AI

Webzcas

Winter is Coming!
Staff member
Joined
Mar 31, 2005
Location
Block S25, South Stand, Ashton Gate, BS3
So no doubt everyone here has heard about the growth of AI over the past 18 months or so, unless you have been living under a rock and disconnected from society and the internet during this time.

Well for me to assist with my work I used to have a monthly subscription to Claude.ai and also Midjourney for art work and suno for music. Part of my new year resolutions is to cull needless subscriptions. So the AI ones which mount up added together were and are prime targets.

Now to run AI models also known as LLM's ( Large Language Model ) you do need a fairly decent machine and fast memory, lots of fast memory is very desirable, should you want them to be responsive.

Of course running an LLM locally on your own machine is not going to be as fast as the likes of ChatGPT et al, which have vast cloud resources at their disposal. But you can get some pretty favourable results.

So what have I done?

I have setup Ollama on my M2 Pro Macbook Pro and have created a webgui to allow me to choose which LLM to run, so I get the full ChatGPT effect. My plan being to upgrade my machine at some stage this year which will allow me to run 30billion parameter models.

A low down to what this means, let me let HAL 9000 ( The name I have given my own AI chatbot ) explain :D

Screenshot 2025-01-13 at 07.39.18.webp


So this is my AI chatbot, the web interface I am continuing to work on. For replacing my sub to Midjourney, I have setup and configured Stable Diffusion, which is absolutely different class.

Below is the very first image I created with it

1736754239052.webp


This was meant to be an Ancient Egyptian. LOL

But finally have come up with some superb images. I borrowed the prompts for Gandalf and Darth Vader, so cannot take sole credit for them. But these were generated on my Mac. The power of Apple Silicon is there for all to see :D

1736754369739.webp


1736754416175.webp
 
Last edited:
AI is a scam, a grift. Industrial scale theft and larceny. It knows nothing, it thinks nothing, it regurgitates stolen material, hallucinates nonsense and presents untruths with a confidence that would be amusing if it weren't so predicated on stealing the sum total of human endeavours for the enrichment of billionaires.

The only good news is that sooner or later this horseshit Ponzi scheme will collapse in on itself.

I'm a big fan of this guy and his work:

You do not have permission to view link Log in or register now.


We occasionally dabble with AI in my line of work (it is artificial, but it's definitely not intelligent), we've seen it make every mistake in the book, including making up entirely new cmdlets, stuff that literally doesn't exist in code, and then when we correct it, it fesses up and admits it hallucinated it.

We use AI as the punchlines to jokes, like, 'Have you asked Chat-GPT about it..... LOL'.
 
Amazing... I don't use AI as they are too limited for me but I wanted to create an editor for my huge family. I'm the only one who speaks English well enough, I have tons of cousins and 2nd 3rd cousins and now that I'm kinda old, everyone is in school and almost to uni, they have lots of requests to help check their English.

I tried asking them to use things like Grammarly so it helps with spelling and proofreading but, it doesn't really give helpful comments to improve the way they explain certain things. Meaning, they always recommend a good grammar structure but they don't know what certain schools or ways of writing want to get good marks for example you must know the style the school or country likes. I'm in Singapore, Singapore English is NOT Grammarly English, for example :)

Long way to get to my question is. Is there a way, to install and teach something like ChatGPT to edit their work in such a way? It would remember all my comments. So if it sees the same thing it will comment the same comment like with language specific translations and explanations. I want to do this and even give it to teachers in my village to help them. And is it going to cost a lot? Sorry OP if thats too much to ask, just never saw anyone do like you :)
 
AI is a scam, a grift. Industrial scale theft and larceny. It knows nothing, it thinks nothing, it regurgitates stolen material, hallucinates nonsense and presents untruths with a confidence that would be amusing if it weren't so predicated on stealing the sum total of human endeavours for the enrichment of billionaires.

