by
Adrian Daub
View all notes
Available query params: commented=1, untagged=1
As the economist Tom Nicholas has pointed out, the American institution of venture capital (VC) emerged out of the attempt to ensure the persistence of capital amid economic windfalls connected to easily depleted natural resources, a monetary message in a bottle passed from one boom to the next. In an economy geared toward the rapid exploitation of immense natural re-sources, VC was the way in which quickly earned money could be transformed into capital for the next boom. The great whaling families of New England secured themselves against the inevitable depletion of whale stocks with venture funds. And the railroad entrepreneurs, who had extracted enormous sums from the state, pooled their money to reinvest it for the moment when those subsidies inevitably dried up.
Today, venture capitalists are the central drivers in what tech calls governing. They are an integral part of the ecosystem, but a part that reaches both backward and forward, into a time before Silicon Valley and a time after it. The money they invest is of dynastic, sometimes even feudal, origin. The arrangement of this money with pluralism, with free competition, with liberal democracy is entirely provisional. This is something that those who manage it seem to sense acutely. Among the CEOs of the large technology companies along the San Francisco Bay, those openly sympathetic to the Trump regime are still in the minority. Among the venture capitalists, however, especially those you hear from in media, many are right-leaning.
In 2020, the investor Paul Graham wrote an essay titled "How to Think for Yourself." One lesson: "If you're surrounded by conventional-minded people, it will constrain which ideas you can express, and that in turn will constrain which ideas you have. But if you surround yourself with independent-minded people, you'll have the opposite experience: hearing other people say surprising things will encourage you to [sic], and to think of more." At the end of the essay, Graham thanks the broad panel of independent thinkers with whom he has surrounded himself: Six of the seven are men between forty and fifty-five years old. Five of the seven work at Y Combinator, a business incubator in San Francisco cofounded by none other than Paul Graham. One is the CEO of the payment service Stripe. The seventh is Peter Thiel.
lol
A myopic perspective on the world has long characterized the tech industry, but success has amplified it. Age and wealth often inhibit contact with people whose lives and careers have unfolded in fundamentally different ways. When great fortune befalls an individual, they might ask themselves some lonely nights: "Do I really deserve all this?" But when a group collectively rushes from success to success, when truly extraordinary success has become the norm by the age of twenty-five, thirty, or thirty-five, and you are forced to then age alongside one another for decades, each constantly aware that your peak lies ever further in the past, then it helps when fellow travelers constantly reassure you that you are still relevant, that the gnawing feeling of irrelevance is, in truth, the fault of others. A concept like "vibes" allows them to understand their aging process, and the alienation that inevitably accompanies it, as identical to the course of the world.
This doesn't mean it's wrong to warn people about Palantir. But our warning shouldn't focus on what the company's tech can do. Our warning should co-cern the fact that that tech is deployed at all. It simply doesn't matter whether Palantir's newtangled systems are more scarily omniscient than the old ones.
If the people to whom this technology is applied include undocumented immigrants in the US or Palestinians in the Israeli-occupied territories, then it is almost immaterial whether the algorithm picks out the "right" people. Those in power have already decided that there are essentially no "wrong" targets in these groups. Palantir is necropolitics, pure and simple. The data sets are merely window dressing. It's not the fact that the data miner might miscategorize someone that makes an individual a victim of Palantir, but rather the fact that Palantir is categorizing them at all.
Certainly, this ready identification with villains, often totalitarian villains, is already plenty creepy by itself-and also explains the large personal and ideological overlap with an administration dominated by online Nazis, who often like to draw inspiration from cinematic heavies. But there is also a fundamental poverty at the heart of an ideology that gets excited about fictional dystopias, villains, and totalitarianism. Not because those are evil, but because they're usually so thinly drawn. The fact that we always see these cinematic villains dominating, but rarely governing, likely strengthens their suitability as a source of inspiration rather than weaken it. We are fairly well informed about Galactic Empire's supply of fully operational battle stations in Star Wars. We know less about how the Empire handles wage negotiations or pensions.
But identification with dystopia is only one aspect of this now well-established language game among tech companies. The other is disidentification with power. One is a witness to one's own power rather than wielding it. One experiences power as an aesthetic effect in two senses—it doesn't feel entirely real, even when it is deadly real to those upon whom it is exercised. And it is, like sensation in Kant's transcendental aesthetic, something that is given rather than made in its precise parameters. Power is something that happens to the powerful.
Julius Blank, Victor Grinich, Jean Hoerni, Eugene Kleiner, Jay Last, Gordon Moore, Robert Noyce, and Sheldon Roberts tounded a new company, Fairchild Semiconductor, from which over time several other microchip companies spun off. Among the "Fairchildren" were Intel and AMD, but so were venture companies like Kleiner Perkins. Like the "PayPal Mafia" of the late 1990s, the "traitorous eight" were fairly homogeneous in their backgrounds, their age profiles, and their management styles. And they were the grandfathers of today's Silicon Valley. The VCs who funded their endeavors were very similar people: investor and inventor spoke the same language, specifically because they shared the same generational experience.
