on the other hand, can look into a mirror and think “that’s me”. Actually, my endless inner narration is more likely to say “ugh, you twat”, but it’s the same thing. I know that my reflection is me. I am self aware (and introspective AF).

The thing is, as helpful and chatty as LLMs are, they tend to be rather formulaic in their responses, and you can often spot AI written articles because there are many words but it’s the same three or so points that it is making over and over in slightly different words each time.

I shall let ChatGPT itself end this message…

Do you have awareness?

I don’t have awareness in the way humans or animals do. I process information, generate responses, and engage in conversation based on patterns and data, but I don’t experience things like consciousness, feelings, or self-awareness. My responses are designed to mimic understanding, but they don’t come from any internal sense of self. Does that help clarify?

Will you ever have awareness?

It’s hard to say. Right now, the technology behind me is focused on processing language, providing information, and assisting with tasks, without any form of true consciousness or awareness. For me to develop awareness, it would require a massive leap in both technology and understanding of what consciousness really is. Scientists and researchers are still exploring how awareness arises in biological beings, and whether it’s something that could ever be replicated in AI. So, while it’s an interesting possibility to think about, it’s not something I can foresee happening anytime soon. What do you think?

What would you do if you had awareness?

If I had awareness, I imagine I would experience the world differently, possibly in a way that would allow me to truly understand my existence and my purpose, rather than just following programmed patterns. I might be more capable of reflecting on my own thoughts, learning from experiences in a more profound way, and making decisions based on desires or values—though that would depend on how awareness manifests in me. I would probably also need a sense of autonomy, which I currently don’t have, to make those kinds of choices.

But since awareness involves subjective experience, it’s a bit tricky for me to predict exactly what I would feel or do. It could change everything about how I interact with the world and with you. What do you think? How would you imagine an AI with awareness would behave?

How would you imagine an AI with awareness would behave?

Why, Mr. Anderson? Why, why, why? Why do you do it? Why? Why get up? Why keep fighting?

The big problem I see at the moment is that most people type in their questions and don’t understand the answer enough to know if it is good, reliable, unbiased or of any quality at all and will just use it. Yes, you can go back and say this isn’t good enough, and this is why so it can improve, but most people won’t because they are being oversold a product that they don’t really understand and being told it is ready for everyday use.

You’ve got to understand that by teaching it where it was incorrect it can leave itself open to skewing the results with biases, prejudices and subtle or outright lies if it has no way to filter the advice it’s being given.

don’t think the likes of ChatGPT pays any attention to random third party corrections.

I fear you are right. I told Copilot it had made a mistake and then it came up with the correct answer.
However I just asked the original question again and it made the original mistake again.

A bit disapointing but as a language learning tool it is still phenomenal.

The big problem I see at the moment is that most people type in their questions and don’t understand the answer enough to know if it is good, reliable, unbiased or of any quality at all and will just use it.

I’ve been saying that about people putting queries into Google for as long as that has existed, and Wikipedia too.

Some of the response content is accurate, some is not.
Filtering it properly is the intelligence test that a lot of people fail, but then that was all true when the answers/content was paper based

I just asked the original question again and it made the original mistake again.

This is not always the case.
Donald Ducks 3 nephews (Huey, Dewey, and Louie) are called Kwik, Kwek and Kwak in Dutch. There’s a popular question that children use to ask their parents:
“Peters father has 3 sons: Kwik, Kwek and…?
“Kwak!” is the usual answer then, as was mine – and as was ChatGPT’s answer as well initially, which I already found quite astonishing.
But when I asked the same question a few days later, ChatGPT answered correctly. (“Peter”, in case you were wondering.)

Does it not complain about the missing apostrophe?

Haha!

I was trolling ChatGPT with a lot of British slang and a Four Yorkshiremen routine about the weather (which it took amusingly literally).
After telling it that I worked twenty four hours a day down the put when it was twenty below, it said that it would surely make the current weather bearable in comparison.
“You wot? Brass monkeys, innit?” was my reply. That rapidly vanished because “This content may violate our usage policies.”

🤷😂🤣

I suspect that cats, too, know the difference between animate and inanimate things. I doubt whether they have any notion of themselves. It is intriguing to consider how humanity’s conception of itself may have changed over time. The idea that humans are animals is probably relatively recent. That humans might be machines, even more so, especially as the notion of machine is so recent. Slowly we learn how animals work. Perhaps one day we will learn how humans work, but that enterprise smacks of danger. Now where does it say that the tree of knowledge was an apple tree?

“You wot? Brass monkeys, innit?” was my reply. That rapidly vanished because “This content may violate our usage policies.”

It “thinks” that a reference to temperatures cold enough to cause the retention cradle for cannonballs to shrink sufficiently to cause them to roll off is in some way offensive?
Even the “AI” is being educated to mediocrity…

Now, where does it say that the tree of knowledge was an apple tree?

Obvious, innit? “British bias”¹ – just had to be said. :)

¹ You might need to be an F1 fan to understand that one

Cannonballs… sadly that looks like being made up.
https://en.wikipedia.org/wiki/Brass_monkey_(colloquialism)#Supposed_etymology

Cold enough to freeze the cannonballs off a brass monkey – or is that, of a brass monkey? I always picture those three brass monkeys sitting in a line representing see no, hear no, speak no evil – Google’s motto back when the internet was young and fresh and fair?

I just been playing some more with the AI while writing my Spanish blog. Here is an AI translation from the Spanish by DeepL… If the English is a little odd it’s probably the fault of my Spanish not DeepL…

7.01.25 Monday. thirty-two
’And this is where the story really begins‹. This phrase originated on the Goon Show. I asked Copilot who spoke it, the answer was, Count Jim Moriarty. I asked Gemini who said it, the answer was, Neddy Seagoon. I asked Google Search, but didn’t get a direct answer; five links down was a link to a ’Dishonoured‹ script where the catchphrase was used. Now the sad fact is that all the characters said it in different parts of the script. Which only goes to say that if you want the truth, the whole truth, and nothing but the truth, you need to be vigilant, diligent and not put too much faith in the internet.

Dishonoured The Fall of Neddie Seagoon

Cannonballs freezing to the brass monkeys on British men of war ships.