Superintelligence might be 40,000 years old?

Article continued from previous page.

Human brain thought to compute at the thermodynamic limit

Apparently, as soon as the food is restored, modern chimps become “lazy” and stop using the sharpened sticks. “Lazy” isn’t the right word — thinking is actually hard work. (Which brings us back to the superintelligence. Computation of any kind requires a lot of energy. The human brain is believed, or at least used to be believed, to compute at close to the thermodynamic limit for computational devices. In terms of energy usage, computers, any computer, are much less efficient than the brain. The require vast amounts of energy in comparison, and produce a lot of heat. This is a problem for our self-evolving WBW handwriting robot, because if it is to become superintelligent, or build an army of self-assembling clones (tiny nanobots or anything else), vast amounts of energy will be required. That energy will need to come from somewhere. Even if the power source is something miraculous, the resulting heat will need to dissipated somewhere. (Currently, most energy comes courtesy of humans. They send out bills for heavy users of energy, and for heavy users of cooling equipment.)

A lot of the evolutionary pressure on the human brain revolves around this need for energy efficiency. There is this built-in “habit” system. It is actually quite sophisticated — it can drive you home once you’ve learned the route while the rest of your brain focuses on other things. (So if you forget to pick up something from the store on your way back from work, you can blame evolution and its attempts to save energy.) This habit “autopilot” is much more energy efficient than the rest of your brain. So any kind of routine tasks get pushed off onto it so that the rest of your brain can shutdown and save energy.

This makes brains in general “expensive” tissue, as has been demonstrated in experiments on fish. (Scientists forced fish to evolve larger brains. The brainer fish had fewer and fewer offspring due to the resource demands of the brain.) In humans, a very efficient liver is needed to support the energy requirements of the brain. (This shows up in pop-culture references like the cult film The Matrix, where a superintelligence has reduced humans to giant batteries. The human liver is, in fact, one of the most efficient chemical batteries out there. This is presumably why humans and not some other animal were being used in the film.)

Making and using tools apparently requires a powerful brain. This is demonstrated by research on chimps. When presented with very simple object manipulation puzzles, chimps struggle. They take hours to solve simple object manipulation puzzles that are solved almost instantly by very young human children. Object manipulation, needed for the construction of tools, is in our genes and innate neural wiring. It’s missing from chimps.

Since brain tissue is expensive, and thinking is apparently expensive, it’s only natural for chimps to stop using tools when they don’t need to. So, chimps will quickly stop hunting with sharpened sticks and revert back to brute force as soon as they possibly can. It saves them from the need for the bigger brains and higher energy consumption needs that come what that.

That being said, sharpened sticks are an obvious tool in the chimps natural environment. If food remains scarce, they’ll continue hunting with it, especially the females and younger males that lack the brute strength of the adult chimp males. (One of biggest genetic differences between chimps and humans is in muscle tissue. We have muscles that are much weaker, but more energy efficient, and, presumably, more precise when throwing projectiles like spears and sharpened sticks. Finesse over brawn, presumably.)

Talking becomes more important when the other chimps are all packing sharpened sticks

Once you’re hunting with sharpened sticks, it makes sense to be bipedal. (Hands are for carrying sharpened sticks.) The traditional view, when science was off by millions of years as to the first tool use, was that bipedalism permitted seeing further over vast differences. This is true, but bipedalism has been rarely evolved in other animals if that is it’s only benefit. (Another theory connects cooking to bipedalism. Cooking is very important, but comes a little later in the story, after sharpened sticks. If you’re playing with sharpened sticks, and certainly with later stone tools, you might accidentally start a fire. And chimps already preferred cooked foods, so it is a relatively straightforward innovation.)

As it turns out, male chimps in societies without sharpened sticks will steal food from females and younger males. Lacking speech, they do this without explanation. Often, it is simply to redistribute the food to other members of the tribe. (But, sometimes they keep the stolen food for themselves.) That doesn’t work in chimp societies where the female and young are packing sharpened sticks.  The male chimp had better explain that he only intends to redistribute the food. Hence, language and better communication would be a crucial asset to avoid getting poked — unless you’re not yet dependent on making sharpened sticks. Once your genus does become dependent on the “competent manufacture of sharp edges” as one author put it, you’ll need language to pass the skills on to the next generation.

Chimps have problems with the concept of higher numbers, supposedly due to their inability to understand recursion. Recursion, in this case, is the idea that a word for an abstract concept (a symbol, if you will), can represent not just a real object or quantity, but it can refer to a group of other abstract concepts. Recursion is critical for teaching how to make a group of sharpened sticks. (There might be many different types of sharpened sticks, with different things attached to each sharpened stick. You need words to explain the general concept of different levels of abstraction around sharpened sticks. And how to connect different types of sharpened sticks together.) A chimp supposedly cannot understand the general concept of infinity, where each new number represents one plus a previously defined number. That would limit the chimp’s ability to manufacture a complex set of tools. This inability to handle symbolism would also prevent the chimp from making cave paintings. (And, indeed, although chimps can understand the concept of smearing paint on canvas, they cannot draw things or use the canvas to express meaning the way even a child can.)

As was only discovered this year, our genus began carving symbols into seashells hundreds of thousands or millions of years before the modern human species existed. This information technology has consequences for the superintelligence.

