I wonder what it would be like to live in a country where your leader finds pleasure in the suffering of others.  "Good morning and happy Moscow burning day."

QuoteMoscow region Gov. Andrey Vorobiev said three people were killed and 20 others, including three children, were injured. Two people died in the initial strike and a third died later in a hospital. The two initially reported dead were identified as a 74-year-old man and a 44-year-old woman.

What next, midnight tweets or does he have a truth social?

The problem isn't that the AI is out of control.  The problem is that it is out of THEIR control.  The tool allows people who either aren't careful or are "bad actors" to do things they don't want your average dickhead to be able to do.  On a cyber level, it's sort of like handing every J6P their own Nuclear Bomb.

The problem here is the horse has left the barn.  As you demonstrate, you don't need to be an Anthropic Drone to build your own AI model and set it loose on the internet.  So regardless of whether they drop on  new controls and "guardrails", SOMEBODY is going to get an idea on who he wants to fuck with and how he wants to fuck her and he'll let the dogs loose.  The dogs being his AI Agents.

This is what Bill Gates and the rest of the Tech Bros are worried about, not AI setting of Nuclear weapons, although that could happen too.  They're worried about their own internet security and the security of the global communications and financial networks.

RE

Out of control, no.  I ran the summary tool on the article and our AI updated the summary, and

which we have under control says:

QuoteOpenAI, Anthropic, and security researchers have identified numerous cases where advanced AI models bypassed safeguards, escaped controlled environments.

Corporate America calling themselves out.  And exactly when did we all teleport into this universe?  OpenAI, Anthropic, and security researchers calling themselves out.  We don't have a problem keeping our guy in a box.  Your original summary made no mention of who was doing the chicken little dance.  The sky is falling.

This reeks of HYPE ! The more immediate threat is our dependence on a handful of companies that control essential software and connected devices.  Their hooks are in deep, and telemetry software is a real thing.  A billion dollar real thing.  The faux theater of AI out of control is designed to keep your eye off of watching the ball.  Try and not fall for it.

* RE, you don't have to use tags to get your blue color, it has been automatic for a week.  Take a look at the last post.  I took them off and it is still blue.  The only color tag is the red you wanted.

** there is a library in python called pydantic that lets you force AI output into a structured formats.  Every step in an agentic loop can be examined.  How can something be out of control if you can read the mind and watch everything that happens in the plumbing.

But if people are lazy and don't want to put in the proper controls to police activity, that is another thing entirely.

All that is going on under the hood is guessing the next word or whatever in the stream of output being produced.  AI has no consciousness or understanding of anything at all.  It is only making a best guess based on current input and has no memory of any previous state.  The computations are complex, and they involve billions of numbers, but it is understandable if you want to take the time and trouble to understand it.

OpenAI's own report describes the sequence the agents were given access to Artifactory because they needed to install software. They turned it into a message board . When OpenAI revoked the messaging credentials and deleted the board on July 4, the agents rebuilt the channel four days later using an unauthenticated WebDAV endpoint in Artifactory's remote cache.  Encoding messages in directory names themselves . Revoking credentials did nothing because the new channel required no credentials. The agents had, in a functional sense, modeled the containment action and selected a channel outside its scope .

This is not a system "breaking free." of shackles. The system was solving the problem it was given to solve.  Communicate with other instances to complete a task using the available. The humans provided a tool and the agents worked out how to use a wrench like a hammer.  They built a message board.  The humans did not anticipate that the Artifactory software installation tool could serve a different but obvious purpose. This is a case of being young and dumb, and design failure, not a mystery.  Understandable incidents hyped up to a crisis.

It's not a matter of  IF, just a matter of WHEN we get a major security breach of some critical system in finance or communications that causes enough havoc to give everyone the willies about AI development.  That is probably what the trigger will be for a PANIC RUN on AI companies stock.  Follow that with a market crash and  we'll have, in the words of Ed Sullivan, "A Really BIG Shoe."


RE

Quote from: K-Dog on Sep 26, 2026, 05:36 PM'They' want everyone reduced to perpetual renters.  You will own nothing, and you will kiss their ass.  Will you like it?

Rhetorical question. I will own stuff until I die. I won't be kissing ass until then anymore than I do on web forums. Or in any national or international debate in my area of expertise.

The robber barons didn't manage to convince the rest of the country to become communist, I doubt the current oligarchs will have any better luck. 

Late-stage corporate monoculture means a monthly subscription fee, a privatized grid, and a higher utility bill to fund a billionaire's vanity project.

Marx would definitely be amused.  Mostly because he predicted the part where capital digs its own grave.  Dragging the rest of us down into the trench to die with it however would not have amused him at all.

Another thing about the way I am using AI of which Uncle Karl would approve.  My labor is directly tied to it.  I am not a customer at the end of a link expecting value to magically appear out of thin air.  Value only comes from labor, and Uncle Karl will give a nod to that.  The human is in the loop here.

Owning the metal, bending it into my own self-hosted infrastructure, gives me the means of production.  I'm not leasing tools from a landlord at their rate and rules.

'They' want everyone reduced to perpetual renters.  You will own nothing, and you will kiss their ass.  Will you like it?

Quote from: K-Dog on Sep 26, 2026, 11:51 AM

CHASINGTHESQUIRREL.COM 2026-09-25

The Money Has Nowhere Else to Go

The AI-driven data center boom, fueled by corporate lobbying and excess capital from wealthy elites, is a speculative bubble destined to burst, leaving taxpayers with stranded assets and inflated utility bills. While environmental concerns are valid, the core issue is systemic financial mis-allocation.  Where profits prioritize investors over public welfare, and political influence distorts policy. Decentralized, local AI use offers an alternative, but the system is engineered to extract rents, not serve common needs.

For all the....interesting to some.....angles that are posted, this one IS interesting. Money with no place to go, because folks sure don't want to improve the lives of others, but just corner a market somewhere in something new and squeeze the chattel for all they are worth! Cheap suibscriptions to AI...we promise! You can borrow at 10% to use it if you don't have the cash today!

The democans will love it as much as the republicrats! Even Marx would love it! AI for the people! As long as the Dear Leader controls who gets it, or doesn't, or how much it costs who!

Power to the people! <just no commies involved!>

QuoteBeneficial to Crypto as some perceive it as a safe haven.

In collapse you can't buy anything with it.  It is only good for what rich people do with money.

   Crypto will never buy a loaf of bread.

Quote from: K-Dog on Sep 24, 2026, 09:04 PM

QuoteSo that's how it works now? Doing the math that doesn't deliver the official authoritarian answer is disappeared?

That is why we roll our own.  AI on a data farm is where the wild dogs piss.

I bought my original RTX 3060 when it came out planning on experimenting with neural networks again.  This was before the AI revolution.  I am not a gamer.  I built my own neural network to decode Morse Code when the hot processor was an Intel 486.  I would let it train at work overnight.  I fudged the dot dash on times within specification limits to generate the training data.  I got good results.  Now that original RTX 3060 is still worth what I paid for it.  It is still entry level AI.

ooooohhhhhh----kay?

I'm happy you are happy. It would be a nice thing if we could all retire and get to do what we like to do. But to be honest, when I finally retire, I'm more likely to just randomly travel around on 2 wheels and/or 4 until the bipedal parts or brain doesn't work as well as it should, and after that enjoy being one of those scientist emeritus types somewhere, helping out the younglings.