What is AI? What is Technology?

We live in confusing times: times defined by terms that everyone all the time is expected to talk about, but which no one at all ever defines. In such a time, it can be sadly easy for ordinary people to feel like the little girl who wrote to the newspaper asking about the rumors of the non-existence of Santa Claus. 

Many of us spend our time these days asking, in one form or another, the question: Is there a Santa Claus? Or, in other words: What do the disturbing, unsettling rumors discussed by my peers actually mean? Are my parents actually lying to me? Is there a conspiracy? How deep does it go? And most of the time, the people answering the question answer only by subtly, deceptively redefining their terms, just like the New York Sun did to poor Virginia O'Hanlon in 1897: Yes, Virginia, there is a Santa Claus in the sense that love and generosity and devotion exist which I know is not what you meant at all and is going to confuse you terribly but what you have to understand is that it will make other adults like me in the know feel affirmed as we continue to keep our terrible secret from you but you are a child and have no choice but to submit to such treatment until you are older and can do it to others.

At The Adventures of Captain Peabody, though, we would never, ever, ever treat you like that! We know that you have questions, say, about what caused the Soviet Union to fall, or what happened in the year 2016, or whether there is a contemporary crisis in Catholic ethics, or whether Christopher Nolan's The Odyssey is a sign of the times, or whether I am in fact a loyal adherent of the Indie Movement, or What Went Wrong, and we strive to answer those questions as simply and accurately and truthfully as we possibly can. 

So, to all the poor confused people out there watching as our society and world transform from a hyper-techno-optimism to a conceptually confused techno-pessimism, I wish to offer this relatively-brief-by-my-standards summary for you. You are reading articles and listening to podcasts and watching people in the GroupMe rant about AI and technology and how it's either going to establish a universal benevolent reign of leisure or end the world, and you wonder: but what is AI? What is technology? But when you ask the smart people, they just tell you that AI is a future super-intelligence that will imprison us all in a Laplacean Hell and also that chatbot that told you it was in love with you or that technology is just when you use a toothbrush and also an inevitable natural process impossible to resist or, in short, I know this is going to confuse you terribly but it will make other people like me in the know feel affirmed as we continue to keep our terrible secret from you but you are a normal person and have no choice but to submit to such treatment until you get a podcast or a tech startup and can do it to others.

Captain Peabody, though, would never treat you like that--and the rumors of his war crimes in WW2 are just that, rumors. So let me just tell you what technology is and what AI is.

What is Technology?

Technology is the human craft of creating, administering, servicing, and directing systems to achieve human ends. For this reason, technology is always and everywhere an exercise of human will and creativity and labor, and always and everywhere political, social, and ethical in nature.

The use of tools in and of itself does not constitute technology. Theoretically speaking, a single man can pick up a single rock and use it to crack open clams until the rock breaks. A chimpanzee can do exactly the same thing. This is not technology.

For technology to exist, a system must be created--that is, something capable of repetitive action and reproduction, but not itself a living thing, existing by and for itself, but a machine, existing by and for other things.

Human technology exists by and for human beings: it exists to achieve human ends, and depends for its continuing existence and reproduction always and everywhere on human craft, which is to say, habitual skilled human labor. This is true of every technology.

Human ends are inherently ethical, since ethics is simply the science of human ends. Human ends, though, are always social and political, and ultimately religious and transcendent, in nature.

Habitual skilled human labor can only exist in a social context where human persons are procreated, fed, educated, and trained. Habitual skilled human labor can only exist in a political context where human labor is coordinated and directed by a governing authority.

When people think of technology today, they have an unfortunate tendency to consider technological objects in isolation, apart from their systemic and social and political causes and ends. We think of a computer sitting on a desk as technology, and not of the global supply chain and manufacturing process and transportation networks and Internet system and national power grid without which it could neither exist nor function.

