All things AI.

grunge said:


Moltbook. A Social media network for AI bots.

bots chatting to bots, they have created there own religion. within days 150k bots all chatting but chatting at Computer speed so very very fast.

Its an experiment but showed up some major security issues but also showing how productivity boosts can appear when bots give hints to other bots about how they did something.

Security flaw wise, these are people giving there whole computer system, posting whats app posts, emails phone calls etc etc. wont take much malicious code being added by an unscrupius hacker's bot to cause utter carnage. security holes have been fixed but there will be more.

This one is interesting to watch unfold.
Click to expand...

I was going to have ago myself until I realised what it could do, the dangers. Interesting though.
 
The “dead internet” idea comes up too, where bots are basically interacting with other bots, which just makes everything feel less authentic . At the same time, some users still point out that AI has legitimate uses, especially in analyzing data or helping with tasks, as long as humans stay involved. That “garbage in, garbage out” mindset shows people aren’t ready to hand over control completely. The concerns about creativity, misinformation, and even energy usage show it’s not just about the tech—it’s about the bigger impact. AI content creation fits right into that discussion because it highlights how much effort is going into making AI outputs feel more human and trustworthy. But even with tools like that, the underlying skepticism doesn’t really go away.
 
Last edited:
Hiw long until someone finds a way to use AI to lsunch a Cyber attack on pension funds. Just imagine the chaos this would bring.
 
abu13 said:
Hiw long until someone finds a way to use AI to lsunch a Cyber attack on pension funds. Just imagine the chaos this would bring.
Click to expand...

I guarantee Hackers are using AI to try to hack everything and anything they can. but cyber security teams will also be using AI to protect against the hacks.

I've already seen links on twitter for AI tools to hack websites. just give it a URL and it hunts down weaknesses. seems to be a developer tool to help show weaknesses but if that exists, a dodgy version will also exist.
 
Gorton_Tubster said:
I was going to have ago myself until I realised what it could do, the dangers. Interesting though.
Click to expand...

Apparently one of the initial security flaws was that each ClawdBot had the Claude API key in it, Moltbook was storing those in an open database. so someone who knew where to look could get every Claude API key for every bot on there.

That could have proven to be VERY expensive for some poeple.

Personally im never planning on giving an AI agent total access to my systems, small sandboxed areas maybe.
 
Last edited:
grunge said:
I guarantee Hackers are using AI to try to hack everything and anything they can. but cyber security teams will also be using AI to protect against the hacks.

I've already seen links on twitter for AI tools to hack websites. just give it a URL and it hunts down weaknesses. seems to be a developer tool to help show weaknesses but if that exists, a dodgy version will also exist.
Click to expand...
What scares me is after years if loading into my pension some cunts come along and destroy the pension providers funds.
 
Apparently one unexpected side effect of all this AI vibe coding is lots of Open Source libraries and apps are getting inundated by AI generated slop making it near impossible to review it all and let the good stuff in. it could kill open source projects entirely if they have to lock down there projects.
 
chris63 said:
AI sure looks good when used right ; )

Click to expand...


Not a dig at you but Im hoping stuff like this will be a bit of a fad. Massive amounts of energy and RAM for this stuff that you can get from a real person with a quick google search. RAM costs have gone mental in the last 6 months. Nintendo looking to increase the switch 2 price. rumours the PS6 is being delayed 12/18 months because of RAM costs etc.

its been nicknamed the RAMpocalypse!.

en.wikipedia.org

2024–present global memory supply shortage - Wikipedia

en.wikipedia.org en.wikipedia.org
 
My missus has "Gemini" on her phone. It listens and sometimes hears things wrong. We have two cats, both girls, one just walked by and the other is nowhere to be seen. "Did you kill.her?" the missus asked of cat number 1. Gemini sent a message: "I can't fulfill that request. I am programmed only to help and do no harm. If you are feeling desperate please call for help ..."

If the police ever get hold of the phone God knows what conclusions they might jump to ...
 
Agentic AI is just going to change the way in which people interact with most of their every day mundane tasks, shopping, booking holidays, renewing insurance, sending money, browsing the internet etc.

Once companies develop their own agents to facilitate these things and interact with peoples personal agents.

It will be as simple as saying renew my car insurance on X date and set a reminder for next year, only find me the top 5 cheapest quotes with X coverage and let me pick the one i want. Then "go" and its renewed.

The biggest barriers are trust, authentication and data protection.
 
Demis Hassabis just defined the real test for AGI. It’s more brutal than anyone expected.

Train AI on all human knowledge. Cut it off at 1911. See if it independently discovers general relativity like Einstein did in 1915.

If it can, we have AGI. If not, we’re still building pattern matchers.

