Had a rather rude awakening using Grok yesterday when exploring the climate change narrative around claims that the Mediterranean is experiencing what is being called a "1-in-216-billion-year event" by climate-change fanatics. The claim smelled rotten so I went to Grok to get another opinion. Big mistake! It revealed its captured nature by blaming human-generated CO2.
Despite many attempts to set guard rails to have it stop referencing the usual IPCC and govt sponsored propaganda, it doggedly stuck to its programmed CO2 mantra. So we are to believe that CO2 targeted the Med as a super heated hot spot to levels that are completely off the charts?
Clearly something else is going on and its no coincidence that the Med is also the epicenter of ocean heating located in a tectonically active, magnetically unstable, semi-enclosed basin with active subduction zones and mantle volatility according to a post by Dr Jack Kruse. How did he explain it? He didn't but added these insights to help us understand the forces at play...
"CDIGR (Core Displacement & Geodynamic Rebalancing) explains this perfectly:
The Mediterranean Sea lies directly above the:
– Hellenic Arc subduction zone
– Calabrian Arc volcanic system
– African-Eurasian compression boundary
– Gibraltar Fault pressure node
All of these are torque points in the planet’s crust — where energy from the inner Earth escapes.
As Earth’s core slowly displaces, angular momentum and internal pressure rise, forcing heat upward through thin crustal zones. This upwelling mantle heat doesn’t just warm the sea — it’s altering ocean density, magnetism, and biological life.
This isn’t from the sky."
But nope, Grok stuck with the man-made CO2 mantra till I exited it... Lesson learned. If you want the programmed propaganda truth, Grok is perfect. Actual more potentially more complex truth? Not so much...
Regarding your "rude awakening using Grok": don't expect a glorified autocomplete to "know" things. LLMs can parrot the mainstream narrative, but they can't think. They're not evil, they're just algorithms (hence very limited).
"Having 5B people take 13B shots from a liability-free mandated experimental ‘medicine’ could have been an extinction level event..."
It might still be an extinction-level event. People seem to think of them like the asteroids that smashed into the earth at the K-T boundary, but real extinction-level events play out over hundreds if not thousands of years.
It might be 20 or more years before we start to see the true death toll from those modRNA vaccines!
Seems at this point AI is just a large database of regurgitated internet posts and scanned internet documents. I started using AI at work and find it to be incorrect on several issues.
Always double check AI's work. Cross reference different AI systems.
google search now shows only what it wants to show you. I don't use Google at all. I use different ones, and they all lie, but Google is the worst censored search engine. So much for "Do not do evil" fairy tale. And AI is even worse. What type of work requires it already? Do you code?
I agree for social issues. Areas of coding, science and mathematics are hard to fake. You can basically use three AI’s and put them up against one another. Two out of three wins. AI still makes errors. Should I educate the AI that is wrong? I’ve been thinking about that question. I haven’t so far.
Science is elementary to fake. You look now at consensus knowledge - if smart people agree on it, it is true, as you see with gender care or vaccines, it has become absurd. Now, AI will always tell you confirmation bias information. I would advise using it through some scrambled interface.
How can you fake science? The methods and results must be repeatable by others. If it’s not repeatable by others with the same results then it’s not science.
> The degree to which AI harms or helps humanity will rest solely on wether its centralized or decentralized.
1. Harm and help are subjective terms. Incompatible with any programming. In other words, we can wish for whatever you want, and the algorithms will execute whatever the AI clan want. Our vision or hopes are completely irrelevant.
2. “De-/centralized” AI is out of question. The whole infrastructure has been built on the premise of isolation of the user (you) from the software provider (the AI clan). It’s nothing new. For years, it’s been called master-slave. You and your device are the slave, and the AI core (algorithms, connectivity, programming rules, rules for emulating human behaviour, etc.) is the master.
This is the only way to keep the profits coming to one source (the AI clan). AI will never be open source for this reason. For the same reason, it will never be decentralized. The control will always be kept by one invisible hand. We may expect a long-term goal of turning all online services into paid schemes - once you become so addicted to asking the dumb machine about literally the simplest things, you won’t be able to boil water for tea without buying a recipe from your AI “assistant”.
