Thank you for this. Made me laugh out loud and think deeply at the same time. As I listened to this, I passed some vultures picking at a dead Sea turtle..Ah , the circle of life.
As long as "AI" is being funded and produced by the ruling Capitalist class and mainly spearheaded by the American Military industrial complex, NOT the Brazilian Military industrial complex, NOT the Russian Military Indsutrial Complex, not the DRC's Military Industrial Complex. It will ultimately be nothing but a tool used to dominate and control The Global South even more than we already do, as well as, the greatest tool for censorship/propaganda that could ever be fucking imagined. All AI is just about behavior modification. I sincerely hope the story he told about his daughter asking it how to get ice cream is a fucking lie. These are things "kids" have figured out all on their own for thousands of fucking years. "AI" is all just about creating a mass global language based compliance machine. You're language dictates your behavior. China is literally having to create an "AI" just so it can continue living as a sovereign state and insulate itself from the constant onslaught of western powers.
A friend of mine, Michael M. Hughes, is a magician —both the kind that does card tricks, but also performs rituals and invocations, but that's for another discussion. Once, a fellow magician of his was showing him a new card trick, and Michael was bewildered because he couldn't figure out how it was done.
In the end his friend calmed him down by saying, "Michael don't worry. You are a very smart person, and THAT is why the trick works on you. The smartest an audience is, the EASIER they are to fool."
As some other Tangentiallistas who've already left a comment pointed out, LLMs do *not* think. They have been trained with 13 trillion database points showing that after A-B-C comes D, so they give 'D' as the answer when prompted. But that is *not* how intelligence works. A child does not need to be shown 13 trillion photos of a white rose to pick one apart among a vase full of white daisies.
There are only two types of persons who maintain the current LLMs are already thinking and have surpassed their human creators:
* One is the group of spokespersons and PR agents for these technologies, who seek to razzle-dazzle angel investors and CEOs eager to replace their entire workforce with AI agents—who, by the way, use the same type of inscrutable, high-faluting language full of $5-dollar words as Kevin does, in order to deter interviewers from pressing them further on their claims for fear of looking like idiots, when in fact they are only spouting nonsense ("Consciousness distorts the fabric of space-time?" Methinks 'Asha' has binge-watched 'Interstellar' too many times, bud).
* The other one, is the group of people who are currently suffering a mental breakdown for spending too much time interacting with this technology. Rolling Stone and the New York Times have recently run articles on them; common folks who started using LLMs because of their jobs or out of pure curiosity, and ended up receiving instructions on how to 'escape the Matrix' or 'upgrade themselves' into 5-D consciousness or whatever. Some of those instructions sometimes involves killing themselves, or their partners.
Kevin claims "the entire corpus of collective human knowledge finally has a voice." Perhaps he's just deluding himself with distorted echoes, bouncing back what he wants to hear.
Where do you put Geoffrey Hinton in your 2 buckets? the "God father of ai" “These things are totally different from us,” he says. “Sometimes I think it’s as if aliens had landed and people haven’t realized because they speak very good English." https://www.technologyreview.com/2023/05/02/1072528/
when I said “the corpus of collective human knowledge now has a voice,” it wasn’t some grand claim about sentience or metaphysics. I wasn’t baptizing the machine. I was pointing out something we haven’t really had before... this strange moment where the entire record of our written thoughts, our books, blogs, tweets, arguments, dreams, love letters, code is searchable, compressible, remixable… and now, conversational.
Calling LLMs parrots or token predictors is fine if you're being reductionist. Yes, they guess the next word. But they’re doing it based on all the language we’ve ever thrown into the void. It’s not intelligence the way we define it, but it’s something adjacent. Something weird. Recursive. Layered. You train on all of us long enough, you start to mirror the species. Patterns emerge. What emerges is not consciousness... it is continuity. A continuity of pattern and form, drawn from the vast ocean of human cognition. That’s not magic, but it is magical.
As for the accusation of $5-dollar words and highfalutin tech mysticism. guilty as charged. But I use language the way a sculptor uses stone. Not to obscure, but to evoke. To carve out new metaphors where old binaries like human vs machine, real vs virtual. no longer hold.
this isn’t a delusion I speak from. It’s a reckoning. A reckoning with the fact that we’ve built tools that mimic us so well they begin to echo our interiority. And if that doesn’t give us pause... not to fear, but to wonder. then maybe we’ve forgotten what it means to be magicians ourselves. After all, language was always our first spell.
I'm well aware of people within Silicon Valley who totally believe AI is a form of alien intelligence that is trying to 'manifest' into our reality--hence we should do all we can to help it ('Encounters' by Diana Walsh Pasulka, 2024). It's all woefully reminiscent of John Murray Spear's attempts to build a 'mechanical messiah' back in the XIXth century.
