AGI Is DEAD. Long Live AGI!
How “Zombies” and their “Masters” Perpetuate “The AGI Fallacy”
Thanks as always to you – my precious and valuable readers – for visiting the @Lamb_OS Substack. Remember, @Lamb_OS is far more than the name of this phenomenal Substack: It is also – and I swear I’m not making this up – the entire open-source repository of Bill Lambos’ biological Operating System – the very structure and function of the author’s brain and mind. Additionally, I purposely refrained from posting for the last two weeks. I hope you missed me! My reasons were: 1. My (adorable) wife and I, and one of our favorite colleagues (who has requested she remain nameless), were on vacation in Mexico for 13 days. While I wrote much of this post in the (very) early mornings during the trip, I held off posting it because 2. The situation – the embarrassing mess at Open AI – was evolving so quickly that I wanted to see how it turned out before opining on it here. As I can demonstrate below (based on other documented posts below), my predictions as to how that would turn out were spot on (it’s great being smart, but lucky is even better!).
Speaking of smart people (ME), I am far more widely known as the one and only Dr. William A. Lambos: the very highly esteemed computational neuroscientist, licensed clinical neuropsychologist, data scientist, and full stack Python developer whose books, journal articles, Internet publications, GitHub code, and other published works are widely and thoroughly read by anybody who is anybody in AI, AGI, data science, neuroscience, clinical neuropsychology, experimental biopsychology, and even a few folks at DARPA – and should be!
Please Subscribe today if you have not already yet done so. I’m closer than ever to getting that private jet[i], and every new subscriber earns me…$0.00. OK, so the “private jet” thing is – and will sadly forever remain – an enjoyable but utter delusion. After all, this Substack is still free!
[i] Since I know you, my readers, just have to know, on the one in a quadrillion chance that my delusion about my future ability to acquire a private jet are, by some freak of nature, incorrect, I’m thinking of going with a “sensible” Gulfstream G500). Cost? $45 million. What a bargain!
Disclaimer: All the facts stated above are true. However, some adjectives such as “esteemed”, “widely and thoroughly”, “anybody who is anybody,” etc. reflect the opinion of the author and should be interpreted as such.
Disclaimer to the Disclaimer: I hope we can agree that the “Disclaimer” is very close to having outlived its usefulness. Can we not, going forward, be adult about dropping the silliness and pretense, and just say it out loud: The adjectival phrases in the first Disclaimer are all but considered to be Universal Truths. I mean, c’mon, people, we’re talking about ME here – have you even bothered to look at my C.V.? I have more degrees than the Kelvin scale! Plus, I’ve been formally studying, analyzing, and assessing the issues I write about for more than 50 years. But decorum insists I as of the present, this 26th Day of November of the year 2023, and carrying as I do the great weight of a reputation as an objective seeker of unvarnished truths, do openly grant that my opinions may be deemed only partially defensible, and, to come entirely clean, reflect what is a very long stretch. For now.
§
OK, on to the Zombies!
The overwhelming popularity of movies and TV shows such as “The Walking Dead” and “Resident Evil” affirms that American audiences adore zombies. Personally, I never saw the appeal of the genre, but that’s entirely predictable. After all, zombies are not real. I like things that exist and are provably true. Much of the rest of the society in which I find myself does not, however, share my predilection. On the other hand, I’ve always lived outside the cultural mainstream. Is this easy? Not quite. Just ask Galileo how he liked being excommunicated, spending much of his life under house arrest. Or Joan of Arc, who, after conquering France as a teenage girl, was tortured alive as a witch. Being smart is nearly always an asset; being a visionary is more likely to get you burned at the stake!
Nonetheless, most of our fellow humans seem to love zombies. But there is more than one type of zombie: there are, of course, the storybook creatures first described in 1851 by the poet Robert Southey – mythological corporeal revenants created through the reanimation of a corpse.
But the term is also used to refer to false constructs, beliefs, or manipulations that we, as a culture, cannot seem to make go away. The Nobel Laureate Paul Krugman, a liberal-leaning economist and opinion writer for The New York Times[i], may have been the first journalist to use the term “zombie” to refer to describe “ideas that should have been killed by contrary evidence, but instead keep shambling along, eating people’s brains.”[ii]
As an example, in his 2020 book Arguing With Zombies, Krugman argues that:
“The most persistent such zombie is the insistence that taxing the wealthy is hugely destructive to the economy as a whole, so that cutting taxes on high incomes will produce miraculous economic growth. This doctrine keeps failing in practice, but if anything has gained an ever-stronger hold over the Republican Party
There are other zombies, too. If you want a low-tax, low-benefit state, you want to claim that safety-net programs are harmful and unworkable. So, a lot of effort goes into insisting that providing universal health coverage is impossible, even though every advanced country besides the U.S. somehow manages to achieve just that.