When compared to the "traitorous eight" or even the PayPal Mafia, the generation of founders and their backers that have dominated headlines and the contemporary image of Silicon Valley appears a bit more diffuse. Still, the age cohort remains a clearly identifiable phenomenon in the Valley, in particular since so many cofounders met in college. Steve Jobs and Bill Gates are the same age; Sergey Brin and Larry Page, who met at Stanford, were both born in 1973. Mark Zuckerberg and the entire original inner circle of Facebook were the same age, which is the same age as Instagram's founders Kevin Systrom and Mike Krieger. It's not really true, as is sometimes said, that through tech our world is dominated by solitary geniuses. It is dominated by cliques of people who are more or less the same age.
makes me think of that quote about generations, napoleon, etc
The botanist, eugenicist, and peace activist David Starr Jordan wrote in 1903 that the Civil War had delivered "the best" of the "race" unto death; those Americans who took their place and sired the youth of his own day, he thought genetically inferior. The wrong kinds of Americans predominated in 1903, because the right kinds of Americans had been needlessly sacrificed between 1861 and 1865. When he wrote his pamphlet, Jordan was president of Stanford, the university's first. In 1965 a Stanford professor gave a long interview to U.S. News & World Report with a similar thesis. In the United States, he claimed, the "abundant welfare state" all but insured that unfit and less intelligent people "will be multiplying at an enormously faster rate than more intelligent people do."' That professor was none other than William Shockley, who had begun teaching at the university after his company failed. And who had turned his yen for hierarchies into speculations about race, the welfare state, and intelligence.
This argument attained its modern form in Charles Murray and Richard Herrnstein's The Bell Curve, which was published in 1994. The two made the same argument as Shockley before them-that the welfare state subsidized "births among poor women ... who are disproportionately at the lower end of the intelligence distribution" - but this time under the banner of "group 10."* According to Quinn Slobodian, this marked the comeback of the IQ concept among American conservatives. And even during times when Silicon Valley Presented itself as far more liberal than it does today, it could still be enthusiastic about IQ.
In that sense, Musk's enthusiasm for Al-generated images of himself is another triumph of the imaginary over the grim reality, which is this: Elon Musk is not a warrior of some galactic empire, not a Roman centurion, not a rapper or an astronaut. He is a middle-aged man with all the problems that come with being a middle-aged man, minus the ones having to do with money. His children don't like him, his subordinates don't like him, his time is running out, and his body can perform the tasks it once accomplished with ease only with the help of drugs. AI promises someone like him more endless role-playing. And, in his understanding, Al also renders meaningless the years or even decades of accumulated expertise that generally lend our biographies their characteristic rigidity and inflexibility.
In fact, if we try to encapsulate what Al promises, we could just say: immaturity. That may not sound particularly tempting, but loitering in immaturity also means emulating those in power. Our historical moment is characterized by the power of immaturity. Or rather: Our time is characterized by immaturity as a technique of power. As soon as a person (assuming they have the right skin color and gender) is old enough that their transgressions are no longer excused as "youthful indiscretions," they already shift into the paradigm where they are simply "part of a different generation" and therefore cannot be held responsible for their words and deeds. The power of our day is always immature. And so are the technologies that wield it.
If this promise sounds a lot like a curse, it's probably because it essentially turns the main threat AI poses into a positive. Because if the trends of recent years continue, Al will primarily have the effect of keeping people in a state of untrained, undetermined adolescence for longer. In Silicon Valley, the jobs falling victim to Al are those entry-level coder jobs that previously represented the first rungs of the career ladder. Whether this is due to the true potential of the technology or instead due to the self-deception of the bosses, Al is destroying that phase of life in which one gains professional experience, contributes, and makes mistakes. In which, as Hegel would say, one "determines" oneself, opens some doors, and closes others.
What happens at that moment is the following: The remote operator in the data center doesn't take over the steering wheel, as passengers sometimes assume, but rather plots a new course for the car on the city map. I don't know why the company opted for this form of human intervention. The important thing is that it opted. Behind the AI of a self-driving car, therefore, lies not only human intercession but something far more significant: a fundamental framing decision about when and how humans intervene in the system. The passenger is completely unaware of this. They drift along, whether they like it or not, in their black box.
The point, therefore, is not whether Al will eventually become more sophisticated or not. Whether Al will be able to do what its creators promise or threaten. The point is that talk of maturity deprives the public of a whole range of decisions and choices, by pretending instead that there are no decisions and choices to be made. When in fact, the maturing technology will involve more and more of these decision points. When the Irish capital, Dublin, decided to build a new subway, Dermot Desmond, the ninth richest person in Ireland, proclaimed that the government should cancel the project. After all, "Al" would make the project "obsolete" anyway. He was referring to cars from manufacturers like Waymo.
In such a statement, "AI" is used as though it were a familiar and agreed-upon concept that—once it's fully developed—will have a clear and unambiguous effect, as it happens precisely the effect the speaker's argument requires. It will negate or render obsolete whatever decisions he dislikes. It will make inevitable the outcome he desires. Dermot Desmond doesn't address how this is supposed to happen, how exactly "Al" is supposed to implement autonomous driving and make public transport obsolete. More importantly, the Irish billionaire didn't seem to grasp what's obvious to anyone whose Waymo car has to be controlled by a data center: that even when the technology is far more advanced than it is now, its deployment isn't a given, but will depend on political, legal, and business decisions - decisions made by programmers and product designers, CEOs, regulators, and politicians. Or perhaps Dermot Desmond does know this; perhaps he's hoping that Irish newspaper readers will forget the fact as quickly as possible.