Technologists constantly exploit this confusion--and never more than now, in an age when virtually no technological object functions at all without a broader network, and where most technological products  are not objects but services. You think of Alexa as a small object that sits on your counter and plays your Spotify playlists; in fact, "Alexa" is a software system networking with and functioning based on a constant Internet connection with further software and information held on servers and data centers distributed throughout the world, and requiring constant inputs of energy and labor and maintenance and repair and updating. Take all this away, and your supposed technological object "Alexa" is literally useless--it is not even aerodynamic enough to make a good substitute for a wiffleball. It could function alright as a doorstop.

Properly understood the basic nature of technology as a system enabled by constant applications of energy and human skilled labor is not a new thing at all, but as old as the human race. It applies as much to the most primitive technology as to the most modern--though admittedly more to the latter than the former. 

A single spear can mean the difference between life and death for a single person in a single battle. Even for that one man with one spear, though, the spear still has to be created. A tree must be felled and wood treated and carved; and if we assume the spear has a bronze tip, metals must be mined, metallurgy practiced, alloys created, a weapon forged. For the spear to be useful, even in one battle, it must be wielded skillfully, that is to say, utilized by a person possessed of craft, or habitual skillfull human labor, . The spear likewise maintained and sharpened, and repaired when it breaks. Taken in these terms, it is more or less impossible that any spear of a given degree of sophistication could exist except in a political society of a certain degree of organizational sophistication.

Even this, though, is an abstraction. We do not, as a matter of fact, ever in history encounter a single person armed with a single spear fighting in a single battle. What we find are societies and political groups utilizing "technologies" that are as more matters of social and political organization as technical knowledge--and which are usually better considered in those terms.

We can, if we want, think about ancient warfare in terms of the technology of the spear or the technology of armor or the technology of bronze. It would be more perspicuous, though, to simply think about the the hoplite: who is a certain kind of person with certain skills existing in a certain kind of social and cultural environment with certain kinds of economic and political structures, and for all these reasons also a kind of technology. 

Individually understood, a hoplite is a soldier fully equipped with bronze armor and wielding a shield and spear; an individually understood hoplite, though, is not a hoplite at all. Properly (which is to say, teleologically) understood, a hoplite is a person capable of fighting as part of a phalanx: which is to say, a tightly-grouped, well-coordinated, fully armored unit capable of interlinking shields and spears and moving and charging as a single entity and thereby crushing without losses more or less any group of less organized and equipped soldiers out there. 

Not every kind of person from every kind of culture or society, however, is capable of functioning as part of such a unit. Someone from a highly individualistic society is likely to be unwilling or unable to subsume themselves into the phalanx. Someone from a highly hierarchical or differentiated society is likely to be unwilling or unable to fight alongside their social inferiors in a phalanx. Someone from a highly egalitarian or familial society is likely to be unwilling or unable to specialize in the skills required to fight effectively as part of a phalanx. 

A professional military, to be sure, can overcome these disabilities to varying degrees by removing people from their original societies and cultures and incorporating them into an entirely separate and all-encompassing political and social and economic structures of their own. Most ancient societies, though, did not possess such professional militaries--and even when they did, they still relied overwhelmingly on their recruits coming from a small range of types of societies and classes. 

Under normal circumstances, only someone raised from their birth in an intensely socialized, politicized, and collectivized society can manage the feat of being a hoplite; let alone the accompanying feats of mass-scale centralized production and distribution of arms and armor required to outfit an army of hoplites. Properly understood, then, the hoplite considered as technology is entirely inextricable from the collectivistic political and cultural model of the polis. Greek culture produced a collectivized, differentiated sort of man used to working and living closely with citizen peers defined by their orientation towards gaining honor through the achievement of collective political and military goals--or in other words, Timon of Athens, the man who the wise Solon in Herodotus considered the happiest to have ever lived, a man who saw his great-grandchildren and died fighting for the state and had a monument reared in his honor. Greek economic structures dug mines and paid metallurgists and developed trade routes to bring metals and skills. Greek political structures compelled wealthy people to set up centralized armories and provide weapons and soldiers for the state and produced constant low-level warfare between cities. And all this together formed the technological system of the hoplite.

