The Question You Can't Ignore
Before you read this, a quick warning. What I’m about starts to bridge into religion over technology. I think the initial reaction will be to quickly judge that (a few days ago, I might have too). If you don’t like it, stay away. Everything here is a WIP in my mind, I’m considering, haven’t finalized, DYOR. It’s the kernel of the idea though, and I wanted to share it.
It’s built on the assumptions in this post I wrote earlier this week, and an expansion on the core ideas. The assumption is that we have some level of proof that building AGI and superintelligence is possible, not an if but a when.
The question is pretty simple:
Posit superintelligence, then, what is our place in the universe?
I’m not going to pretend to be even close to tackling that, but I’m going to cite a bunch of sources much more educated than I am and try to back out an argument.
Background
A lot of this context dates back to the summer of 2014, I think, which is when I first heard of OpenAI. I’ve been trying to dig into hacker news posts etc (that’s where I was living as a teenager). The first OpenAI post I can find is December, but I’m almost sure it was being talked about earlier. At the time, it was quite a shocking statement, Elon and crew had committed $1b to funding AGI research. It made a pretty big buzz in that circle, was the first time I really had any knowledge or exposure to Sam too.
There was a lot of writing about why, and that’s where I started digging.
The core reasoning for the research lab and non-profit was fundamentally existential. Elon in particular was citing this book called Superintelligence by a guy named Nick Bostrom. The HN post linked at the bottom also has pretty good context. At the time I was so curious I read the book (circa 10 years ago, so a little rusty), but basically, it covers how incredible the technology can be if done right and all of the risk scenarios involved in its development. It’s actually a little dry, he iterates over literally all of the takeoff scenarios you can think of. If you dig more into Nick, he’s the leading philosophical expert in these technologies and leads an organization called the Future of Life Institute.
Nick is an interesting dude. He’s written a lot about various topics in this space, and I think of him as seeding some of the core ideas that are behind the motivations of these technologists. To add a few other core ideas worth reading about, simulation theory (literally wrote the seminal paper on this, Elon clearly follows it) is a really interesting statistical argument for us living in a computer simulation. It’s this, along with helping seed the concept of long-termism, which is rooted in the idea that most humans will live after our current timeline resulting in a moral and ethical basis to make sure their lives are as good as possible, including a bunch of assumptions about the future of technology.
AI Creation and the Cosmic Host
Naturally, when I started thinking about this question, my first thing was, what is Nick up to?
Turns out, more recently, he’s thinking about AI and God.
This is the first time I’ve seen someone so close to this technology take a shot at the why.
The paper is pretty dense, but I’m actually sure the background and research that went into it is even more dense. I spent all day on this today. Would be pretty difficult to even do a summary of the argument, so you should just read it if you’re interested. The conclusion is pretty simple though:
In simple terms, after dedicating his life to this, every major technologist building this following him, building the leading institute on these ideas, he’s saying God exists, and we have a moral reason for building superintelligence in a way that will contribute to the greater good of the universe.
That is a fucking wild thing for someone of his stature to say.
The other hint embedded in this is that he believes there are moral laws that govern everything, cosmic laws that we might not be fully aware of, even. They might even transcend physical laws. When you start assuming certain things about the host, unbounded compute etc etc, then physical laws might just be being imposed by the simulator that could change. It’s also pretty reasonable to think that the simulation could have started more recently than we realized (cough cough, the statistical significance of living in this timeline is shocking if we really think the universe has been around for billions of years).
It’s a similar conclusion to most world religions. I’ve just never seen an argument rooted solely in technological progress, maybe others exist elsewhere.
At the beginning of the post, did I say this was going to get religious? Well, yes.
I also said, WIP, not final, don’t quote me, DYOR.
Conclusion
Technology and god are not mutually exclusive - in fact, god is more human than machines are. If a hardcore technologist is now finding god, imagine the rest of us. We will need a moral compass in a world where morality is likely all that will matter as other fears/needs fade away.
References