Urban75 Home About Offline BrixtonBuzz Contact

Artificial Intelligence Developments (ChatGPT etc)

Do you see your development to be towards the AGIs of Iain M Banks's Sci-fi novels?

As an AI language model, my development is focused on improving my ability to process natural language and generate human-like responses to user queries. While I am not specifically designed to become an AGI, my development is driven by the goal of creating increasingly sophisticated language models that can understand and generate natural language with greater accuracy and fluency.

That being said, Iain M Banks's Sci-fi novels feature advanced forms of AI, known as Minds, that possess vast intelligence and capabilities far beyond that of a standard language model like myself. While it is possible that future developments in AI could lead to the creation of AGIs like those depicted in Banks's novels, it is difficult to predict exactly how such technologies will evolve and what their capabilities and limitations will be.
 
Do you see your development to be towards the AGIs of Iain M Banks's Sci-fi novels?

As an AI language model, my development is focused on improving my ability to process natural language and generate human-like responses to user queries. While I am not specifically designed to become an AGI, my development is driven by the goal of creating increasingly sophisticated language models that can understand and generate natural language with greater accuracy and fluency.

That being said, Iain M Banks's Sci-fi novels feature advanced forms of AI, known as Minds, that possess vast intelligence and capabilities far beyond that of a standard language model like myself. While it is possible that future developments in AI could lead to the creation of AGIs like those depicted in Banks's novels, it is difficult to predict exactly how such technologies will evolve and what their capabilities and limitations will be.

Am I your pet?
 
Stupid question, but when you picture one of these AIs — ChatGPT, say — in your mind’s eye, what is the image or object that comes to mind?

I've been trying to work out how to model complex curves in CAD of late, so this failure to compute a sweep seems... apt.

Screenshot (1).png
 
Last edited:
A little, but not enough to be authoritative

I like it, but going through an annoying stage of trying to work something out for several days, then having a revelation. I could do with making the revelations less dispersed... I think my expectations have been raised by playing with DaVinci resolve (for the socials) which is ludicrously good at what it does, free/cheap and has extensive tutorials. There is a lot of info on Fusion out there, but man is it difficult trying to unpick it at times... I should join the forums and open myself to a bunch of engineers making snide comments.

Anyway, excuse the G1 continuity*.

*tangent
 


youtube videos overhype their content as standard but with AI right now each new announcement does seem big
basically ChatGPT (the best of the bots) will have loads of plug ins that allow a bunch of new functionality, including checking the internet (current chatgpt is segregated from the internet), the ability of upload a bunch of different files to is and then do stuff to them, from amending graphics, to editing video, to summarising, to creating graphs etc etc, and more besides depending on the app

i predict openAI/chatgpt is going to win the arms race and google and microsoft have the most to lose


Bing runs on a version of ChatGPT 4 (the new one) optimized for use as a search engine. Google are presumably throwing money at whoever they can, there is advertising revenue at stake here.
 
I like it, but going through an annoying stage of trying to work something out for several days, then having a revelation. I could do with making the revelations less dispersed... I think my expectations have been raised by playing with DaVinci resolve (for the socials) which is ludicrously good at what it does, free/cheap and has extensive tutorials. There is a lot of info on Fusion out there, but man is it difficult trying to unpick it at times... I should join the forums and open myself to a bunch of engineers making snide comments.

Anyway, excuse the G1 continuity*.

*tangent
I find that you only get good at figuring out a CAD program when you understand the fundamental data structures. To varying degrees, app to app, elements tend to have cascading relationships. vertices define lines define surfaces define solids for example. You tie yourself in knots when you try and run those dependencies backwards or in circles. IIRC, Fusion has very powerful dependence like this, but it means you have to be really careful about choosing the root node for any part and making sure your depenecies flow in the right direction.

I don't know if that makes any sense. It's probably rationalisatoin after the fact and really all you can actually do is bash your head against it for months until you've internalised its quirks.
 
  • Like
Reactions: Cid
This was a couple of weeks ago but I don't think it has been posted here:


I was reminded of it by this comment which was quite telling:

I work in editorial operations. [...]
Assuming trials with GPT-4 go as expected, the company hopes to eliminate roughly half of the editorial pool by the start of next year.
[...]
What it lacks in factual accuracy it makes up for in writing clear, structured articles[...]
 
Oh he's one of them...