And here is the really magical thing--the technological thing, we might say. At their height, a phalanx of hoplites properly trained and outfitted and psychologized could not only defeat but effortlessly annihilate any less trained and equipped and organized group of soldiers whatsoever. They could accomplish feats in warfare that were legendary, mythic, beyond belief. No number of guys with spears could stand up to them.

Everyone in the Mediterranean copied hoplite warfare, including the Greek's enemies. In doing so, though, they emphatically did not only copy what we think of as technology: for the simple reason that that didn't work. You could take all the hoplite's equipment and give them to a group of warriors from another culture, and they would still be annihilated. Hence, when people copied Greek warfare, they usually copied many other things as well: including the Greek language, Greek literature, the Greek polis, Greek gods, and Greek pederasty. The Roman Republic copied all those things, and also utilized Greek hoplite warfare--until they found a better military system that could defeat hoplites. 

Nearly all that was said above can and must be said of every technology; and none more than contemporary digital technologies, which both depend upon and create a certain kind of atomized, professionalized, technologized person existing within certain kinds of mass-scale financialized economic systems and bureaucratized and plutocratic political systems and pornographic and advertising-based cultural systems.

Americans are trained to think of technology as essentially "labor saving" and indeed "labor replacing"--which it sometimes is. Sometimes, though, the opposite is true: technology can and often does function to create new forms of labor and even to increase the overall demand for human labor. There is a real and obvious sense in which an industrial spinning machine is a labor-saving device relative to a handloom; the amount and degree and concentration of labor required to run early industrial machines, though, and the consequent demand for cheap textiles produced by that, certainly created a great deal more labor overall.

Labor, though, is not precisely quantifiable; just as important is the qualitative features of labor, transforming it, say, from being more skilled to less skilled, or more dangerous to less dangerous, or more individual to more collective, or vice versa. In the case of early industrial machines, they created labor that was more concentrated, more intensive, vastly more dangerous, but also vastly less skilled. A washing machine saves a housewife a lot of difficult labor: someone, though, has to make the machine and repair it when it breaks, and the two forms and degrees of labor by different people with different skillsets and needs cannot be easily quantified or compared. And so on.

Regardless of the degree or kind of human labor required, though, every technology requires human labor to exist and function--usually rare and specialized and organized forms of labor. This labor is carried out by human persons raised and trained and incentivized and directed in some sort of social and economic and cultural and political system. 

To summarize: every technology both depends upon a certain kind of human social and political and economic system. This indeed is what makes it so hard to understand, and so frequently misunderstood even by those who have the most to do with it. Physical science is aimed at understanding natural substances, not human social and cultural entities. Yet technology is a human social-cultural entity, not a natural substance.

On the most essential level, then, every technology is a social, political, and economic entity, not a scientific or technical one. It can only be understood in human terms. To understand why a given technology is produced at a given time and place, how it functions, and what effects it is likely to have, it is absolutely necessary to understand the values and goals of the society and culture and political entity that produces and sustains it.

What is AI?

Here is something I have said before in this space: "AI" in contemporary usage is not a descriptive term for a new technology, but an advertising term for an essentially political program of reallocating massive economic and social resources toward developing and finding applications for and forcing large-scale use of a heterogenous range of pre-existing digital technologies, products, and services.

I still agree with this statement. Taken by itself, though, it is liable for misunderstanding--most of all in the dichotomy it poses between "technology" and "political program." That a technology is produced by and requires for its functioning very particular political and cultural and economic systems should not surprise us. 

That this sometimes means that a given technology is simply not worth the political and cultural and economic changes and ethical effects on people it is produced or will produce is a very basic and frequent corollary of that reality. Innumerable technologies have been either not developed or not implemented or abandoned because of this. That in our society there are people who think that one can never make a statement of this form out of fear of sinning against some immanent progress-demon or imminent machine-god simply means that there are many people in our society who are completely insane. But you already knew that, didn't you?