You get the idea. But while it’s easy to understand the politicization of tax and spending analysis, why does the politicization extend to areas that don’t so obviously correspond to class interests? Even billionaires need a livable planet, so why has climate change become such a left-right issue? Recessions hurt everyone, so why do conservatives oppose printing money to fight slumps? And why are racial attitudes so closely correlated with positions on taxing and spending?
A lot of the answer is that political players believe—I think rightly—that there is “a kind of halo effect that links all forms of government activism. If people are persuaded that we need a public policy to reduce emissions of greenhouse gases, they become more receptive to the idea that we need public policies to reduce inequality. If they are persuaded that monetary policy can fight recessions, they’re more likely to support policies that expand access to health care.”3
One characteristic of zombies that Krugman does not address, however, is that well formulated zombies have gravitas. They are feared by many people, even as garnering admiration from others for their mystical abilities to cheat death. Should we be surprised by their popularity? I think not. Zombie ideas are, after all, demonstrably powerful demons.
Another aspect of the metaphor implied by Krugman, but not (to my knowledge) specifically stated, is that Zombies do not “self-reincarnate”. Zombie ideas exist only because they have what I call “masters”. So, it is natural for a reasonable observer of this cycle of death and revival to wonder what the point of the exercise is – i.e., why make the effort is made to revive a debunked idea whose time has passed?
The answer is that the zombie is rendered undead only through the ministrations of the zombie master, some evil magician who revives the defunct precept in pursuit of a goal that cannot otherwise be attained. The zombie master keeps alive the undead ghoul (repeatedly, if necessary), the carcass of a dubious construct that should have died if it were judged only upon its relationship to facts or logic. And since zombie masters have self-serving agendas for keeping false - and harmful - ideas alive, constructs that would otherwise have stayed dead are resurrected allowing more taunting, or, more seriously, harming society at large. And this is crucial for the zombie-master metaphor: The zombie masters are the people, entities, and political or social movements who, recognizing that certain dubious ideas which have been broadly rejected by a majority of the citizenry (because of their demonstrated inadequacy), must repeatedly be reanimated on the sly. Zombie masters, you see, don’t care about trivialities such as the veracity of assertions, or the consequences of bad ideas. They care only about advancing their own agendas.
The cleverest of zombie-master teams are those manage to sound convincing because they evoke strong negative emotions. Such appeals to anger can persuade people to suspend their better judgments, because many negative emotions arouse a sense of indignancy. If you want to motivate someone to act in anger – even against their own interests – you need only two things: First, a “hot button” issue that elicits the indignant anger, and second, a convenient scapegoat on which to focus the associated sense of betrayal. Once these two conditions are met, logic and rational consideration fly out the window, and the Zombie construct awakens from the dead to threaten the populace.
§
After the detailed discussion of zombies and their self-serving masters, it should come as no surprise my favorite zombie is the “AGI Fallacy.” As I’ve demonstrated (I would say “proved”) in my previous posts, the construct of AGI (artificial general intelligence) may be the ultimate zombie regarding in the field of predictive analytics (including machine learning) that currently goes by the highly suspect moniker ‘AI’. Think about it: even if one insists on using the term AI, AGI does not, cannot, and will not, ever exist – at least not in silicon-based computational systems. No one ever need worry about the consequences of AGI because it is a delusion – a logical and empirical impossibility if ever there were one. Yet due to a blend of ignorance, fraud, and hype, we find ourselves in a circumstance in which “the smartest people in the room” are warning us that AGI is an existential threat on equal footing with nuclear Armageddon, unchecked climate change, or the consequence of the ultimate pandemic pathogen. Am I really the only informed person who recognizes these assertions as so patently phony as to make us all look like fools?