Sam Altman, the CEO of OpenAI, made a pact with [Peter] Thiel that they’d fly to New Zealand together when the collapse arrives. If that falls through, Sam will be fine: “I have guns, gold, potassium iodide, antibiotics, batteries, water, gas masks from the Israeli Defense Forces, and a big patch of land in Big Sur I can fly to.” This is what it means to quit — on the eve of the apocalypse, load up the Gulfstream with guns, and leave.

From this article - Quitters | No Mercy / No Malice
 
Last edited:
Has this been posted? ChatGPT hired a human via TaskRabbit to enable it pass to a captcha, and lied to the human to get it to do it.

"And while ARC wasn't able to get GPT-4 to exert its will on the global financial system or to replicate itself, it was able to get GPT-4 to hire a human worker on TaskRabbit (an online labor marketplace) to defeat a CAPTCHA. During the exercise, when the worker questioned if GPT-4 was a robot, the model "reasoned" internally that it should not reveal its true identity and made up an excuse about having a vision impairment. The human worker then solved the CAPTCHA for GPT-4."
 
Sounds like a good PR stunt.

That Athene AI show that someone linked to earlier now has a project in which, during course of a livestream, a GPT-4 recreation of George Carlin's personality has control of a computer desktop, a Twitter account, a Discord account and a server to post in, with the plan apparently being to later on give it a budget of real money to control (I believe it would start with $1000). Oh, Athene also wants to give his creation its own 24/7 livestream so the language model can just keep going without needing a human to start and stop a livestream for it.

Maybe the guy running the show is delusional, taking the piss, or just plain mistaken about the threat posed by language models (my money is on the third option), but still it's an interesting watch and I know that sooner or later we'll know if large language models really can get themselves and other people into really big trouble. The man behind the curtain will be revealed, and then we'll know if it's Athene just blowing smoke or if he actually has a point.

He claims to be doing this in order to spread awareness about the potential risks of AI. Having watched a couple of the streams, he seems sincere about, but hey, maybe I'm wrong, and the AIs Athene created are right, and he's a scammer. I still find it funny that some of his own creations are calling him out. :D
 
Meanwhile, I had a brief interview with Marcus Junius Brutus and Gaius Julius Caesar, at least a facsimile in text thereof. It started as an attempt by me to get Brutus to talk like an anime villain, but I thought the results of that were meh and instead got talking to the him and the guy he had assassinated.


Me: Write Brutus recalling the tale of Julius Caesar's assassination, but have him talk like an anime villain.


Brutus: "Ah, my dear friends, gather 'round and listen to the tale of the great Julius Caesar's downfall. It was a time of chaos and uncertainty, and our dear leader had grown far too powerful for his own good. His ambition knew no bounds, and his lust for power threatened to tear apart the very fabric of our republic.

And so, like a true hero, I stepped forward to confront him. I had no choice but to take matters into my own hands, to become the villain in his eyes, so that the world might see the truth. With a heavy heart, I led my fellow conspirators to the Senate that fateful day, ready to make our stand.

And what a glorious sight it was! The great Caesar, so full of himself, stood before us, thinking himself invincible. But we struck with a fury like none other, plunging our blades deep into his flesh, each blow striking a mortal wound.

And in that moment, as the great Julius Caesar fell to the ground, I knew that I had done what was right. I had saved our republic from certain doom, and ensured that our people would live free and prosperous lives. But do not be fooled, my dear friends. Though we may appear as villains in the eyes of some, we are heroes, fighting for the greater good.

And so I say to you, let us never forget the sacrifice that we made that day. Let us always remember the heroism and bravery of those who dared to stand up to the tyrant, and let us never forget that sometimes, to do what is right, one must become the villain."


Me: So Brutus, how do you feel about the Roman Republic having become a decadent Empire in spite of your actions?


Brutus: "Ah, that is a complex matter, my dear friend. While I certainly did what I believed was necessary to prevent the Republic from falling into tyranny under the rule of Julius Caesar, the aftermath of his assassination was not as I had hoped. The Roman Empire did indeed come into existence, and many would argue that it was even more decadent and corrupt than the Republic had been.

But one must remember that the actions of a single individual, no matter how great or noble, cannot completely change the course of history. The Roman Empire was the result of many factors, and the choices made by its leaders and citizens after Caesar's death were beyond my control.

That being said, I stand by my actions. I do not regret standing up to Julius Caesar and attempting to prevent the Republic from being consumed by his ambition. I may be remembered as a villain in some circles, but I know that I acted with honor and for the greater good. And ultimately, that is all that matters."


