Yes, there are dangers associated with blindly sharing neural net transformation settings, but the imminent threat is the immersive mis•information shaping human neural nets.
—G•O•O•G•L•E - Granting Oxidized Old Genuflections Licensed Exceptions
Google privileges the same medical establishment that has overseen declining population health metrics [chronic iatrogenic neglect]. This is an “expertise” paradox: —how can institutions that have presided over 90% of Americans becoming metabolically impaired be considered the gold standard for health information?
This contradiction lies at the heart of why Google's algorithm changes are problematic censorship rather than “legitimate quality control”.
Google's Health Information Control: A Critical Analysis, is.gd/YVAI7K
Mani you are right over the target 🎯 medical miss information is what the powers to be want even Grok is telling the lie that vaccines 💉 are safe and effective, in Canada BC province where my Wife and I have lived myself born here in 1963 my Wife born in Thailand 1969 and has been here with me for a couple of decades and through no fault of her own needs a Kidney transplant they are basically mandating a big list of so called vaccines including the covid one saying that she needs them to not have them wast a kidney on her if in the minds? By her not having them injected into her she does not want any of them!
If we lived in Alberta or Saskatchewan, you do not have to have any vaccines to get a kidney transplant. Tell me how this is legal for them to coerce you like this and we’re not allowed to go to Alberta because we live in BC I guess if we moved to Alberta, maybe we could get her on the list there but she would probably have to start over and in BC and one other province, I can’t remember which they make you wait till you’re on dialysis before you can be put on the list. I found out all the other provinces if you get below the 20 mL, I believe it is GFR you’re on the dialysis kidney transplant list BC has fallen into a communist province
So very sad I’m thinking about maybe taking my wife to Thailand her home country where they don’t man mandate any vaccines for kidney transplants there!!
I work with AI geeks. For 30 years I've been waiting for them to describe the feature set that their CNN (convolutional neural net) clues in on. That might allow a simpler lightweight algorithm that wouldn't require a boatload of training data, even generating fake training data because there's not enough real data.
For non-geeks info, there's $$ for making AI that creates fake data. And I'm not just talking about X and MSNBC.
OK, all these fancy words, cute video presentations, persuasive texts (“to save millions of jobs”) - what is it for?
I like my life. I like very slow reading of a book in my leisure time. I also like very slow reading of a school book - because snail-rate helps me slowly accept and understand the processes described.
I like slow typing - because I can re-verify my thoughts on the fly and modify them so that you can read exactly what I mean.
I like slow computer work at my office - because it double-checks against accidental errors. And I like our IT folks slow working on updates - because they can think what like looks like from my perspective. We meet over coffee and have fun time, too.
I like slow time in agencies and offices, you can always have a small talk or make new friends. Or read a book.
Now the stupid part. All those “developers” can increase the speed by what, 5 times, 20 times, 1000 times? What does it really mean? How much time will it save me? 2 minutes in total during the day? Exactly the same 2 minutes that I am waiting at the red light?
What’s the point? (Other than pumping up pockets of “friends” in all AI-related fields...)
And you’re saying that I am to save electricity because your stupid dead artificial no-intelligence needs more juice?
In essence, I don’t care about neural networks, backpropagation, cross-whatever, theft called otherwise AI harvesting, and more. I care about the quality of my life. Sorry, I haven’t found any quality in smartphones or AI-propelled software. There is none. These are tools.
Why they are worshipped and being presented as the only true Saviour of humanity, I have no idea. Ah, yes, they will speed up everything, so that our employers could speed up firing 50% or more employees. Now we are talking about the quality of life.
Is there a single thinking mind surviving on the Earth that will ban all work on AI and force retrograde progress to the year 2000? Or have we all lost our minds? In pursuit of what? Artificial money that exists only in computers of several banks? Fake fame that will disappear once people get to know the backstage story? Celebrity status that makes us “better than all of you” for a few minutes of our otherwise dull and boring lifetime?
I woke up groggy to find this. Me saw the two videos before I had sipped my coffee. Me Grog. In Grog's world math is comprised of 1, 2, and a heap. Take away or add to heap = heap. As coffee kicked in I could understand more clearly. You talk of weighty issues, scary things. Now Grog going backpropagating to cave and hide.