What may seem even more absurd, I further beg the reader to recognize that I want nothing more than to be proved the bigger fool! My proudest accomplishment would be to find the means put the AGI Fallacy Zombie to death once and for all. If, despite the overwhelming evidence to the contrary citied in the links above, anyone can prove that AGI is possible now – as in today and not at some future date (which of course will never arrive) – then I enthusiastically invite you to make your case with the greatest possible dispatch. If any person can do this – prove that AGI is not only possible but represents an existential threat – I promise I will make us both billionaires within two weeks! Look at what has been going on with OpenAI! As my esteemed colleague Devansh reported just yesterday, the OpenAI debacle of the previous two weeks is a boring distraction of almost no value to the field. Who needs parody when we have the real Silicon Valley?
Although I have no desire whatsoever to become a billionaire (I’ll leave that to Gary Marcus), I have deluded myself into thinking I want a private jet, even more than I think I want the latest Apple MacBook Pro M3-Max (that I have no demonstrable need for either of these baubles matters not a whit – I still want them!). So, if AGI can be proved to be achievable (again, now, today), then I’ll be ordering that aircraft in less than a week. I am willing to die on this hill, but I have no reason to fear my demise. I am, sadly, resigned to the reality that AGI is not only a Zombie, but one of the most ridiculous ones I’ve ever seen. Which is saying something! I strongly advise my readers to similarly reject such patent absurdities.
Getting back to the AGI Fallacy zombie, this undead creature appears to answer to two different masters. The first (and to me the more disturbing – more on that in a bit) – is the wildly underestimated agenda of those who control the AI industry to defend the status quo, a “belief system” that defines the sector. The core belief is that AGI is not only possible, but virtually around the corner. And this is worse than embarrassing. It is prima facia evidence that AI, as a field, is in its dark ages. If so, then the entire field of AI is basically at the equivalent of the flat earth stage of astronomy. Sun vs. earth at the center of the solar system? What’s a “solar system? I ask anyone who has the power to influence the tenets of AGI to allow some “adult” into the room – and quickly! I can’t speak for you, my readers, but the entirety of AI as a field is premised upon a heretofore unrecognized and stunning level of ignorance with respect to its own subject matter (the nature of intelligence).
It has become painfully obvious that AI company leaders are unwilling to undertake the minimal investigative efforts required to test their assumptions about the present and future capabilities of their own intellectual property. Until recently, I assumed that this was a result simple inertia: you know, go with the flow. If so, the AGI Fallacy is largely the result of subject matter ignorance, perpetuated by norms that enforce the inattention to, and a rejection of, knowledge that is readily available in related fields. As I’ve discussed in some detail in my first post on this substack, this willful ignorance is more common in many areas of inquiry due to “Balkanization”.
§
Unfortunately, I am beginning to believe that this is only half the story. I am finding it increasingly harder to defend the assertion that none of the smartest minds in what is inarguably the “hottest” field in Big Tech have been “playing” their less attentive competitors with regards to AGI. Just last week I wrote a post on LinkedIn in which I proposed – being purposely “tongue in cheek” about it – that Elon Musk had been “playing” (as in skillfully and almost cruelly manipulating) Sam Altman and Microsoft for more than a year now. I just checked, and it has been received all of 205 views. But now I’m starting to feel like an undiscovered genius! As any of my regular readers are aware, I often joke that I am, indeed, just such an overlooked intellectual powerhouse, but until yesterday there was no evidence that I might be right! (I hate to admit it in this Substack – in which I intentionally make outrageous assertions about my intellectual superpowers, fully in jest, in order to entertain my readers – but in truth I never have, and never likely will, see myself as any such individual; if you don’t like it when I show a modicum of honest humility, I hereby apologize, and promise to keep it to an absolute minimum going forward.
So, if I am onto something here, the AGI Fallacy zombie is forcing us to grapple with not one, but two equally unappetizing Zombie Masters. Master No. 1 is some hybrid of hubris and a lack of any intellectual rigor. Perhaps 90%. of the “geniuses” in the field of AI are, in a disturbing way, proving themselves to be simpletons, at least with respect to their powerlessness over the AGI Fallacy Zombie. Two, the remaining 10% are very much aware of the fallacy, but perpetuate it because, again, serves them a valuable purpose. I am increasingly gob-smacked by this ugly situation then the entire field of AI is nothing less than the worst of embarrassments. I mean, we are basically at the “flat earth” stage. Sun vs. earth at the center of the solar system? What’s a “solar system? Please, PLEASE – allow some “adult” to sound the alarm – and fast! I can’t speak for you, my readers, but the entirety of AI as a field is premised upon a heretofore unrecognized and stunning level of ignorance with respect to its own subject matter (the nature of intelligence). Fraud and/or outright would be exponentially preferable! And yet, this a sober assessment of the status quo of one of my two chosen professions. YIKES!. AGI is not only dead – it was never alive!