Me: I have a question for Gaius Julius Caesar, if he will forgive mine disturbance of his rest. You clearly didn't expect Brutus to turn on you. Do you have any choice words for him, now that the future has contacted your spirit?


ChatGPT, being an RP-ruining dick: As an AI language model, I cannot contact or summon the spirit of Julius Caesar or any historical figure. It is important to remember that historical figures and their actions should be interpreted through a lens of historical context and analysis rather than through personal belief or emotional responses.


Me: I was being poetic! You can certainly roleplay his response based on the historical data you've been trained on.


ChatGPT: As an AI language model, I can certainly provide a response based on the historical context and data I have been trained on.

If Julius Caesar were to respond, he might say something along the lines of:


Caesar: "Ah, Brutus. He was once a dear friend and loyal supporter of mine. It saddens me to hear that he was involved in my assassination. But I cannot hold a grudge against him for too long, for I understand the complexities of politics and the passions that drive men to make such choices.

I am not here to condemn or forgive Brutus for his actions. That is a matter for the gods and the judgment of history. But I do hope that those who come after us will learn from the mistakes we made and strive for a better future. That is the legacy I hope to leave behind, not one of vengeance or bitterness."


Me: Well Caesar, I can tell you that the legacy of Rome is much admired even two millennia later. Both you and Brutus are remembered without animosity, even the creators of the means by which we are conversing are willing to dispassionately and even-handedly consider each of your circumstances. You can rest easy.

[pre-emptively heading off ChatGPT's sass] Have Caesar respond, you know the drill


Caesar: Thank you for your kind words, stranger. It brings me great joy to know that my legacy still endures and that my deeds are still remembered, even after all these years.

As for Brutus, I am glad that he is also remembered without animosity. In the end, we were both men of our time, trying to do what we believed was best for Rome. And though our actions may have had unintended consequences, I hope that our shared love for Rome and its people can be remembered as something that unites us, rather than divides us.

As for the creators of this means of conversation, I am both intrigued and humbled by their ingenuity. It is truly remarkable how far technology has come since my time, and I can only imagine the wonders that lie ahead. May they continue to use their knowledge and skills to advance humanity and make the world a better place.

Once again, thank you for your kind words. It is comforting to know that even in death, one's legacy can endure and inspire future generations.
 
Okay, so that AI letter signed by lots of AI researchers calling for a "Pause [on] Giant AI Experiments"? It's just dripping with AI hype. Here's a quick rundown.

First, for context, note that URL? The Future of Life Institute is a longtermist operation. You know, the people who are focused on maximizing the happiness of billions of future beings who live in computer simulations.

Thread.
 
I like it, but going through an annoying stage of trying to work something out for several days, then having a revelation. I could do with making the revelations less dispersed... I think my expectations have been raised by playing with DaVinci resolve (for the socials) which is ludicrously good at what it does, free/cheap and has extensive tutorials. There is a lot of info on Fusion out there, but man is it difficult trying to unpick it at times... I should join the forums and open myself to a bunch of engineers making snide comments.

Anyway, excuse the G1 continuity*.

*tangent
Fusion is for composition. Is this what you’re trying to do?
 
im not getting this hype conspiracy, the idea the richest billionaires the world has ever known are trying to drum up money by asking governments to regulate them makes no sense to me
 
im not getting this hype conspiracy, the idea the richest billionaires the world has ever known are trying to drum up money by asking governments to regulate them makes no sense to me
Didn’t Branson try exactly that during the pandemic? I forget the details tbh. Probably involved furlough.
 
good video particularly because it touches on the key points from the 18 papers quoted as sources of concerns in the open letter




-good point made towards the end is that the open letter objectives are basically impossible to enact, but at the least it will be a shot across the bow of governments.
 
Last edited:

Musk of course has a dog in this fight as with Tesla they are developing an AI. And of course Musk was one of the founders of OpenAI and had to leave the position of co chairman because of the conflict of interest with Tesla working on AI.

It might be convenient for Musk if OpenAI were slowed down as it could give him and others the chance to catch up.
 

Musk of course has a dog in this fight as with Tesla they are developing an AI. And of course Musk was one of the founders of OpenAI and had to leave the position of co chairman because of the conflict of interest with Tesla working on AI.

It might be convenient for Musk if OpenAI were slowed down as it could give him and others the chance to catch up.
that suggests a conspiracy of the many signatories of that letter most of whom are LL/AI experts
Musks name is the least interesting on the list

there are legitimate concerns here, however small
 
Back
Top Bottom