The real difficulty with AI, though, is that it does not really refer to a technology in the common sense at all. Its marquee products in the last few years, it is true, have mostly been a range of what used to be called "machine learning" techniques and derivatives of what in the 2010s were known as "neutral net" algorithms centering on a range of what are now called "Large Language Models" and derivative and analogous systems. These products, though, even taken together, have made up only a fairly small proportion of investment in what are popularly called AI products and systems--and an even smaller proportion of profits. Taken in those terms, AI would be best described as a bevy of short-term financial instruments.

Even if we accept AI as a rough term for these technologies, though, we cannot even begin to understand the broader social and economic and political system of which ChatGPT is a mere minor derivative and unprofitable red-headed stepchild. 

If AI is to continue being a useful term, then, I would suggest that it might be better to use AI in a broader but I think more useful way, for the actual technological and political and economic and cultural system that has allowed the bulk of global GDP growth for the past several years to center on servicing 3-5 unprofitable companies making products that virtually everyone hates.

By the mid 20th century, the United States of America was the industrial and agricultural powerhouse of the world--a status cemented even further by the ruination of all its rivals in WW2. American farms grew more food with less labor than anywhere else in the world; some of this food was shipped and sold overseas, some of it bought and stored by the government in granaries, and the rest fed the commercial and industrial workers who produced the best heavy equipment and consumer goods in the world and sold them. Thanks to the 1944 Bretton Woods agreement, the US Dollar, backed by gold, became the global reserve currency, ensuring that the US government would remain flush with cash and US businesses with investment. Wall Street was the global financial center, attracting and dispensing investment throughout the productive economy.

In the late 20th century, though, all this changed. American agriculture entered a deep and terminal decline spurred by massive debt spurred in turn by predatory technological and financial systems spurred ultimately by the continuing and escalating desire to grow ever more food with ever less labor. American consumer and heavy industries were first undercut and then annihilated by cheaper and at times more sophisticated consumer goods from Europe and Asia. The global reserve currency became, for all intents and purposes, oil, a resource controlled largely by unstable Middle Eastern Muslim governments.

And yet, despite the collapse of virtually every basis of American global economic power, the American government, the American military, and Wall Street all continued to flourish. The American government found that its sovereign debt could be bought up by allied political powers interested in propping up the American global system: Europe, Japan, eventually China. The American military-industrial system had already become the arms supplier du jour for the world, a status reinforced further by the collapse of their main rivals in the Soviet defense industry. 

Wall Street, meanwhile, learned to flourish not despite, but through the collapse of the American productive economy. Hostile takeovers and liquidations of productive industries competed with new financial instruments premised on gambling for its own sake: with both finally culminating in the orgy of the 2000s housing bubble. When that bubble collapsed, the American financial system as it had existed, bound up with production across the world and the savings of the middle class, collapsed with it--and has not come back.

It was at this time, properly speaking, that AI was born. 

AI in this inclusive sense, it must be said, is not simply another name for the Internet, or smartphones, or even social media. All of these existed well before 2008, and were in their earlier stages closely dependent on real-world productive industries. People paid real money for smartphones and personal computers and Internet connections; and while people did not pay for Facebook, Facebook was not that big of a deal. Certainly Facebook was not the economy.

All this changed when the economy itself collapsed in 2008: and that, and not the release of ChatGPT, is the real inflection point in the American history of the 21st century.

Here, then, is the definition: AI was and is the attempt to keep the American financial system going post-2008 by systematically decoupling it from both real-world production and middle-class assets, and instead tying it to the success of tech companies at attracting venture capital and harvesting data.