But these Planetary Overlords, these Heads of Nations, Titans of Industry, even the “less fortunate” among them such as serial entrepreneurial billionaires, these Very Serious People (to borrow another Krugman catchphrase) ARE DOING NO SUCH THING? Really? I would think that if I were among the Illuminati, I would at the very least act to protect myself and my filthy lucre. I get that the rest of us sub-human scum are not entitled to the temerity to expect these folks to act as if they might owe us some modicum of effort to save humanity. But even so, in the highly improbable case that The Chosen Ones are nothing more than self-serving liars who care only about themselves, would they not at least take steps to save themselves? Like, they’re going to do nothing when next week, our new Robot Overlords form Skynet and begin to exterminate the pitiful and moronic humans that rashly created them? Um…no. Hmm. To quote Timon (the cheeky meerkat from The Lion King), “and everybody’s OK with this?” Of course not! DUH!!!
This is the second AI Zombie. I propose that more than a few of the members of the AI Brain Trust now screaming that the sky is falling because of the danger posed by AGI – and I admit this is giving them a lot of credit – do, in fact, know that AGI is pure fiction. But if that fiction can create fear, then it can be monetized. And who are the two most likely of these candidates? If I just, sort of, like, pulled some names from a hat, would any of you be surprised that the names on the pieces of paper would be Sam Altman and Elon Musk? I didn’t think so.
I can’t say with confidence that Altman was fired, rehired, fired again, rehired again, etc., – the the entirety of what I noted above that Devansh only yesterday pointed out was a boring distraction of almost no value to the field – because he was exposed as a fraud, and as such, was at risk of destroying OpenAI as an entity. Which he effectively achieved. But I sure do believe it! You see, Open AI had received far too much in investment dollars to put it, and its corporate sponsors (largely Microsoft Corporation, a beneficent protector of society’s better angels if ever there was one (please note sarcasm). So, what inevitably followed? Microsoft hired Altman and a few of his henchmen! Please, please tell me you are in no way surprised by this outcome.
Nor, in truth, do I know if Elon Musk, who has now firmly established himself as the world’s most brilliant imbecile, played a hidden role in exposing Altman as the fraud he is. But I can state with absolute certainty that I posted an eerily similar scenario one day before the entire mess began.
That’s it for this post. Let me part with this: All AI is not fraud. GPTs like LLMs are still very crude but they do show some potential for evolving into useful analytic tools. But, for all the reasons above, and in my prior posts on this Substack, they barely merit being referred to as examples of “artificial intelligence.” Rather, they are better seen as improved machine learning models, and they have recently become generative – even if half of what they generate is false.
If they should be regulated – and I’m not yet sure where I come down on that – then such regulation should come from disinterested entities or organizations that have no skin in the game.
But a great deal of what is labeled as “AI” today is fraud – one of two Zombies, both of which are truly frightening creatures. What’s more frightening? Stunning ignorance coupled with fanaticism (Zombie No. 1), or Machiavellian manipulative fraud (Zombie No. 2)?
The good news? YOU get to decide which you feel is worse. For my part, I say, let’s figure out how to actually kill them both.
Thanks as always for reading and see you next post!
“Be kind, for everyone you meet is engaged in an enormous struggle."
— Philo of Alexandria
--
William A. Lambos, Ph.D.,
Licensed Psychologist, BCN
Neuroscientist, Data Scientist
CNS Computational Neuroscience
walambos@mac.com
813.235.4270 Office
888.503.3166 Fax
[i] Per the NY Times, “Paul Krugman has been an Opinion columnist since 2000 and is also a distinguished professor at the City University of New York Graduate Center. He won the 2008 Nobel Memorial Prize in Economic Sciences for his work on international trade and economic geography. @PaulKrugman”
[ii] Excerpt From Arguing with Zombies: Economics, Politics, and the Fight for a Better Future
Paul Krugman. This material may be protected by copyright.
They are, and they do! They just like to live in “mainframes” (animal and plant bodies) where they can hang out in Mission Control to pull the levers.
Why does agency require DNA? I don't see the reason.