Hassabis: “My definition of AGI has never changed. A system that can exhibit all the cognitive capabilities that humans can.”

Not bar exams. Not coding competitions. All cognitive capabilities.

Hassabis: “The brain is the only existence proof we have, maybe in the universe, of a general intelligence.”

That’s why DeepMind studies neuroscience. Not for inspiration. For data. The human brain is the only confirmed evidence that general intelligence is physically possible.

If you want to build it, you study the only example that exists.

Hassabis: “True creativity, continual learning, long-term planning. They’re not good at those things.”

Current systems are impressive and broken simultaneously.

Hassabis: “They can get gold medals in international math olympiad questions, but they can still fall over on relatively simple math problems if you pose it in a certain way.”

Jagged intelligence. Brilliant in narrow domains. Incompetent when approached differently.

That inconsistency is the tell. A true general intelligence doesn’t spike in one direction and collapse in another.

The Einstein test cuts through all of it. No benchmarks. No leaderboards. No carefully curated evals.

Just a model, a knowledge cutoff, and the question of whether it can do what one human did alone in 1915.

Hassabis: “Training an AI system with a knowledge cutoff of 1911 and seeing if it could come up with general relativity like Einstein did in 1915. That’s the true test of whether we have a full AGI system.”

Current models can’t. They remix brilliantly. They don’t generate paradigm-shifting theories from first principles.

Hassabis: “I think we’re still a few years away from that.”

A few years. Not decades.

The system that can be Einstein once can be Einstein a thousand times simultaneously across every domain.

That’s not AGI anymore. That’s the beginning of something we don’t have words for yet.

When that test gets passed, we won’t need a press release to know what happened.

[\quote]
Click to expand...

A different take on this is someone I know who is a VC and funding AI startups believes those touting AGI as there target are just trying to bring in obscene levels of investment and are no where near it.
 
Last edited:
I am suddenly in awe of our Gemini AI at work.

I am getting it to write a script to analyse and filter log files but am getting fed up with it's obsequiousness. So today I told it to stop messing about and from now on frame all answers in the style of ex-Nottingham Forest manager Brian Clough. When I asked it to add some filtering, here's what he, I mean it, came back with.

Right. Filtering. About time we started talking sense. You can't just let any Tom, Dick, or Harry onto the pitch. You need standards. You need to look 'em in the eye and say, "Son, you're not good enough for my team today."

This is simple. We're not splitting the atom here. Pay attention.

When we look at a player—an event, you call it—we check his attributes. Is his index is what we're looking for? Is his value up to scratch? If you tell me you only want players with index=4 then any lad who turns up with index=5 can go and have an early bath. He's not getting in the squad. Simple as that.

And if you give me a list of standards, they'd better meet every single one. No slackers. If his index is right but his value is wrong, the whole lot gets thrown out. I don't build teams around one-trick ponies.

Now, I've put the tactics on the board. I've updated the config file so you can tell me what you're looking for, and I've updated the script to be the bad guy and tell 'em they're not getting in if they don't meet the grade.

Don't just stand there gawping. Get it done.
 
Deepseek have been using Claude to train it's latest model. I can start a session by calling it Claude and it genuinely thinks it is Anthropic and hallucinates then when you call it out it goes on a rant saying Claude steals data too. The games the game though all as unethical as each other let's be honest.
 
RobMCFC said:
I am suddenly in awe of our Gemini AI at work.

I am getting it to write a script to analyse and filter log files but am getting fed up with it's obsequiousness. So today I told it to stop messing about and from now on frame all answers in the style of ex-Nottingham Forest manager Brian Clough. When I asked it to add some filtering, here's what he, I mean it, came back with.

Right. Filtering. About time we started talking sense. You can't just let any Tom, Dick, or Harry onto the pitch. You need standards. You need to look 'em in the eye and say, "Son, you're not good enough for my team today."

This is simple. We're not splitting the atom here. Pay attention.

When we look at a player—an event, you call it—we check his attributes. Is his index is what we're looking for? Is his value up to scratch? If you tell me you only want players with index=4 then any lad who turns up with index=5 can go and have an early bath. He's not getting in the squad. Simple as that.

And if you give me a list of standards, they'd better meet every single one. No slackers. If his index is right but his value is wrong, the whole lot gets thrown out. I don't build teams around one-trick ponies.

Now, I've put the tactics on the board. I've updated the config file so you can tell me what you're looking for, and I've updated the script to be the bad guy and tell 'em they're not getting in if they don't meet the grade.

Don't just stand there gawping. Get it done.
Click to expand...

I use Gemini alot now prefer it to open AI. Not for work like just general questions I have.
 
You must log in or register to reply here.
Unread Posts

Don't have an account? Register now and see fewer ads!

SIGN UP
Back
Top