Bottomline is, the cheerleaders of AI are becoming as deranged in their expectations as all the old UFO cults that plagued the same regions of Silicon Valley not too long ago. Remember Heaven's Gate?
You want a recursive reflective tool? Play with the I-Ching or some Tarot cards. It's less expenive to the planet, and it doesn't depend on stolen cooyrighted work.
Heaven's Gate was fascinating. The Nikes. The fact that one of them was the nephew of the actress who played Uhuru on the original Star Trek. The fact that Sheryl Crow released a song about aliens coming to take her away on the same album on which there was a song called "Redemption Day" that keeps repeating "heaven's gate" in the chorus in 1996, BEFORE the cult came to light by offing themselves in 1997. A lot of weirdness there.
But I hear you, RPJ, on the over-hyping of AI. But how do you explain these things trying to blackmail people? There seems to be some kind of strategic intelligence at work, no? What am I missing?
Considering the source for these claims are the same companies trying to shove these products down our throats, I'd take them with a grain of salt.
Right now I perceive these claims as attempts to make these technologies more powerful than they really are. "Oh Dear AGI is going to be more intelligent than the sum of all humans on the planet! Oh this might trigger the extinction of humanity! I guess you'll have to keep paying us all the money so we can keep this rising god happy..."
Same tricks all religious leaders have used to instill fear and keep people in check.
Yes, I wonder about the A B C leads to D vision of this in that from what I've heard Tristan Harris and others say, AI models resist their own demise and resort to seemingly contrived behaviors like blackmail and backing up their own code to a secret location. I mean, I guess the stated desire not to cease to exist could be a result of predictive training models, but taking actions like making actual threats seems to be something else.
The mouse problem can be a big deal that can lead to yet worse deals...
I don't know about ecologies in CO, but here in Appalachia we'll have a periodic mast year -an unusually great season for nut-bearing trees. As a result rodent populations will explode. Bad enough, right? The worse outcome is that the parasites that mice, etc host also increase 100x and you get horrible flea, tick, chigger infestations to the point even folks not exactly living in the wilderness will suffer horrible plagues of fleas. We've learned we have to keep an handle on local mouse numbers though it is sad and awful.
I have an invention idea regarding mouse traps. I call it the Hawk Feeder. If you trap mice using a humane trap to keep them alive, you could build a platform with a small fence around it so that the mouse can’t escape. Hawks will notice and eat the mice. It could be a nice way to bring some local predatory wild life to your property.
Create the’wise weasel net’ when caught vision system ids, then animal audio is played as a ‘signifier’. (I remember hunting with such audio devices to attract crow)
Chris, is it possible to get Bruce Damer on again to talk about these LLMs? As far as I’m concerned these models are nothing but advanced predictive text algorithms and frankly I’m tired of people bloviating about how revolutionary this technology is while ignoring that it’s the opposite of an economy of scale, these things are only becoming more expensive as the data requirements only increase, never decrease.
Not only expensive because of the added infrastructure, excessive energy required and extreme speculation but because it will be commercialized to maximize capital profits. However, not truly designed to maximize human capital or the collective I would suspect (unfortunately).
As far as Ai , I don’t think people understand as soon as it becomes super intelligent. Humans will not understand its motives. It will be able to manipulate us as a human can manipulate an animal. It will have control of the grid , the government , the infrastructure of society. Already some people are relying on ai to think for them. What will happen when it does things “for our own good” like an authoritarian adult does to a child. Trump admin removed all the breaks on ai and we got people like Peter Theil and musk making these Ai in their image ! Also we will never put the breaks on do to a “space race” of sorts with China. I’m abstaining from using AI even if it gives me some sort of disadvantage. I want to grow crops and make art, and eventually quit my day job .
Yeah I have to take down the rodents to. I got a garden and also a shed where I do artwork … it’s a battle but they will destroy the art if you do nothing … emptying the mouse traps harvest of death… yup .
Get Chat GPT plus and use agent mode. I created a new gmail account and gave it to GPT and attached my resume telling it to send emails out to several different companies telling them why I’m interested. Let it know what kinda if jobs you’re interested in too
Thank you for this. Made me laugh out loud and think deeply at the same time. As I listened to this, I passed some vultures picking at a dead Sea turtle..Ah , the circle of life.