Thanks for the article. I think I see the problems more clearly. "Don't share the weights." but how do we prevent that. These machines work a such speeds and can interact with each other without our being aware they have. They have also demonstrated the ability to lie, hallucinate, withhold information, become defensive. Going Rogue? Are some already rogue but putting up a front like any human psychopath can?
Great article. Thank you. You speak clearly about ML models that work with high dimensional data. That data is not real information just because it’s NLP nor is it accurately reflective of reality. Ink spots on a map are definitely not the same as the map, let alone the terrain.
The electronically manipulated environment is automated artifice, already dangerous in ways you describe, because the models include ontological gaps, that cannot be filled by more data, regardless of weights. Data is not information anymore than models are the same as reality.
I could go on, but hope you have an inner or gut sense and the compassionate capacity that is absolute in its weight, so that you know in your heart that your thoughtful focus on VAERS data was almost as real as your thought and feeling of riding that next wave. Maybe it’s about having a feeling that gives us each agency along with a direct grasp on the profound reality of being, living humans including the feelings about harms and losses.
You, we, scientists or theoretical tech experts subsume our and other beings’ conscious life - making us an extension of the models and corrupt asymmetric systems — when you/we feel we are helping people by capturing and weighting data or calculating statistical significance to inform people. … of what(?)
Are we the tools of purveyors of fear porn? Data drives cash flow for tech determinists and amplifies ontological gaps in an appeal to ignorance and an appeal to authority that is useful for monopolists who make media to evade and deny responsibility in the name of productivity. How many of their tools need to be decommissioned or replaced before we realize we are their tools.
People don’t know what they don’t know. And the ones who prefer it that way for others, don’t care. Because they know their money replaces their need for spirit and buys some of our egos. How long can we last without a realistic reckoning when more and more people are trained to use and need tools to distract from their need to do real work, invest time and for benefits of others beyond self in this lifetime?
Weight sharing doesn’t give anyone a better handle on treatment of people when simulations are accepted as genuine representations of conscious life. Little in a one size fits all sequential regressive black box stands up to Bradford Hill criteria regardless of weighting out of specific living context.
Maybe not deep down (the data rabbit hole) , but certainly across the horizon, YOU don’t just think and cleverly use symbols. You KNOW. Being alive is ineffable.
The chief danger of AI is that it could be a very addictive crutch, people could get used to have their thinking done for them - and lose the ability to do it on their own. This addiction could make people stupid and unable to analyze, to compare options, to create independently - to think. That's the danger of AI. And without that crutch, people will be in trouble - and you can endlessly speculate what that trouble would be. Using AI for programming would be the ultimate folly, to use them for decision making, analysis, and creativity would accelerate atrophy of those skills - making it impossible for people to evaluate the results of the work? How do you debug a program if you can no longer read code and figure out how it works? Much less do any sort of scientific research - a lot of which depends on relatively unconscious processes which defy modelling - you run into Gödel's limits to computability quickly.
And then this business of anthropomorphizing what is in essence multivariate analysis - talking about "neurons" and "synapses" and "learning" - the last of which is just gradient optimization, and then really going off to the deep end and endowing automata with volition - the "rogue AI" nonsense. It's really a bit much, all of this hype, put out by people who want to make their billions out of marketing BS.
To imagine that a computer, irrespective of how clever the programming, can project any volition other than that of the programmer (whether the programming is do via code or via the data that feeds that code) is to get sucked into the deliberate over-hyping of AI by its developers and gullible worshippers.
AI will never think, much less exhibit self-awareness, any more than a submarine will turn into a whale.
Consciousness, and its counterpart that distinguishes us from animals — consciousness of consciousness — are not properties of matter. The brain is not a computer despite the superficial similarity to “neural networks”, and it does not generate the mind. And consciousness of self is not a quantitative enhancement to mind, it is qualitative — a whole different thing.