Really @ChopleyIOM , scam? For me it has replaced the likes of Stack Overflow. Certainly not infallible, but a superb assistant IF you know how to use it :-)
 
Long way to get to my question is. Is there a way, to install and teach something like ChatGPT to edit their work in such a way? It would remember all my comments. So if it sees the same thing it will comment the same comment like with language specific translations and explanations. I want to do this and even give it to teachers in my village to help them. And is it going to cost a lot? Sorry OP if thats too much to ask, just never saw anyone do like you :)

Yes, there is. The method is called RAG which stands for Retrieval-Augmented Generation.

Have a read here:
You do not have permission to view link Log in or register now.
 
Really @ChopleyIOM , scam? For me it has replaced the likes of Stack Overflow. Certainly not infallible, but a superb assistant IF you know how to use it :-)

Yes but what you're doing with it isn't 'AI', it's an evolution of machine learning which has kind of been a thing forever, although admittedly we have far more processing power to chuck at it these days which makes LLMs feasible. (And lest we forget, LLMs are basically just 'word guessers'.)

Don't get me wrong, it's impressive technology, but it absolutely isn't 'artificial intelligence', and to see it constantly being branded as such gets on my tits.

Chat-GPT isn't AI, Gemini isn't AI, Grok isn't AI - none of it is, and none of it is remotely commercially viable, the most impressive thing about Open AI is the rate at which they're burning through billions of dollars and how long they can keep finding suckers to carry on propping up their edifice of nonsense.
 
I have been using Jan AI, which runs locally from my computer for a few months now. Some of the models require 32GB memory to load and big downloads on top! It's fast and easy to use with voice or text with answers in seconds.

There are three areas which I have used specifically for gambling.

First - Slots. I can say suggest a NetEnt slot to play or Microgaming with 243 ways. Often it suggests a game I haven't played for a few years like the other night Dr Watts Up which led to over 500X in the bonus. Asking for a Megaways suggestion is always fun as there are so many now. It knows so much about each game it sometimes throws in a pun like 'Go get that G O L D' in Bonanza for example.

Second - Live Blackjack. Using a private table, I ask for a new shoe and how many decks there are. As the cards come out, using voice prompts, the AI model on command helps with decisions. As an example, I can ask how many cards under SEVEN remain or how many ACEs are there. Of course the cutting card puts things in the favour of the casino.

Third - Crazy Time. I can load the last 3 months of results into the software from Tracksino. I know the wheel is RNG, but it adds some extra fun to proceedings.
 
I have been using Jan AI, which runs locally from my computer for a few months now. Some of the models require 32GB memory to load and big downloads on top! It's fast and easy to use with voice or text with answers in seconds.

There are three areas which I have used specifically for gambling.

First - Slots. I can say suggest a NetEnt slot to play or Microgaming with 243 ways. Often it suggests a game I haven't played for a few years like the other night Dr Watts Up which led to over 500X in the bonus. Asking for a Megaways suggestion is always fun as there are so many now. It knows so much about each game it sometimes throws in a pun like 'Go get that G O L D' in Bonanza for example.

Second - Live Blackjack. Using a private table, I ask for a new shoe and how many decks there are. As the cards come out, using voice prompts, the AI model on command helps with decisions. As an example, I can ask how many cards under SEVEN remain or how many ACEs are there. Of course the cutting card puts things in the favour of the casino.

Third - Crazy Time. I can load the last 3 months of results into the software from Tracksino. I know the wheel is RNG, but it adds some extra fun to proceedings.

Ahhh yes, 'databases', I've heard of those :)

I tried asking Chat-GPT about fruit machines, it got nearly all the questions completely wrong, but it did so with an impressive degree of confidence.
 
So no doubt everyone here has heard about the growth of AI over the past 18 months or so, unless you have been living under a rock and disconnected from society and the internet during this time.

Well for me to assist with my work I used to have a monthly subscription to Claude.ai and also Midjourney for art work and suno for music. Part of my new year resolutions is to cull needless subscriptions. So the AI ones which mount up added together were and are prime targets.

Now to run AI models also known as LLM's ( Large Language Model ) you do need a fairly decent machine and fast memory, lots of fast memory is very desirable, should you want them to be responsive.

Of course running an LLM locally on your own machine is not going to be as fast as the likes of ChatGPT et al, which have vast cloud resources at their disposal. But you can get some pretty favourable results.

