People Don't Actually Want to be Replaced

The Gap in Silicon Valley's Philosophy

The goal of technology has always been to make people better at what they currently do. It's never existed as a mechanism to subvert human excellence or to replace people entirely. The only form of modern technology that has come close to any kind of subversion is social media; however, even that has its benefits to excelling humanity. Before I go any further, I want to make a clear distinction between Artificial Intelligence as we know it now and Silicon Valley's vision for Artificial General Intelligence. Artificial Intelligence as we know it now has no ability to replace people. If you don't believe me, I encourage you to stare at the ChatGPT textbox without touching your keyboard and see if it starts doing things. As it turns out, the largest difference between LLMs and human intelligence is the ability to proactively do things without needing to be explicitly told what to do. Humans' number one edge currently is our own agency. AGI, on the other hand, poses what I can only describe as a threat to humanity. There's a lot of talk about how AGI may not be desirable because it could get out of control and exterminate the human race to serve the greater goal of solving a particular problem. However, I don't think that is at all a realistic scenario. Rather, I think the reality will be much worse.

Socrates, via Plato's Apology, said, "To fear death, my friends, is only to think ourselves wise without being wise, for it is to think that we know what we do not know." As it turns out, AGI wiping out humanity may not be the worst-case scenario. The subversion of the ability for humans to be able to think and make decisions for themselves is the unintended consequence of the abundance we would undoubtedly have with AGI. There is an idea perpetuated by Elon Musk that the road to AGI will be rough, but once we get through it, there will be infinite abundance for all. Meaning we could have a sort of utopia where everyone has a universal high income. To most people, this seems like quite the dream come true. Not having to work? "I don't like work," one might say. "I like spending time with friends and family, exploring the world, and playing video games." The problem with this is the liking of these activities comes directly from the disliking of work. If you break up people's general experiences into two categories, building society and experiencing society, then it's a lot easier to understand why not building society or feeling like you're simply someone who takes from the world without contributing to it would subtract from the experiencing of said society. The joy of experiencing society is that you're the one who partially built it. If you don't believe me, I encourage you to cook yourself a meal and then eat it and notice how enjoyable it is, even if it is in spite of it not tasting great; you still get a sense of accomplishment in eating it. Your experience of building what you're now enjoying is innately valuable.

How much disassociation is too much?

Karl Marx wrote in Das Kapital that "The object which labour produces – labour's product – confronts it as something alien, as a power independent of the producer." Say what you will about the validity of Marx's ideas, but it's hard to deny that he had a point here. As a stark liberal capitalist, I'm not going to sit here and tell you that Marx's ideas are correct. However, I do think there is an important lesson to be learned here. Marx was writing during the Industrial Revolution, where the labour force was being replaced by machines at quite a rapid pace. He was seeing in real time the effects of people becoming more disassociated from their work. The march towards AGI will undoubtedly be a lot faster than the Industrial Revolution, and while it won't necessarily have the physical impact on people that the Industrial Revolution did, it will have a much more profound impact on people and their ability to feel a part of society. Because the promise of AGI is not just that people will be more separated from their work, it's that it won't be their work at all.

There are many proponents of the idea that AI shouldn't be touched or used at all. Most of these people either list baseless environmental concerns or express their issues with AI art and big corporations taking from small, independent artists. However, in calling these people's concerns baseless, I'm not trying to completely discredit them. In fact, these people tend to be far more bearable than the anti-humanist AGI proponents. Largely because I'm a human, and anti-humanism doesn't quite sit right with me. The issue with being completely against AI is that it ignores its true potential. Due to the hype around AGI, people tend to think of it as this sort of problem-solving oracle, capable of solving all of humans' self-inflicted problems. This is not AI's true magic, however. For the first time in history, we have a machine that is capable of understanding human language and communication. For the first time, the gap between humans and machines is closing. Our ideas that we need to put into practice on a computer no longer face the barrier of language. For the first time, if you can speak it or write it, you can make it happen. It is quite literally the first time in history someone could just make something happen by speaking it. The computer is still a tool, but you don't need to be a programmer to use it. Sound familiar? This was the promise of Apple in 1984. The promise of the Macintosh was that the computer would be a tool that anyone could use. A machine for the masses.

But I can't talk to Aristotle.

Steve Jobs articulated the idea of being able to capture the ideas of someone beyond their lifetime and being able to ask questions about those ideas to a machine. Information would no longer be a problem. You can just ask a machine as if it were a person and get an answer. What this means is that you can build anything. If you want to build a car now, all you need is the tools. The information and knowledge to build a car is already there. What this means is that people are innately more connected to their work and labour. They no longer have to rely on big corporations with large teams of people to do the dirty work. They can do it themselves. I'm not proposing that my ideal society is one where everyone is a mechanic. Rather, I'm simply proposing the idea that AI could end up going either way. People will either be completely disassociated from their work or they will be more connected than ever.

The Steve Jobs vision of being able to ask Aristotle a question and get an answer will, in fact, be possible. Many may argue it already is. In the same way that the Mac made the computer accessible to the masses, AI will make the masses accessible to the computer. Computers will be able to understand your intent and understand your goals. They will be able to cut out the middleman and give you the agency to build your own things.

Utopia

What does this actually look like in practice? Picture someone who wants to start a small bakery. Today, they have to hire a lawyer to incorporate, a designer to build a website, an accountant to handle the books, a marketing agency to find customers, and a developer to take orders online. Most people who could be great bakers never make it past the second step. In a world where AI is a tool rather than a replacement, that same person can talk through their legal structure, sketch their own brand, keep their own books, and reach their own customers. They still bake the bread. The bread is still theirs. But the wall of specialists that used to stand between them and the thing they actually wanted to make is gone.

This is the part the AGI utopia gets backwards. The promise of universal high income assumes the bread doesn't matter, only the eating of it. If you've ever made anything with your hands, you know that the eating is downstream of the making. Take away the making, and you don't have a person enjoying bread. You have a person consuming calories.

The social safety net matters here for a specific reason. It's about lowering the cost of trying. If failing at your bakery doesn't mean losing your healthcare or your housing, more people will start bakeries. More people will write books, build companies, learn trades, change careers at forty. The safety net acts as a trampoline of sorts.

Education would have to change too, but probably not in the ways people are currently panicking about. The point of school in this world is not to fill people with information they could ask a machine for. It's to teach people how to think clearly, how to ask good questions, how to tell the difference between something that's true and something that just sounds true. The skill that matters most in a world of infinite knowledge is judgement. And judgement is built the way it always has been: by trying things, failing, and trying again.

Conclusion

The true power of AI is that it doesn't have to think for you. It can think with you. The computer used to be a wall. You needed a translator, a programmer, to tell it what you wanted, and a small army of specialists to turn those translations into anything useful. AI tears the wall down. For the first time, the distance between an idea and a finished thing is just the time it takes to describe it.

The cook still cooks. The writer still writes. The builder still builds. What changes is that the cook can now talk to a machine the way they would talk to Aristotle, or a master chef, or a friend with thirty years of experience. The knowledge that used to take a lifetime to accumulate is sitting on the other side of a conversation.

This is the choice in front of us. We can build AI that takes the steering wheel, or we can build AI that hands us a better one. One version makes humans into passengers. The other makes them drivers for the first time in history. There's a version of the future where most people watch the world get built around them by machines they don't understand, and there's a version where most people build the world themselves. The technology doesn't decide which one we get. We do.

Technology has always been about making humans better at what they already do. AI doesn't have to be the exception to that rule. It can be the most extreme example of it.