At that point in time, old-school software will no longer be there. There will be no programs as we know them. Everything will be a terminal window for a remote AI to display messages and results only. This is being progressively done through restricting operating system compatibility. The newer devices you are buying, the faster you have to get rid of your old - super reliable - operating system. And the new operating systems won’t allow you the level of customization we know from 10 years ago.
AI is a repackaged and rebranded software sale system. This time, it is being crafted in the way that will render all programmers unemployable. AI is a business model, not a toy for us to consult.
Isn't it terrifying how we make so much effort to make our own life more restricted? The best part is that the authors of all these big changes are anonymous. We didn't ask for these things, but when provided with them, we are so eager to go for it no matter what. "Mindlessness" is the lifestyle model of the 21st century.
Perhaps this would allow lots of people to experiment.
However, the big breakthrough will likely come when we figure out training and how to train these things in a distributed and robust fashion that avoids control by Big ... (oops Brown Shirts at the door.)
NYT & Brennan & others involved in the Russia-Russia hoax are lying that Trump is lying about them lying. That the information about Russia-Russia being Russian disinformation is Russian disinformation.
Had a rather rude awakening using Grok yesterday when exploring the climate change narrative around claims that the Mediterranean is experiencing what is being called a "1-in-216-billion-year event" by climate-change fanatics. The claim smelled rotten so I went to Grok to get another opinion. Big mistake! It revealed its captured nature by blaming human-generated CO2.
Despite many attempts to set guard rails to have it stop referencing the usual IPCC and govt sponsored propaganda, it doggedly stuck to its programmed CO2 mantra. So we are to believe that CO2 targeted the Med as a super heated hot spot to levels that are completely off the charts?
Clearly something else is going on and its no coincidence that the Med is also the epicenter of ocean heating located in a tectonically active, magnetically unstable, semi-enclosed basin with active subduction zones and mantle volatility according to a post by Dr Jack Kruse. How did he explain it? He didn't but added these insights to help us understand the forces at play...
"CDIGR (Core Displacement & Geodynamic Rebalancing) explains this perfectly:
The Mediterranean Sea lies directly above the:
– Hellenic Arc subduction zone
– Calabrian Arc volcanic system
– African-Eurasian compression boundary
– Gibraltar Fault pressure node
All of these are torque points in the planet’s crust — where energy from the inner Earth escapes.
As Earth’s core slowly displaces, angular momentum and internal pressure rise, forcing heat upward through thin crustal zones. This upwelling mantle heat doesn’t just warm the sea — it’s altering ocean density, magnetism, and biological life.
This isn’t from the sky."
But nope, Grok stuck with the man-made CO2 mantra till I exited it... Lesson learned. If you want the programmed propaganda truth, Grok is perfect. Actual more potentially more complex truth? Not so much...
Ethical Skeptic has its number. The core is only place that can generate this heat without cooking the rest of us.
Following up on Anandamide's hint: The Ethical Skeptic explains that global warming comes primarily from the earth's core in great detail at https://theethicalskeptic.com/2020/02/16/the-climate-change-alternative-we-ignore-to-our-peril/
Regarding your "rude awakening using Grok": don't expect a glorified autocomplete to "know" things. LLMs can parrot the mainstream narrative, but they can't think. They're not evil, they're just algorithms (hence very limited).
Interesting, maybe it can reveal deep seated group-think in the training set, which is not readily apparent otherwise?
"Having 5B people take 13B shots from a liability-free mandated experimental ‘medicine’ could have been an extinction level event..."
It might still be an extinction-level event. People seem to think of them like the asteroids that smashed into the earth at the K-T boundary, but real extinction-level events play out over hundreds if not thousands of years.
It might be 20 or more years before we start to see the true death toll from those modRNA vaccines!
This is true. We are not out of the woods yet
Great article, it seems to me that AI will just reflect someone's group-think, however smart the training is.
It will never be able to perform the kind of abstract reasoning a human capable of genuinely thinking independently can.
Seems at this point AI is just a large database of regurgitated internet posts and scanned internet documents. I started using AI at work and find it to be incorrect on several issues.