So what have I done?

I have setup Ollama on my M2 Pro Macbook Pro and have created a webgui to allow me to choose which LLM to run, so I get the full ChatGPT effect. My plan being to upgrade my machine at some stage this year which will allow me to run 30billion parameter models.

A low down to what this means, let me let HAL 9000 ( The name I have given my own AI chatbot ) explain :D

View attachment 204725

So this is my AI chatbot, the web interface I am continuing to work on. For replacing my sub to Midjourney, I have setup and configured Stable Diffusion, which is absolutely different class.

Below is the very first image I created with it

View attachment 204726

This was meant to be an Ancient Egyptian. LOL

But finally have come up with some superb images. I borrowed the prompts for Gandalf and Darth Vader, so cannot take sole credit for them. But these were generated on my Mac. The power of Apple Silicon is there for all to see :D

View attachment 204727

View attachment 204728
Can you AI image me 32 ways of 6OAK diamonds on Bonanza at 15x multiplier?
 
Third - Crazy Time. I can load the last 3 months of results into the software from Tracksino. I know the wheel is RNG, but it adds some extra fun to proceedings.

How do you know it's RNG? When I looked behind it, there were no motors or breaking mechanisms.
 
Second - Live Blackjack. Using a private table, I ask for a new shoe and how many decks there are. As the cards come out, using voice prompts, the AI model on command helps with decisions. As an example, I can ask how many cards under SEVEN remain or how many ACEs are there. Of course the cutting card puts things in the favour of the casino.
Deck penetration is so low online I don't even bother.
 
I love seeing AI struggle with text.

Or anything, really.

This is why I take issue with the 'intelligence' aspect of the descriptor, because it has none at all. LLMs just guess at what word is statistically most likely to come next.

1736783340208.webp
 
AI is a scam, a grift. Industrial scale theft and larceny. It knows nothing, it thinks nothing, it regurgitates stolen material, hallucinates nonsense and presents untruths with a confidence that would be amusing if it weren't so predicated on stealing the sum total of human endeavours for the enrichment of billionaires.

The only good news is that sooner or later this horseshit Ponzi scheme will collapse in on itself.

I'm a big fan of this guy and his work:

You do not have permission to view link Log in or register now.


We occasionally dabble with AI in my line of work (it is artificial, but it's definitely not intelligent), we've seen it make every mistake in the book, including making up entirely new cmdlets, stuff that literally doesn't exist in code, and then when we correct it, it fesses up and admits it hallucinated it.

We use AI as the punchlines to jokes, like, 'Have you asked Chat-GPT about it..... LOL'.

Haha something we can agree on at last!

That was a good article, never realised how much money has been pumped in to create these platforms/products, the media have never mentioned it, I thought it would be just a few tech nerds doing their stuff with the more powerful computers - and wham bam thankyou mam.

According to Starmer we're going to be world leaders, so he must mean beating amazon, Microsoft and google etc... unless they're now counted as uk firms.😵‍💫
 
Haha something we can agree on at last!

That was a good article, never realised how much money has been pumped in to create these platforms/products, the media have never mentioned it, I thought it would be just a few tech nerds doing their stuff with the more powerful computers - and wham bam thankyou mam.

According to Starmer we're going to be world leaders, so he must mean beating amazon, Microsoft and google etc... unless they're now counted as uk firms.😵‍💫

I predict NHS data being sold to private firms to SUPERCHARGE THE UK'S AI PUSH, few quid in the government's coffers and the population of the UK gets sold out, they'll brand it as 'an amazing new deal for patients' or something.

Fiddling while Rome burns.

As ever with UK government pronouncements on anything that has computers in it, whichever political party is in power, it's people who know very little talking nonsense about stuff that will never happen, although you can be sure that there are plenty of snouts already lining up to plunge into the trough.

If you like Ed Zitron's writing his latest polemic about tech in general, not just AI, is a superb, if rather sobering read.

You do not have permission to view link Log in or register now.