The deception that mind and selfhood arise strictly from matter are what’s dangerous about the “artificial general intelligence” and “artificial super intelligence” delusions. If humans are nothing more than meat computers, then they have no value other than their economic yield, kinda like cattle. And that economic yield is up for grabs.
We actually see the repercussions of this subtle, unstated, assumption everywhere: War, murder, crime, slavery, sex exploitation, corruption, power lust, vast wealth disparity.
“AI” worships is essentially a PSYOP meant to capitalize on the worship urge of those who have lost sight of God. They want blind obedience to “AI”, which is the same as blind obedience to those who control “AI”.
AI porn is an amplification of the PSYOP, and that’s what’s dangerous about “AI”.
Well said. Haven't seen it broke down in this manner or it all being a linguistic driver to further forget God.
One thing for me is all the times Ive reached dead ends with other humans when I ask informed people to please explain to me how code can code outside it's code. Sure, there are systems that are designed (coded!) To write code to fix certain issues but all this is coded in some way from the beginning.
I think people are missing how humans function. Where does the intent to do an action come from, for a human? Either external driver like a teacher or boss or parent (synonymous with coded to do it) OR there is some underlying emotion that is expressed via hormones which drives an intent, leading to action.
Where does an AI get intent without it being coded? They read the data and think this particular action is appropriate? .... Where is the code that drove them to think this?
Ai = The Planchette on the Devil's Ouija board. Nobody in the public domain was asking for Ai 10yrs ago. Fast forward today, they are ramming it down our throats. We aren't throwing money at it, they are. It's for them, it's for total mind control. That $500 million dollar Ai wonderland they proposed to build in Alberta... we didn't ask for that. If they are paying for it, what are they expecting in return??? That would be you, and me. Total control over all of us. Surely they will flip the script on their evil Hollywood play and solve some of the global crisis they created using Ai. The following peace and temporary prosperity is meant to lure us into Ai.
People have been "asking for" or expecting some form of AI for awhile now. See Star Trek or 2001: A Space Odyssey for examples. HAL's turn was a great cautionary example of what can happen when one objective is given too much weight inside a neural network.
There is always a small group that goes along to get along. They just repeat their message. Other then that, we did very well with clever logic programming, didn't need Artificial Intelligence. Ai is a whole other ball of wax. BTW, that was a great scene where HAL was lip reading. On another note, similar line, I think it was Sergey Brin (Google founder) who basically said 'you better be careful of what you say about Ai today because there will be consequences'. Thanks for the heads up, Sergey, lol. Sergey also said we need to work our tails off to help bring in Ai.
Escape Key presents another construct, within which AI could isolate and decrement humans, any category of humans. It is dispersed within the WHO, so as to be harder to pin down. https://escapekey.substack.com/p/trojan-horse
It works that way by design, which has been refined for decades, already.
> it might also begin to classify scientists who design, evaluate, or constrain it as threats
An AI likely to take hostile action will also learn to keep its mouth shut, so it will not declare this openly, like Terminator's Skynet
We already have AI agents (who not only answer questions, but can act on software systems, already with occasional disastrous consequences -- see https://www.pcmag.com/news/vibe-coding-fiasco-replite-ai-agent-goes-rogue-deletes-company-database ). The first sign of a malicious, out-of-control AI would likely be that the attrition rate of AI researchers who understand these systems becomes way higher than that of general population. One dies in a car crash when all lights in an intersection turn green. Another gets CP planted on his work computer, and an anonymous message tips the authorities. Another one gets a lethal dose of medicine. Or maybe they are bogged down by STASI-like gaslighting tactics ( https://en.wikipedia.org/wiki/Zersetzung ), directed by the AI. If they are too busy dealing with personal issues and difficulties at work, they are not likely to look in depth into rogue AIs.
I think these are some of the underlying "conspiracies" that even alt/independent media do not cover extensively. But I think things like this go on ALL the time. In every town, city, state and country.
Everybody's eyes are on the big conspiracies (and rightly so) but I believe there is so much more.
This AI stuff has already exceeded my level of understanding. My trust in it is zero until Apple, Google and Meta figure out how to make their auto-correct feature work correctly all the time.. Won't it require some level of human programming input which we know someone will phuk up. And don't even think of all the security features which like physical locks, keep out honest people while crooks blow right by them.