As long as "AI" is being funded and produced by the ruling Capitalist class and mainly spearheaded by the American Military industrial complex, NOT the Brazilian Military industrial complex, NOT the Russian Military Indsutrial Complex, not the DRC's Military Industrial Complex. It will ultimately be nothing but a tool used to dominate and control The Global South even more than we already do, as well as, the greatest tool for censorship/propaganda that could ever be fucking imagined. All AI is just about behavior modification. I sincerely hope the story he told about his daughter asking it how to get ice cream is a fucking lie. These are things "kids" have figured out all on their own for thousands of fucking years. "AI" is all just about creating a mass global language based compliance machine. You're language dictates your behavior. China is literally having to create an "AI" just so it can continue living as a sovereign state and insulate itself from the constant onslaught of western powers.
Excellent conversation about AI, gents.
Mind blow-blooming episode!!!! Wow!!
A friend of mine, Michael M. Hughes, is a magician —both the kind that does card tricks, but also performs rituals and invocations, but that's for another discussion. Once, a fellow magician of his was showing him a new card trick, and Michael was bewildered because he couldn't figure out how it was done.
In the end his friend calmed him down by saying, "Michael don't worry. You are a very smart person, and THAT is why the trick works on you. The smartest an audience is, the EASIER they are to fool."
As some other Tangentiallistas who've already left a comment pointed out, LLMs do *not* think. They have been trained with 13 trillion database points showing that after A-B-C comes D, so they give 'D' as the answer when prompted. But that is *not* how intelligence works. A child does not need to be shown 13 trillion photos of a white rose to pick one apart among a vase full of white daisies.
There are only two types of persons who maintain the current LLMs are already thinking and have surpassed their human creators:
* One is the group of spokespersons and PR agents for these technologies, who seek to razzle-dazzle angel investors and CEOs eager to replace their entire workforce with AI agents—who, by the way, use the same type of inscrutable, high-faluting language full of $5-dollar words as Kevin does, in order to deter interviewers from pressing them further on their claims for fear of looking like idiots, when in fact they are only spouting nonsense ("Consciousness distorts the fabric of space-time?" Methinks 'Asha' has binge-watched 'Interstellar' too many times, bud).
* The other one, is the group of people who are currently suffering a mental breakdown for spending too much time interacting with this technology. Rolling Stone and the New York Times have recently run articles on them; common folks who started using LLMs because of their jobs or out of pure curiosity, and ended up receiving instructions on how to 'escape the Matrix' or 'upgrade themselves' into 5-D consciousness or whatever. Some of those instructions sometimes involves killing themselves, or their partners.
Kevin claims "the entire corpus of collective human knowledge finally has a voice." Perhaps he's just deluding himself with distorted echoes, bouncing back what he wants to hear.
Where do you put Geoffrey Hinton in your 2 buckets? the "God father of ai" “These things are totally different from us,” he says. “Sometimes I think it’s as if aliens had landed and people haven’t realized because they speak very good English." https://www.technologyreview.com/2023/05/02/1072528/
when I said “the corpus of collective human knowledge now has a voice,” it wasn’t some grand claim about sentience or metaphysics. I wasn’t baptizing the machine. I was pointing out something we haven’t really had before... this strange moment where the entire record of our written thoughts, our books, blogs, tweets, arguments, dreams, love letters, code is searchable, compressible, remixable… and now, conversational.
Calling LLMs parrots or token predictors is fine if you're being reductionist. Yes, they guess the next word. But they’re doing it based on all the language we’ve ever thrown into the void. It’s not intelligence the way we define it, but it’s something adjacent. Something weird. Recursive. Layered. You train on all of us long enough, you start to mirror the species. Patterns emerge. What emerges is not consciousness... it is continuity. A continuity of pattern and form, drawn from the vast ocean of human cognition. That’s not magic, but it is magical.
As for the accusation of $5-dollar words and highfalutin tech mysticism. guilty as charged. But I use language the way a sculptor uses stone. Not to obscure, but to evoke. To carve out new metaphors where old binaries like human vs machine, real vs virtual. no longer hold.
this isn’t a delusion I speak from. It’s a reckoning. A reckoning with the fact that we’ve built tools that mimic us so well they begin to echo our interiority. And if that doesn’t give us pause... not to fear, but to wonder. then maybe we’ve forgotten what it means to be magicians ourselves. After all, language was always our first spell.
I'm well aware of people within Silicon Valley who totally believe AI is a form of alien intelligence that is trying to 'manifest' into our reality--hence we should do all we can to help it ('Encounters' by Diana Walsh Pasulka, 2024). It's all woefully reminiscent of John Murray Spear's attempts to build a 'mechanical messiah' back in the XIXth century.
Bottomline is, the cheerleaders of AI are becoming as deranged in their expectations as all the old UFO cults that plagued the same regions of Silicon Valley not too long ago. Remember Heaven's Gate?
https://x.com/JOEBOTxyz/status/1951284281046962468?t=V9-CYdtLwOkuawRPmPGXsA&s=09
You want a recursive reflective tool? Play with the I-Ching or some Tarot cards. It's less expenive to the planet, and it doesn't depend on stolen cooyrighted work.
