Later On

One can see where Vernor Vinge got the idea of the Singularity. Take a look at this article by John Wilford in the NY Times:

Digging deeper in a South African cave that had already yielded surprises from the Middle Stone Age, archaeologists have uncovered a 100,000-year-old workshop holding the tools and ingredients with which early modern humans apparently mixed some of the first known paint.

These cave artisans had stones for pounding and grinding colorful dirt enriched with a kind of iron oxide to a powder, known as ocher. This was blended with the binding fat of mammal-bone marrow and a dash of charcoal. Traces of ocher were left on the tools, and samples of the reddish compound were collected in large abalone shells, where the paint was liquefied, stirred and scooped out with a bone spatula.

In the workshop remains, archaeologists said they were seeing the earliest example yet of how emergent Homo sapiens processed ocher, one of the species’ first pigments in wide use, its red color apparently rich in symbolic significance. The early humans may have applied the concoction to their skin for protection or simply decoration, experts suggested. Perhaps it was their way of making social and artistic statements on their bodies or their artifacts.

Of special importance to the scientists who made the discovery, the ocher workshop showed that early humans, whose anatomy was modern, had also begun thinking like us. In a report published online on Thursday in the journal Science, the researchers called this evidence of early conceptual abilities “a benchmark in the evolution of complex human cognition.”

The discovery dials back the date when modern Homo sapiens was known to have started using paint. Previously, no workshop older than 60,000 years had come to light, and the earliest cave and rock art began appearing about 40,000 years ago. The exuberant flowering among the Cro-Magnon artists in the caves of Europe would come even later; the parade of animals on the walls of Lascaux in France, for example, was executed 17,000 years ago.

The cave people in South Africa were already learning to find, combine and store substances, skills that reflected advanced technology and social practices as well as the creativity of the self-aware. The paint-makers also appeared to have developed an elementary knowledge of chemistry and some understanding of long-term planning earlier than previously thought.

The discovery was made at . . .

Continue reading.

So 100,000 years ago humans already had started technology and even, on some level, manufacturing. They clearly must have had language by then. So why did it take so many thousands of years to advance, looking at (say) how far we have come in the past 10,000 years? It’s like the pace picked up tremendously. And if you compare the last 5,000 years to the previous 5,000, or the last 2000 years to the 2000 before that, or just the progress since 1000 AD until today—a mere 1,011 years—it seems quite clear that the pace of change and technological development is exponentially increasing. To maintain the pace, one pretty much has to hypothesize superintelligence. Thus we get the Singularity.

But why did the pace pick up so quickly? Why so slow at the start and so rapid now? I’ve been reading William H. McNeill’s A World History (3rd edition) plus the book The Human Web that he wrote with his son, and fairly recently read his book The Pursuit of Power: Technology, Armed Force, and Society since 1000 AD. So far as I can tell from that reading, the key change is the bandwidth of our communication channels.

Those early humans mixing paint had a great idea, but how to get it out? How can they let more people know, so more people are working on it, making discoveries, improving techniques, pooling experience, and in general getting a move on?

They can’t: no written language, no way to travel as “travel.” Their discovery lived within the tribe, and perhaps when the tribe split, two tribes knew it. But so far as spreading the knowledge? Slow going.

Once agriculture began with food surpluses that allowed many to turn their attention to things other than food, the pace picked up a lot: now people were free to get together and work on things. One of the things they developed was writing.

With the invention of writing, the communications channel widened considerably. It becomes easier to store and access knowledge, and to transfer it from place to place because the transporter no longer had to learn what was to be communicated: they can simply carry information in written form.

Having both writing and spare time, people did indeed start to communicate their ideas effectively and efficiently, pooling their efforts and continually building upon what had been previously accomplished (thus the exponentiality—and it utilizes recursion, in a way:  it’s through the exponential growth of technical knowledge that we achieve exponential growth of the communications channels that allow for the…  you see where that goes: very steep curve).

I think the growth of the capacity of the total human communications channel is the vehicle of change. And clearly today the total human communications channel is ENORMOUS, with information flooding the globe all hours of the day or night, both direct personal communications between individuals and work groups and also broadcast communications that go to many simultaneously. At the same time we have enormous capacity to store information and (equally important) can readily retrieve it.

So things move ever faster. It does not strike me as a stable situation. I am skeptical that superintelligences will come to pass, but if they should, we must wonder whether our superintelligent creations will care for us. Perhaps they will detest us and our weaknesses (take a look at Congress today: how would Superintelligence (the networked totality of the superintelligences—for surely they will quickly—at their speeds—decide to communicate: as I’ve outlined above, the benefits are obvious) view that?). And in any event Superintelligence—and it will surely exist, for it’s obvious that superintelligences, having decided to communicate, would be able to figure out how—is likely to develop its own goals and pursue those.  (I think the hope is that we will be able to outwit it if it comes to that…)

I’m sure attempts will be made to incorporate some sort of Asimov’s Laws, but things are not always so easily controlled, and once the first artificial intelligences that exceed our abilities are building the next generation—and then, even more quickly, those build the next, and so on… Well, it seems to me that we won’t really know (or be able to understand) what they’re up to. We already have trouble figuring out programs now that involve neural nets and evolutionary algorithms. Take a look at “Creatures from Primordial Silicon” (PDF).

UPDATE: I am just at the part (page 231) where Charlemagne has nominally recreated the Roman Empire and is crowned as Emperor of the Romans by the Pope. “Co-operation between Byzantines and Franks was never close. Political distrust was exacerbated by persistent religious friction, centering on the proper role of images in Christian worship.”

I thought it odd that such a trivial issue—undetectable on the scale of natural events and thus totally a cultural construct—can have such an impact on us. Culture is our own creation (as humans) and we’re having such a conflict over something we ourselves made? And cultural definitions are low-energy events—as I say, undetectable on the scale of natural events—but they seem to totally control our exertions and energies.

But of course exactly that difference—among many others—continue to be burning issues and have whole peoples at daggers drawn to this day. We are unable to think like a species.

Do you think Superintelligence will be so caught up as we in issues purely of human creation? Or will Superintelligence perceive, plan, and operate at the scale of natural events: physical reality rather than cultural constructs? We don’t give much thought to, say, ant culture, though we do study and understand it. But on the whole we prefer our own culture. Will Superintelligence prefer its own culture to ours? Or that of ants, for that matter?

And will Superintelligence fail to think like a species? Not likely.