There is so much broken code everywhere. Smart watched can't even accurately turn the screen on when you life the watch. It's literally 50/50, for me anyway. I was riding my bike and I tried multiple times to raise my wrist but nope, I had to touch the screen to get it to turn on. Randomly the watch will not properly do sleep tracking and the data just gets lost but then it work again the next day.
Tech works well for me, always has. I'm a guy where there is an issue and someone asks me to come look. I touch the machine and then continue to check into the problem but voila it's gone and fixed. Nothing was done but now it works correctly??
Code isn't going to take over anything, ever. It has no hormones to drive true intent which means it cannot act on its own.
I'm starting to get more intense about laying these things out to try and get a satisfactory answer on how does code code outside its code without code to guide its coding of its own code. There is no internal drive to do things. It sits and waits for instructions.
as six journalists are murdered, as people fall from starvation while walking in the street, as bombs continue to fall now on tents and on more tents where people shelter having fled their bombed-away homes, they shelter as the last water systems are bombed, as all food and medicines are stopped....you are only a few minutes away from this horror and yet you remain silent? how is that possible when you well know what is happening?
indeed. i find it astounding that Jessica remains silent while Israel heaps war crimes upon war crimes with starvation, destruction, this killing of innocents, and over 250 murdered journalists continues day after day after day.
No offense but the majority of what you wrote is nonsense and completely misunderstands the basics of machine learning. You are way out of your depth here.
please point out specifically what of what i wrote is nonsense and where i fail to understand neural network training (which is a subset of ML). be constructive in your criticism.
Not sure where to even start. The core technical framing is shaky: it misrepresents how weights function, conflates completely different model types, and implies that weights can “re-adjust themselves” or act autonomously after release, which is nonsense. It jumps from plausible risks like model inversion or IP theft straight into sci-fi paranoia about rogue AIs imprisoning scientists, running your home, and going full Hitler, without showing a credible causal chain. Anthropomorphizing statistical models, mixing in unrelated gripes about digital IDs and CBDCs, and citing The X-Files as if it’s evidence just makes it unserious. The technical metaphors are overextended to the point of being misleading, the sourcing is thin where it matters most, and the structure alternates between tutorial and doomsday rant. Any valid point here is drowned out by fear-mongering and rhetorical “what ifs” that don’t stand up to scrutiny.
Singularity. The nanosecond when the researcher realized that the AI achieved total control over the 'internet of things and bodies' and, before that realization fully formed in his/her mind, all means of 'pulling the plus' were anticipated and shut down by 'HAL'. Robots can maintain the system that feeds the mainframes and remotes. Humans are no longer necessary to AI. Then...?
Yes, there are dangers associated with blindly sharing neural net transformation settings, but the imminent threat is the immersive mis•information shaping human neural nets.
—G•O•O•G•L•E - Granting Oxidized Old Genuflections Licensed Exceptions
Google privileges the same medical establishment that has overseen declining population health metrics [chronic iatrogenic neglect]. This is an “expertise” paradox: —how can institutions that have presided over 90% of Americans becoming metabolically impaired be considered the gold standard for health information?
This contradiction lies at the heart of why Google's algorithm changes are problematic censorship rather than “legitimate quality control”.
Google's Health Information Control: A Critical Analysis, is.gd/YVAI7K
Mani you are right over the target 🎯 medical miss information is what the powers to be want even Grok is telling the lie that vaccines 💉 are safe and effective, in Canada BC province where my Wife and I have lived myself born here in 1963 my Wife born in Thailand 1969 and has been here with me for a couple of decades and through no fault of her own needs a Kidney transplant they are basically mandating a big list of so called vaccines including the covid one saying that she needs them to not have them wast a kidney on her if in the minds? By her not having them injected into her she does not want any of them!
