No Employees
My first job out of grad school was, regrettably, at a Bitcoin startup. There are a lot of bad things to say about that experience, but if I try hard, I can think of two good things that I have to say about it. The first is that being in that world visceralized the machinations of Silicon Valley and venture capital. Seeing how the sausage is made clears up all the smoke and mirrors.
The second is that, among the people I worked with, there were maybe four people who were both good people at heart and competent. And in the areas they weren't competent, they were at least thoughtful and introspective enough to take feedback well and grow. The CEO was not one of them. But working with that small group was the first of experience I had, out in the wild world, demonstrating the value of community building. Or if you wouldn't call it a community, some other word for fostering cohesion, alignment, and a shared understanding among peers working toward a common goal.
This isn't a shitpost about that CEO, except for this one part. After I had been working at this company for about a year, the CEO was invited onto a no-name blockchain podcast for some idle pontification. The podcast was livestreamed on YouTube, and the CEO asked us employees to tune in. Then, in response to a question like, "What would you do differently about this company if you could start over?" He replied—in a way that made it excruciatingly clear he had thought about this a lot—that he would have hired no employees and run the company entirely via blockchain smart contracts. At the moment he said that, the live viewer count on the stream was 4. The only people watching were his employees.
Putting aside how oblivious and conceited that insult was, it underscored for me that the way bosses (writ large, but particularly in tech spheres) think about new technology is very different from everyone else. Bosses resent having to rely on employees for their skills. Any technology or financial contrivance that can substitute capital for having to deal with human needs and desires is a critical investment opportunity. This has become awfully obvious in the last three years, as companies have raced to replace workers with LLMs.
So when I read an author or movie director beffudled by the idea that someone would want to use AI to "make art," I think to myself, of course, they want what you can make without your pesky opinions in the way. The same people can't educate their children, and so are beholden to the education system and the mess of politics that entails. Wouldn't it be great if some financial contrivance or regulatory loophole allowed them to opt out and make their own school staffed by people who will bow to their whims? Even better, what if those staff members were obsequious robots?
I enjoy reading the author Kyle Chayka, who among other things wrote the books Filterworld and The Longing for Less. At one point he was active on the same social media platform as me. He posted something about some decentralized autonomous organization (DAO) project, marveling over how cool the particular consensus mechanism was. In bad taste, I replied with a negative take on DAOs and blockchains in general as fundamentally designed and intended for immoral purposes. Like crime. Like bypassing regulations. Like financial joyriding. Like replacing employees.
And, in the way that short-form text can be, I came off as mean. Like I was calling him bad for thinking it was cool. It's all a bit sad, because, from a purely technological standpoint, decentralized consensus really is cool. And if cryptocurrency wasn't buying politicians, or funding oppressive regimes, or being the backbone of modern crime, what would remain is genuinely novel. If you divorce LLMs from their economics and the way these bosses want to use them to replace skilled labor, fuck up energy markets, make people stupider and more suicidal, and accelerate climate change, the way they work is technically very cool! And if we had ten more years to stew with small-scale LLMs, using them out of curiosity for idle pursuits as they matured in capability, reliability, and efficiency, instead of being force-fed chatbots and AI slop in every possible setting, attitudes might be different. (My neighborhood public elementary school advertised an after-school program in which they teach seven year olds how to use Midjourney. How could they possibly think that will go well?)
But we can't divorce the technology from its commercialization. This is a tragedy of modern society, that all new technology is fashioned into a financial weapon before it is viable. The instantaneous jump to commercialization molds a novel idea in the image of finance. It creates leverage that forces our collective efforts into ensuring a new technology generates enough revenue to justify the investment. The hermeneutics are rotten because in this scripture adopting the technology is preordained, and profit is God.
I try to think of examples that have bucked this trend, and the only one that come to mind is a technology that failed an explicit push for hypergrowth, and fallen into a steady state of niche, genuine usefulness: 3D printing. I should really get into that hobby.