Over the past 18 years, it should be pointed out, tech companies have come up with an absolutely absurd amount of entirely different ways to attract venture capital and harvest data. Very few of those methods, unfortunately, have ever worked in the most basic sense of producing something that human persons actually wanted to acquire in large enough quantities to achieve their ends to profitably make a profit and pay back their investments. Along the way, admittedly, an awful lot of nifty technological objects and systems have been created, and actual scientific and medical technology has continued to advance while benefiting intermittently from tech-inspired infusions of capital. Taken on such terms alone, though, in terms of its actual products, the technologization of the economy has certainly not been a rousing success.

As many many people have pointed out over the past six years, virtually none of the "whales" of the venture capital economy have ever succeeded in becoming reasonably profitable on their own through producing and selling actual products to consumers. This, though, has mattered less and less for the simple reason that the financial system itself has been less and less interested in any kind of productive economy. Uber has never been profitable on its own terms, but its stock has gone up, which has made short-term money for investors: and this, and not the productive economy, is what our financial system has come to depend on, and hence what corporations have come to depend on for their existence and functioning year-to-year.

Even more drastically, though, the biggest corporations in the world have more and more come to center not on producing any saleable product at all, but on harvesting and processing data. Data, though, does not make money--and the overwhelming majority of it is useless for any purpose whatsoever. Targeted advertising, the basis of profit for both Facebook and Google, has been again and again proved useless and worse than useless for the purposes of actually getting people to buy products more than randomized ads. Internet engagement has been cratering continuously for the past ten years. Most data harvested by systems has for decades either been stored and not used for anything or else sold to governments that mostly do nothing with it too except when they occasionally use it to target and blow up political dissidents and random people that they could have easily blown up anyway.

Anyway, I don't mean to be prejudicial about blowing up political dissidents or investing billions in unprofitable start-ups when I say that, however fun or interesting or even useful both may be in their place, neither is capable of simply replacing the entire productive economy of America and the world. And yet that, somehow, is the pickle we find ourselves in today.

From one perspective, the rise of AI over the past five years has been simply a desperate attempt to keep this process going after everyone everywhere, including the people running the system, stopped believing in it. From another, it is an even more desperate attempt to cash in on all the supposed benefits of tech company development since 2008 at once, the fantastic hope that in LLMs and other neural-net architectures tech companies could finally find a way to take all that useless data and all that investment capital and turn it into something that could really and finally do things: things like replacing all programming and cultural and educational and artistic and  jobs in the economy in one fell swoop and/or creating new surveillance and military systems that could find and target and blow up any dissident anywhere in the world up to and including entire nations and/or making a superintelligent alien machine-god of power who would kill us all. 

So far, it's been a bit less like this and a bit more like Elizabeth Holmes promising continually for 10+ years that she really had or was really going to have have or eventually if you just trusted her and loved her would certainly absolutely always have had a single small device that could perform 200+ tests cheaply on a single drop of blood and using that to con people out of more and more venture capital and run the same tests on the same devices as everyone else just really inaccurately and so get a bunch of patients killed. 

Over the past few years, a bunch of books have certainly been shredded, a bunch of loud data centers built, a bunch of NVIDIA chips stored in warehouses in preparation for data centers that will never be built, a bunch of Gazans targeted and murdered by algorithms, a bunch of African children forced to mine rare earth minerals, an Iranian girls' school struck, an Iranian war lost, Houthi missiles targeted onto Saudi tanks by Claude, an awful lot of terrible art created, and quite a few people driven stark screaming mad--but have we really gotten the payout that would make 25 years of economic and political collapse and financalization and dismantling of the productive economy and predation and pornography and addiction and belief and hope and fruitless investment worthwhile? Have we really found the one device that will do any blood test and fit on people's desks? Have we really found the One Weird Trick to replace an entire global productive economy composed like all economies before it of agriculture and industry and trade with a single really buggy chatbot? 

But hey, don't you trust Elizabeth Holmes? Don't you trust Sam Altman? Don't you trust that one Anthropic guy? Don't you trust Elon Musk? Don't you trust Donald Trump? Don't you--

And that, dear Virginia, is what AI is.