------------

I will never forgive these people for what they’ve done to the computer, and the more I learn about both their intentions and actions the more certain I am that they are unrepentant and that their greed will never be sated. I have watched them take the things that made me human — social networking, digital communities, apps, and the other connecting fabric of our digital lives — and turned them into devices of torture, profitable mechanisms of abuse, and find it disgusting how many reporters seem to believe it's their responsibility to thank them and explain why it's good this is happening to their readers.

  • You do not have permission to view link Log in or register now.
    who would burn our planet to the ground, steal from millions of people and burn billions of dollars in pursuit of power, and I believe the same can be said of people like Dario Amodei of Anthropic and Mustafa Suleyman of Microsoft.

  • Tim Cook is a wolf in sheep’s clothing, slowly allowing the rot to seep into Apple’s products, slowly adding bothersome subscription products and useless AI features to chip away at the user experience. Apple’s app store and its repeated support of exploitative microtransaction-laden mobile games
    You do not have permission to view link Log in or register now.
    in adults
    You do not have permission to view link Log in or register now.
    ,
    You do not have permission to view link Log in or register now.
    . Because Apple’s products are less shitty, it gets a much easier time.

  • Sundar Pichai is the Henry Kissinger of technology — a glossy executive that escapes blame despite having caused harm on a global scale.
    You do not have permission to view link Log in or register now.
    should be written about like a war crime, and those responsible treated as such.

  • Satya Nadella has aggressively expanded Microsoft’s various monopolies, the most egregious of which is the Microsoft 365 suite — a monopoly over business software that everybody kind of hates that Microsoft prices to undercut the competition, effectively setting the conditions of most business software as either “cheaper than Microsoft” or “slightly better than Microsoft.” Nadella has overseen layoffs of tens of thousands of people in the last three years alone, and despite his bullshit “
    You do not have permission to view link Log in or register now.
    ” culture treats his employees and customers as equally disposable.

  • Mark Zuckerberg is a putrid ghoul that has overseen the growth and proliferation of some of the single-most abusive and manipulative software in the world. Meta has grown to a market cap of $1.5 trillion dollars by intentionally making the experience on Instagram and Facebook worse, intentionally frustrating and harming billions of people.


These are the people in charge. These are the people running the tech industry. These are the people who make decisions that affect billions of people every minute of every day, and their decision making is so flagrantly selfish and abusive
You do not have permission to view link Log in or register now.
.

These men lace our digital lives with asbestos and get told they’re geniuses for doing so because money comes out.
 
Yes, there is. The method is called RAG which stands for Retrieval-Augmented Generation.

Have a read here:
You do not have permission to view link Log in or register now.
Thank you. It seems really technical to me but I'm gonna read this a lot :)
 
Honestly this is like diving down a rabbit hole. The potential is amazing. I mean just from the tech side, having your own chat bot that is proficient in the areas you work in is a game changer. Not having to worry about removing credentials, file paths etc before querying it, knowing full well everything shared is secure on your own machine.
 
Honestly this is like diving down a rabbit hole. The potential is amazing. I mean just from the tech side, having your own chat bot that is proficient in the areas you work in is a game changer. Not having to worry about removing credentials, file paths etc before querying it, knowing full well everything shared is secure on your own machine.
The thing with all this stuff is right now is the worst it will ever be, it will only improve. Right now everyone is trying to shoehorn "AI" into products and add AI buzzwords to their marketing and a lot of the use cases for LLMs are rubbish. There are definitely some excellent use cases currently but these are overshadowed by all the dumb ones.

I disagree with the notion that its a ponzi-like grift that will eventually crumble. I don't particularly like a lot of stuff relating to AI but I see it as an inevitable part of the future that is only going to continue to be refined and become more powerful as it approaches AGI.

It's already happening but one of the most annoying things so far is completely artificial youtube channels popping up and its only going to get worse in the next few years. Crappy AI slop taking money and views away from real people.
 
Yes, I think you and I are on the same wavelength when it comes to AI, Mulven.

The tech that is in the public domain and open source, is still very much in its infancy. But it is incredible as to the assistance and help it can provide already. For developers there is github copilot, which is for want of a better word, SCARY!!
 

Users who are viewing this thread

Meister Ratings

Back
Top