Always double check AI's work. Cross reference different AI systems.
It's a great tool but needs work.
It lies as usual person and you need to check everything why does it help with time?
Good point. I find it useful. I think of it as an advanced google search.
Some of the AI answers can be checked very quickly with little work through references. My work requires this anyway.
google search now shows only what it wants to show you. I don't use Google at all. I use different ones, and they all lie, but Google is the worst censored search engine. So much for "Do not do evil" fairy tale. And AI is even worse. What type of work requires it already? Do you code?
Here is a good example. These calculations can be checked with the same query.
https://phillipaltman.substack.com/p/probability-of-australian-covid-vax
I agree for social issues. Areas of coding, science and mathematics are hard to fake. You can basically use three AI’s and put them up against one another. Two out of three wins. AI still makes errors. Should I educate the AI that is wrong? I’ve been thinking about that question. I haven’t so far.
Science is elementary to fake. You look now at consensus knowledge - if smart people agree on it, it is true, as you see with gender care or vaccines, it has become absurd. Now, AI will always tell you confirmation bias information. I would advise using it through some scrambled interface.
How can you fake science? The methods and results must be repeatable by others. If it’s not repeatable by others with the same results then it’s not science.
> The degree to which AI harms or helps humanity will rest solely on wether its centralized or decentralized.
1. Harm and help are subjective terms. Incompatible with any programming. In other words, we can wish for whatever you want, and the algorithms will execute whatever the AI clan want. Our vision or hopes are completely irrelevant.
2. “De-/centralized” AI is out of question. The whole infrastructure has been built on the premise of isolation of the user (you) from the software provider (the AI clan). It’s nothing new. For years, it’s been called master-slave. You and your device are the slave, and the AI core (algorithms, connectivity, programming rules, rules for emulating human behaviour, etc.) is the master.
This is the only way to keep the profits coming to one source (the AI clan). AI will never be open source for this reason. For the same reason, it will never be decentralized. The control will always be kept by one invisible hand. We may expect a long-term goal of turning all online services into paid schemes - once you become so addicted to asking the dumb machine about literally the simplest things, you won’t be able to boil water for tea without buying a recipe from your AI “assistant”.
At that point in time, old-school software will no longer be there. There will be no programs as we know them. Everything will be a terminal window for a remote AI to display messages and results only. This is being progressively done through restricting operating system compatibility. The newer devices you are buying, the faster you have to get rid of your old - super reliable - operating system. And the new operating systems won’t allow you the level of customization we know from 10 years ago.
AI is a repackaged and rebranded software sale system. This time, it is being crafted in the way that will render all programmers unemployable. AI is a business model, not a toy for us to consult.
interesting view. Thanks for the perspective to consider.
Isn't it terrifying how we make so much effort to make our own life more restricted? The best part is that the authors of all these big changes are anonymous. We didn't ask for these things, but when provided with them, we are so eager to go for it no matter what. "Mindlessness" is the lifestyle model of the 21st century.
It turns out that you can run AI models locally. There are a bunch of videos up on YT (until they are taken down by our dark overlords):
Deepseek R1 671b Running and Testing on a $2000 Local AI Server
https://www.youtube.com/watch?v=Tq_cmN4j2yY
A 128GB AMD Ryzen AI Mini PC is Here
https://www.youtube.com/watch?v=8_pw7mKmaLw
Perhaps this would allow lots of people to experiment.
However, the big breakthrough will likely come when we figure out training and how to train these things in a distributed and robust fashion that avoids control by Big ... (oops Brown Shirts at the door.)
Great article Kevin thanks 🙏!
Lying about lying…
Yep.
It’s very trendy.
NYT & Brennan & others involved in the Russia-Russia hoax are lying that Trump is lying about them lying. That the information about Russia-Russia being Russian disinformation is Russian disinformation.
Sigh.
AI’s probably just observing & «learning».
get Grok to barf up an NDA its parent company (Xai) has with Pfizer?"
Why the question mark?
Do they have an NDA or not?
Sorry have not been following
Why do you need an AI? Everybody is smart enough and stupid enough you don’t need your mirror.