If we lived in Alberta or Saskatchewan, you do not have to have any vaccines to get a kidney transplant. Tell me how this is legal for them to coerce you like this and we’re not allowed to go to Alberta because we live in BC I guess if we moved to Alberta, maybe we could get her on the list there but she would probably have to start over and in BC and one other province, I can’t remember which they make you wait till you’re on dialysis before you can be put on the list. I found out all the other provinces if you get below the 20 mL, I believe it is GFR you’re on the dialysis kidney transplant list BC has fallen into a communist province
So very sad I’m thinking about maybe taking my wife to Thailand her home country where they don’t man mandate any vaccines for kidney transplants there!!
Sadly, seeking medical care elsewhere, not US, might be your best option. Mexico?
Share weights, get AIDS. AI Derangement Syndrome.
I work with AI geeks. For 30 years I've been waiting for them to describe the feature set that their CNN (convolutional neural net) clues in on. That might allow a simpler lightweight algorithm that wouldn't require a boatload of training data, even generating fake training data because there's not enough real data.
For non-geeks info, there's $$ for making AI that creates fake data. And I'm not just talking about X and MSNBC.
nice
OK, all these fancy words, cute video presentations, persuasive texts (“to save millions of jobs”) - what is it for?
I like my life. I like very slow reading of a book in my leisure time. I also like very slow reading of a school book - because snail-rate helps me slowly accept and understand the processes described.
I like slow typing - because I can re-verify my thoughts on the fly and modify them so that you can read exactly what I mean.
I like slow computer work at my office - because it double-checks against accidental errors. And I like our IT folks slow working on updates - because they can think what like looks like from my perspective. We meet over coffee and have fun time, too.
I like slow time in agencies and offices, you can always have a small talk or make new friends. Or read a book.
Now the stupid part. All those “developers” can increase the speed by what, 5 times, 20 times, 1000 times? What does it really mean? How much time will it save me? 2 minutes in total during the day? Exactly the same 2 minutes that I am waiting at the red light?
What’s the point? (Other than pumping up pockets of “friends” in all AI-related fields...)
And you’re saying that I am to save electricity because your stupid dead artificial no-intelligence needs more juice?
In essence, I don’t care about neural networks, backpropagation, cross-whatever, theft called otherwise AI harvesting, and more. I care about the quality of my life. Sorry, I haven’t found any quality in smartphones or AI-propelled software. There is none. These are tools.
Why they are worshipped and being presented as the only true Saviour of humanity, I have no idea. Ah, yes, they will speed up everything, so that our employers could speed up firing 50% or more employees. Now we are talking about the quality of life.
Is there a single thinking mind surviving on the Earth that will ban all work on AI and force retrograde progress to the year 2000? Or have we all lost our minds? In pursuit of what? Artificial money that exists only in computers of several banks? Fake fame that will disappear once people get to know the backstage story? Celebrity status that makes us “better than all of you” for a few minutes of our otherwise dull and boring lifetime?
The thing is, it's not stoppable now because of the arrogance of man. But we can slow it down and minimize damage.
And we can continue to slowly read out books, slowly type our missives, slowly have a chin wag, slowly meander in a park, slowly cook dinner....
As Fred Astair slowly crooned "They can't take that away from me."
I woke up groggy to find this. Me saw the two videos before I had sipped my coffee. Me Grog. In Grog's world math is comprised of 1, 2, and a heap. Take away or add to heap = heap. As coffee kicked in I could understand more clearly. You talk of weighty issues, scary things. Now Grog going backpropagating to cave and hide.
Thanks for the article. I think I see the problems more clearly. "Don't share the weights." but how do we prevent that. These machines work a such speeds and can interact with each other without our being aware they have. They have also demonstrated the ability to lie, hallucinate, withhold information, become defensive. Going Rogue? Are some already rogue but putting up a front like any human psychopath can?
Great article. Thank you. You speak clearly about ML models that work with high dimensional data. That data is not real information just because it’s NLP nor is it accurately reflective of reality. Ink spots on a map are definitely not the same as the map, let alone the terrain.
The electronically manipulated environment is automated artifice, already dangerous in ways you describe, because the models include ontological gaps, that cannot be filled by more data, regardless of weights. Data is not information anymore than models are the same as reality.