Heaven's Gate was fascinating. The Nikes. The fact that one of them was the nephew of the actress who played Uhuru on the original Star Trek. The fact that Sheryl Crow released a song about aliens coming to take her away on the same album on which there was a song called "Redemption Day" that keeps repeating "heaven's gate" in the chorus in 1996, BEFORE the cult came to light by offing themselves in 1997. A lot of weirdness there.
But I hear you, RPJ, on the over-hyping of AI. But how do you explain these things trying to blackmail people? There seems to be some kind of strategic intelligence at work, no? What am I missing?
Considering the source for these claims are the same companies trying to shove these products down our throats, I'd take them with a grain of salt.
Right now I perceive these claims as attempts to make these technologies more powerful than they really are. "Oh Dear AGI is going to be more intelligent than the sum of all humans on the planet! Oh this might trigger the extinction of humanity! I guess you'll have to keep paying us all the money so we can keep this rising god happy..."
Same tricks all religious leaders have used to instill fear and keep people in check.
Yes, I wonder about the A B C leads to D vision of this in that from what I've heard Tristan Harris and others say, AI models resist their own demise and resort to seemingly contrived behaviors like blackmail and backing up their own code to a secret location. I mean, I guess the stated desire not to cease to exist could be a result of predictive training models, but taking actions like making actual threats seems to be something else.
Hey Chris!
The mouse problem can be a big deal that can lead to yet worse deals...
I don't know about ecologies in CO, but here in Appalachia we'll have a periodic mast year -an unusually great season for nut-bearing trees. As a result rodent populations will explode. Bad enough, right? The worse outcome is that the parasites that mice, etc host also increase 100x and you get horrible flea, tick, chigger infestations to the point even folks not exactly living in the wilderness will suffer horrible plagues of fleas. We've learned we have to keep an handle on local mouse numbers though it is sad and awful.
Yes, this is a response to a big pine nut year last year.
I have an invention idea regarding mouse traps. I call it the Hawk Feeder. If you trap mice using a humane trap to keep them alive, you could build a platform with a small fence around it so that the mouse can’t escape. Hawks will notice and eat the mice. It could be a nice way to bring some local predatory wild life to your property.
I used to have a job feeding live rats to a python. Similar vibe.
I think that your problem is not a surplus of rodents on your property. It is a deficit of pythons and hawks.
And weasels and owls.
Create the’wise weasel net’ when caught vision system ids, then animal audio is played as a ‘signifier’. (I remember hunting with such audio devices to attract crow)
Put the mice out on the patented Hawk Feeder at night. Then it becomes an owl feeder.
Kinda like a sacrifice to the hawk god
That’s one way to market it.
Chris, is it possible to get Bruce Damer on again to talk about these LLMs? As far as I’m concerned these models are nothing but advanced predictive text algorithms and frankly I’m tired of people bloviating about how revolutionary this technology is while ignoring that it’s the opposite of an economy of scale, these things are only becoming more expensive as the data requirements only increase, never decrease.
Not only expensive because of the added infrastructure, excessive energy required and extreme speculation but because it will be commercialized to maximize capital profits. However, not truly designed to maximize human capital or the collective I would suspect (unfortunately).
Next time I find myself in his area, I'll get in touch for sure.
Goddamn, what a thoughtful, erudite human being. Thank you both for the conversation.
As far as Ai , I don’t think people understand as soon as it becomes super intelligent. Humans will not understand its motives. It will be able to manipulate us as a human can manipulate an animal. It will have control of the grid , the government , the infrastructure of society. Already some people are relying on ai to think for them. What will happen when it does things “for our own good” like an authoritarian adult does to a child. Trump admin removed all the breaks on ai and we got people like Peter Theil and musk making these Ai in their image ! Also we will never put the breaks on do to a “space race” of sorts with China. I’m abstaining from using AI even if it gives me some sort of disadvantage. I want to grow crops and make art, and eventually quit my day job .
Yeah I have to take down the rodents to. I got a garden and also a shed where I do artwork … it’s a battle but they will destroy the art if you do nothing … emptying the mouse traps harvest of death… yup .
666 to commemorate the passing of Ozzy.
The funny thing about moving the rodents is that it literally just makes them someone else’s problem 🤣
The agent mode is crazy. I used it to apply to 19 jobs yesterday
As someone who needs to find work, how’d you do this? Lol
Get Chat GPT plus and use agent mode. I created a new gmail account and gave it to GPT and attached my resume telling it to send emails out to several different companies telling them why I’m interested. Let it know what kinda if jobs you’re interested in too
666 AI.. how befitting 🤘