Everyone a Darcy
Work and AI
Just a few thoughts on Conti’s recent article “Big Tech’s War on Human Achievement”:
That AI leaders believe in transhumanism doesn’t make the strongest forms of transhumanism (supersession by machines or merging with machines) possible metaphysically or desirable popularly. VR goggles are possible metaphysically (and actually exist) but few people want to use them and using them has no social value right now. And the metaphysical challenge of psychological continuity with a machine is still very much addressed in the transhumanist literature by hand-waving analogies like the Star Trek transporter mechanism or the Ship of Theseus. Yes, “transhumanism” encompasses more than the idea of merging and that merging can be accomplished (as by smartphones) as much by habit as by a direct interface. But it is almost like the literary authorial fallacy to take anything Dario Amodei says straightforwardly and judge AI on the basis of it. We may separate Henry Ford’s ideas from the use to which we are putting cars, Edison’s from the use of film, Watson’s from the use of genetics, Egypt’s from use of gold, Thomas Jefferson’s from the use of his legal framework, etc.
Conti divides human disciplines into those that can (mathematics) and can’t be (philosophy) done well by machines. But then he talks as though those that can’t be done by machines will “simply die.” Why hasn’t he just proven that philosophy will survive since it requires the “disposition of [human] love.” His causal chain there is that we need human philosophers to have good philosophy and somehow having AI around will deplete the motivation for character-building philosophical work that is the womb of the good philosopher. But in so many disciplines right now, younger learners cut their teeth on solved problems - recapitulating them - and in the process changing themselves into virtuosos. The violin has been “solved,” and yet people are still challenging themselves to become violinists, and virtuosos at that.
His example of Secretariat is a great one. Sure, most can’t remember the jockey’s name, but we know Secretariat was ridden and that he was the result of countless human hours of irreducible effort and expertise in breeding, nutrition, training, and care. Secretariat was a “thoroughbred” whose very subspecies and existence was the result of human planning and care, almost like a multi-generational artifact. Secretariat himself can be a hero, and one day a robot that is easier to individuate may be heroic. It is currently hard to make Claude a hero in the way it is hard to make the National Thoroughbred Racing Association a hero. But Demis Hassabis is a hero to many, as could the humans behind (and riding) Secretariat if we took the time to know them.
Yes, always turning to an easy answer is a moral failing that deprives one of character formation. But this temptation existed even before the era of AI or the rise of the YouTube explainer. If Conti were simply saying that AI makes it harder to be virtuous, that would be defensible, and he would have an impetus to reflect on virtue formation in the era of AI. But it sounds like he wants to remove temptation instead of building a new kind of virtuous person. The indefatigable diligence of ants doesn’t/shouldn’t discourage me from being diligent.
Conti’s thoughts on vocation are interesting, but I wonder to what extent they come from a concept of vocation carved out by modernity. These are jobs based on specific subject-matter expertise, credentials, and skill formation. What is the vocation of a gentleman like Mr. Darcy? His vocation is the cultivation of virtue in the exercise of his responsibilities to the land, estate, the people who depend upon it, and ultimately his country. His vocation involves upholding justice as a magistrate. His vocation involves finding and wooing a companion to carry on this work and to illustrate felicity to others. Nothing that Mr. Darcy does, other than maybe the annual bird shoot, should be called a “hobby” and yet he has no “profession.” Even when his tenants are satisfied; his steward has a bundle of tasks to fulfill; his love is comfortably working to make something beautiful, form the character of a friend or child, or learn a new piano piece; and he wanders out to commence a new gardening experiment, calling this a “hobby” seems to diminish the value of such work. We constantly divide, name, grow, and then take things to scale that start as pet ideas, “hobbies,” and amusements. I assume the new heavens and the new earth will be full of endless horizons to explore where we engage in meaningful pursuits. Heaven is not a world of hobbies.
I grant the problem of possible demoralization, the arc to which AI bends, as Sacasas describes it. But this is also the essence of commensurability thinking - putting humans and AI on the same scale. I really don’t linger over how a perfect circle can be drawn by a protractor even though I’m grateful for precisely milled machines. But I still will spend five minutes watching a video of schoolkids compete to see how perfectly they can draw a chalk circle by hand.
The overall framing of Conti’s article has a surface plausibility - that AI prosperity is like communism in that it will remove “the striving that brings romance and meaning to our lives.” Yet first, this requires our equating the human-enforced coercion of communism (the Wilt Chamberlain thought experiment) with the natural result of capable technologies. And second, we don’t need scarcity to find meaning; this is the wrong kind of post-lapsarian thinking. A feast isn’t great because it must end.