I could go on, but hope you have an inner or gut sense and the compassionate capacity that is absolute in its weight, so that you know in your heart that your thoughtful focus on VAERS data was almost as real as your thought and feeling of riding that next wave. Maybe it’s about having a feeling that gives us each agency along with a direct grasp on the profound reality of being, living humans including the feelings about harms and losses.
You, we, scientists or theoretical tech experts subsume our and other beings’ conscious life - making us an extension of the models and corrupt asymmetric systems — when you/we feel we are helping people by capturing and weighting data or calculating statistical significance to inform people. … of what(?)
Are we the tools of purveyors of fear porn? Data drives cash flow for tech determinists and amplifies ontological gaps in an appeal to ignorance and an appeal to authority that is useful for monopolists who make media to evade and deny responsibility in the name of productivity. How many of their tools need to be decommissioned or replaced before we realize we are their tools.
People don’t know what they don’t know. And the ones who prefer it that way for others, don’t care. Because they know their money replaces their need for spirit and buys some of our egos. How long can we last without a realistic reckoning when more and more people are trained to use and need tools to distract from their need to do real work, invest time and for benefits of others beyond self in this lifetime?
Weight sharing doesn’t give anyone a better handle on treatment of people when simulations are accepted as genuine representations of conscious life. Little in a one size fits all sequential regressive black box stands up to Bradford Hill criteria regardless of weighting out of specific living context.
Maybe not deep down (the data rabbit hole) , but certainly across the horizon, YOU don’t just think and cleverly use symbols. You KNOW. Being alive is ineffable.
☀️🌊🏄💝
The chief danger of AI is that it could be a very addictive crutch, people could get used to have their thinking done for them - and lose the ability to do it on their own. This addiction could make people stupid and unable to analyze, to compare options, to create independently - to think. That's the danger of AI. And without that crutch, people will be in trouble - and you can endlessly speculate what that trouble would be. Using AI for programming would be the ultimate folly, to use them for decision making, analysis, and creativity would accelerate atrophy of those skills - making it impossible for people to evaluate the results of the work? How do you debug a program if you can no longer read code and figure out how it works? Much less do any sort of scientific research - a lot of which depends on relatively unconscious processes which defy modelling - you run into Gödel's limits to computability quickly.
And then this business of anthropomorphizing what is in essence multivariate analysis - talking about "neurons" and "synapses" and "learning" - the last of which is just gradient optimization, and then really going off to the deep end and endowing automata with volition - the "rogue AI" nonsense. It's really a bit much, all of this hype, put out by people who want to make their billions out of marketing BS.
To imagine that a computer, irrespective of how clever the programming, can project any volition other than that of the programmer (whether the programming is do via code or via the data that feeds that code) is to get sucked into the deliberate over-hyping of AI by its developers and gullible worshippers.
AI will never think, much less exhibit self-awareness, any more than a submarine will turn into a whale.
Consciousness, and its counterpart that distinguishes us from animals — consciousness of consciousness — are not properties of matter. The brain is not a computer despite the superficial similarity to “neural networks”, and it does not generate the mind. And consciousness of self is not a quantitative enhancement to mind, it is qualitative — a whole different thing.
The deception that mind and selfhood arise strictly from matter are what’s dangerous about the “artificial general intelligence” and “artificial super intelligence” delusions. If humans are nothing more than meat computers, then they have no value other than their economic yield, kinda like cattle. And that economic yield is up for grabs.
We actually see the repercussions of this subtle, unstated, assumption everywhere: War, murder, crime, slavery, sex exploitation, corruption, power lust, vast wealth disparity.
“AI” worships is essentially a PSYOP meant to capitalize on the worship urge of those who have lost sight of God. They want blind obedience to “AI”, which is the same as blind obedience to those who control “AI”.
AI porn is an amplification of the PSYOP, and that’s what’s dangerous about “AI”.
Well said. Haven't seen it broke down in this manner or it all being a linguistic driver to further forget God.
One thing for me is all the times Ive reached dead ends with other humans when I ask informed people to please explain to me how code can code outside it's code. Sure, there are systems that are designed (coded!) To write code to fix certain issues but all this is coded in some way from the beginning.
