About ten years of so ago I discovered object-oriented ontology (OOO) and Bruno Latour. That plunged me into a philosophy period during which I ended up taking a really Big Look at things. I ended up sketching a cosmology/ontology, Living with Abundance in a Pluralist Cosmos: Some Metaphysical Sketches. I ended up arguing that, to date, the universe has seen the emergence of three arenas of abundance. I’ve taken the term “abundance” from Paul Feryerabend: the universe is abundant, it just keeps generating lots and lots of stuff.
I’ve identified these three realms on this nice chart which I round in Wikipedia’s entry on Universe. Roughly speaking, the universe began 14 billion years ago, giving rise to the arena of Matter. Four billion years ago Life emerged. While animals do have culture – the higher primates, certainly, do beaver dams count as culture? I don’t know – it’s human culture that ushered in a new arena, that of Culture. Wikipedia dates Oldowan tools to about 2.6 million years ago. That’s when we can locate the origins of human culture. Whether it’s a million years later or earlier hardly matters on this time scale.
Look at the time scales involved. Speech emerged millions of years ago. Writing only 1000s of years ago, 10,000 at the outside. Calculation is 100s of years, 1000 at the outside. And computation? Darwin published On the Origin of Species in 1859; Clausius formulated the idea of entropy in 1865. But the digital computer is a mere 70 years old or so. Cultural evolution is accelerating.
And what if we measure those intervals in generations, where a generation is 30 years? Let’s simplify things:
Speech = 1M years ago = 33,333 generations ago
Writing = 10K years ago = 333 generations ago
Computation = 1kya = 33 generations ago
Computing = 100 years ago = 3 generations ago
A mere three generations to remake the cognitive matrix from top to bottom; that’s compressing things quite a bit. Remember, culture is passed on through imitation and education between generations. But you can see where that evolution seems to be heading. It takes an individual 15 to 20 years to internalize the basic ideas and norms of their culture. Barring unforeseen advances in genetic engineering and pedagogy, that process cannot be compressed any more.
What’s happening now? Perhaps it is only the consolidation of Rank 4. Maybe it’s that, but also the emergence of a new rank. Hays and I talked about a fifth rank in private, but never published about it. Conceptualizing Rank 4 was difficult enough.
Still, the question must be asked: Are we witnessing the emergence of Rank 5 culture? And behind that lurks an even more radical question: Are we seeing the emergence of a new Arena of Abundance, the Fourth Arena?
We argued that Rank 4 was grounded in and driven by computing. At the moment it appears that machine learning, deep learning, artificial neural nets, all of them, are doing amazing things, things that are prompting apocalyptic fever in some quarters. Is AGI (artificial general intelligence) just around the corner? Will that yield a world of super-intelligence computers? What then?
I’m sure some of that is just hype and projection, tech-bros and others getting too far out over the edges of their skis. Here we have Holden Karnofsky breathlessly proclaiming we’re living in “the most important century” humankind has ever had. He’s projecting that, over the next 1.4 billion years humans and our intelligent machines are going to people the whole freakin’ galaxy. Really?
Karnofsky’s vision strikes me as something like Olaf Stapledon plus computers. It’s science fiction that’s escaped from the realm of imagination and into the real, not as actual, but as potential, as future. He senses that something Big is afoot and seeks to rationalize it.
Is that so very different from what I’m doing? I can’t help but wonder whether or not something of cosmic magnitude is in fact going on. However skeptical I am about all this AGI-is-coming hype, I am in fact deep into the process of drafting a paper arguing that machine minds are possible and that sketches out the broad requirements for them. Who knows?
More later.