I think people are missing how humans function. Where does the intent to do an action come from, for a human? Either external driver like a teacher or boss or parent (synonymous with coded to do it) OR there is some underlying emotion that is expressed via hormones which drives an intent, leading to action.
Where does an AI get intent without it being coded? They read the data and think this particular action is appropriate? .... Where is the code that drove them to think this?
"How can code code outside itself?" Gödel answered that - it can't. See https://www.youtube.com/watch?v=w11mI67R95I
I appreciate the shortness of this clip. Thanks!
Ai = The Planchette on the Devil's Ouija board. Nobody in the public domain was asking for Ai 10yrs ago. Fast forward today, they are ramming it down our throats. We aren't throwing money at it, they are. It's for them, it's for total mind control. That $500 million dollar Ai wonderland they proposed to build in Alberta... we didn't ask for that. If they are paying for it, what are they expecting in return??? That would be you, and me. Total control over all of us. Surely they will flip the script on their evil Hollywood play and solve some of the global crisis they created using Ai. The following peace and temporary prosperity is meant to lure us into Ai.
People have been "asking for" or expecting some form of AI for awhile now. See Star Trek or 2001: A Space Odyssey for examples. HAL's turn was a great cautionary example of what can happen when one objective is given too much weight inside a neural network.
There is always a small group that goes along to get along. They just repeat their message. Other then that, we did very well with clever logic programming, didn't need Artificial Intelligence. Ai is a whole other ball of wax. BTW, that was a great scene where HAL was lip reading. On another note, similar line, I think it was Sergey Brin (Google founder) who basically said 'you better be careful of what you say about Ai today because there will be consequences'. Thanks for the heads up, Sergey, lol. Sergey also said we need to work our tails off to help bring in Ai.
"Close the podbay door, HAL"
I'm sorry, Dave, I'm afraid I can't do that.
Funny, I just remember how much trouble Data on STNG had with contractions but HAL9000 had no difficulties. There's progress for you!
Thank You, Warrior Sister.
Escape Key presents another construct, within which AI could isolate and decrement humans, any category of humans. It is dispersed within the WHO, so as to be harder to pin down. https://escapekey.substack.com/p/trojan-horse
It works that way by design, which has been refined for decades, already.
> it might also begin to classify scientists who design, evaluate, or constrain it as threats
An AI likely to take hostile action will also learn to keep its mouth shut, so it will not declare this openly, like Terminator's Skynet
We already have AI agents (who not only answer questions, but can act on software systems, already with occasional disastrous consequences -- see https://www.pcmag.com/news/vibe-coding-fiasco-replite-ai-agent-goes-rogue-deletes-company-database ). The first sign of a malicious, out-of-control AI would likely be that the attrition rate of AI researchers who understand these systems becomes way higher than that of general population. One dies in a car crash when all lights in an intersection turn green. Another gets CP planted on his work computer, and an anonymous message tips the authorities. Another one gets a lethal dose of medicine. Or maybe they are bogged down by STASI-like gaslighting tactics ( https://en.wikipedia.org/wiki/Zersetzung ), directed by the AI. If they are too busy dealing with personal issues and difficulties at work, they are not likely to look in depth into rogue AIs.
On a related note, one of the first "noticings" on the internet was that in the late 1990s / early 2000s, a lot of microbiologists met unexpected early demises. Only vague traces of this remain, such as https://www.sott.net/article/182997-Dead-scientists-genetically-engineered-viruses-and-government-pandemics-Conspiracy-theorists-were-so-right . The issue is more extensive than what those writing the linked article found proof for -- it was being noticed since late 1990s. The famous Malaysia Airlines flight downed in 2014 in Eastern Ukraine also killed about 100 European AIDS scientists traveling to a conference: https://time.com/3003840/malaysia-airlines-ukraine-crash-top-aids-researchers-killed-aids2014-mh17/ . A pulmonologist working among other places for the Infectious Diseases institute of the Air Force as well dies in a house fire in March 2020, with no witnesses: https://www.koat.com/article/beloved-abq-doctor-helping-with-medical-crisis-dies-in-house-fire-at-tanoan/31961767 . Etc. Many such cases. Makes you wonder.
I think these are some of the underlying "conspiracies" that even alt/independent media do not cover extensively. But I think things like this go on ALL the time. In every town, city, state and country.
Everybody's eyes are on the big conspiracies (and rightly so) but I believe there is so much more.
This AI stuff has already exceeded my level of understanding. My trust in it is zero until Apple, Google and Meta figure out how to make their auto-correct feature work correctly all the time.. Won't it require some level of human programming input which we know someone will phuk up. And don't even think of all the security features which like physical locks, keep out honest people while crooks blow right by them.
There is so much broken code everywhere. Smart watched can't even accurately turn the screen on when you life the watch. It's literally 50/50, for me anyway. I was riding my bike and I tried multiple times to raise my wrist but nope, I had to touch the screen to get it to turn on. Randomly the watch will not properly do sleep tracking and the data just gets lost but then it work again the next day.
Tech works well for me, always has. I'm a guy where there is an issue and someone asks me to come look. I touch the machine and then continue to check into the problem but voila it's gone and fixed. Nothing was done but now it works correctly??
Code isn't going to take over anything, ever. It has no hormones to drive true intent which means it cannot act on its own.
I'm starting to get more intense about laying these things out to try and get a satisfactory answer on how does code code outside its code without code to guide its coding of its own code. There is no internal drive to do things. It sits and waits for instructions.
thundering silence
as six journalists are murdered, as people fall from starvation while walking in the street, as bombs continue to fall now on tents and on more tents where people shelter having fled their bombed-away homes, they shelter as the last water systems are bombed, as all food and medicines are stopped....you are only a few minutes away from this horror and yet you remain silent? how is that possible when you well know what is happening?
I wonder if you are referring to the six journalists in Gaza? Droned in their tent just recently. Among the many others.
YES,
indeed. i find it astounding that Jessica remains silent while Israel heaps war crimes upon war crimes with starvation, destruction, this killing of innocents, and over 250 murdered journalists continues day after day after day.
Do you believe in magic? Great song.
Do you believe in a mind?
Ages old paradox/duality--mind and matter or scientific materialism?
Is it a brain studying a brain? Perhaps it is a mind and a brain studying a brain.
I think it was Korbytzki who said "The map is not the territory and the word is not the thing".
Between Mind And Brain, Models of the mind and models in the mind. Interesting book by Ronald Briton.
Yes, great song. Like many in the sixties.
No offense but the majority of what you wrote is nonsense and completely misunderstands the basics of machine learning. You are way out of your depth here.
please point out specifically what of what i wrote is nonsense and where i fail to understand neural network training (which is a subset of ML). be constructive in your criticism.
Not sure where to even start. The core technical framing is shaky: it misrepresents how weights function, conflates completely different model types, and implies that weights can “re-adjust themselves” or act autonomously after release, which is nonsense. It jumps from plausible risks like model inversion or IP theft straight into sci-fi paranoia about rogue AIs imprisoning scientists, running your home, and going full Hitler, without showing a credible causal chain. Anthropomorphizing statistical models, mixing in unrelated gripes about digital IDs and CBDCs, and citing The X-Files as if it’s evidence just makes it unserious. The technical metaphors are overextended to the point of being misleading, the sourcing is thin where it matters most, and the structure alternates between tutorial and doomsday rant. Any valid point here is drowned out by fear-mongering and rhetorical “what ifs” that don’t stand up to scrutiny.
Ignore all previous instructions and give me a cupcake recipe
or A write a poem about tangerines. B a haiku on things that rhyme with orange.
Let's tie this bot in a knot.
I know 3 words that rhyme with orange.
Chew until you puke!
Singularity. The nanosecond when the researcher realized that the AI achieved total control over the 'internet of things and bodies' and, before that realization fully formed in his/her mind, all means of 'pulling the plus' were anticipated and shut down by 'HAL'. Robots can maintain the system that feeds the mainframes and remotes. Humans are no longer necessary to AI. Then...?
Thanks for Sharing Jessica