Human genes clearly shaped by our technologies

A few other technologies were essential. Humans have the ability to communicate over long distances using drums and smoke signals (and, later, symbols in fine clothing and trinkets — which also relates to the superintelligence.). Drums (and other instruments) likely coordinated the hunt. Drums certainly scared and manipulated animals. Another genetic difference between humans and chimps involves a special ankle bone that permits humans to run long distances. Humans are natural marathon runnings, and can run other animals to exhaustion.

This may be a natural hunting technique for an animal armed with sharpened stick. Humans are impervious, especially in an armed group, to even the most powerful unarmed animal. The only escape is for the animal to run away, where they might be stampeded off a cliff (as was still practiced until recent times) or run down to exhaustion. But we promised volcanoes and vampires. Time for those to entire the picture.

Killer volcanos a bigger deal than killer asteroids?

It turns out meteorites and asteroids aren’t the only things which cause mass extinctions. Every few thousand years a supervolcano goes off. (The two phenomena might be related. It is thought the Cretaceous dinosaur extinction asteroid set off a supervolcano or two in its aftermath.)

Over the course of human (and chimp & gorilla) evolution, there may have been many such catastrophic supervolcanic blasts.  Genetic evidence shows our species (and that of chimps & gorillas) is severely inbreed. We (humans, chimps, and gorillas) have been down to only 20,000 individuals or less, probably at multiple points in our evolutionary history. It’s one reason we’re all so closely genetically related. (Humans were subjected to additional bottlenecks long after diverging from chimps. Mitochondrial evidence shows all humans descended from a single individual at at least one point in our long evolutionary history. Saying all humans descended from a single ancestor at one point is not quite the same as saying we were down to a single human. That single ancestor might simply have had a very advantageous mutation that spread throughout the population, and competed out everyone who wasn’t descended from that individual. That being said, there is strong evidence for repeated bottlenecks that brought us down to a small number of individuals.)

There were a number of reasons for this bottleneck. Compared with chimps, humans are missing a key receptor in their cells. Generally, the only way a species loses a “critical” receptor is through a catastrophic disease. (The infectious agent was using that receptor as its entry point. Individuals missing the receptor were the only survivors.)

But multiple supervolcanic blasts probably explain most of the genetic bottlenecks. There were many of these over the course of the evolution of our genus.

A big question concerns the evolution of the truly modern human subspecies around 40,000 years ago, which then quickly dominated the globe. It replaced the Neanderthal humans, which had diverged from modern humans perhaps a million years ago and had been successful in their own right over that period.

Did humans explode out of Africa due to a genetic change 40,000 years ago?

Was there a genetic change 40,000 years ago that allowed modern humans to explode out of Africa? (And is this what we meant by the first superintelligence?) Recent evidence suggests there was. Based on skull evidence, testosterone levels dropped around 40,000 years ago in modern humans. This allowed closer cooperation within tribes. (Existing anatomical differences already permitted modern humans to speak more clearly than Neanderthals, but that is a much older innovation.) The finer clothes and better weapons of moderns humans may have resulted from this closer cooperation.

But what triggered this closer cooperation? Were they banding together to protect themselves from more primitive humans? Intriguingly, there was a supervolcanic explosion around 40,000-50,000 years ago. Extreme cold weather may have triggered the testosterone drop amongst these hot-weather-adapted humans. Only humans that worked together in societies may have survived.  (There were earlier supervolcanic eruptions. Perhaps an earlier cold weather spell forced the initial human-chimp split by forcing the human line to adopt sharpened sticks to survive.)

It is not clear why killer asteroids get all the attention in the media. Supervolcanoes are a much more frequent event. Perhaps we are more comfortable with asteroids, as we can potentially develop technologies (and blockbuster Hollywood action films) to destroy an asteroid given enough warning. (The early warning systems are generally not yet in place, given the number of number of surprise meteorites impacting Earth each year.) Volcanoes, however, cannot be predicted at all.

(That may not quite be true. Microphones may be able to detect the resonant frequency of a volcano to give brief warnings in the future. But the warning, for a supervolcano affecting the entire planet, would be far too short. And what to do? Nuke the supervolcano? We’ve argued in the past that pre-emptive Earthquakes might release just enough energy to stave off larger events. We’ve also argued for—very expensive—deep earth pressure sensors to help predict earthquakes. These might also detect pressure building up for a supervolcanic eruption. Perhaps, using nukes or some other technique, the volcano could be prematurely detonated, nipping the more dangerous supervolcanic eruption in the bud. These technologies are for the future.)

Suffice it to say, we live on a very dangerous planet. As the true evolutionary history of our species will attest, there is no such thing as the status quo on this planet. Preservation of existing species is very difficult, except through data. (We’ll likely take up this topic again.)

Vampires are related to supervolcanos, steam-powered computers & the superintelligence???!?

So, volcanos check. We still promised vampires.

Well, as it turns out, there was a recent (but relatively mild) volcanic supererruption. This is the so-called “Year without summer” of 1816, which triggered world-wide famine, religious movements, and other phenomena. (The resulting shortage of horses may have led to invention of bicycles, and, ultimately, automobiles.) “Year without summer” is actually a misnomer — “Year of extreme weather” would be more accurate. Hot summer days would be followed by crop-destroying frosts in July. (The cause of the weather disturbances were not understood at the time.)

This is where we get to our vampires. (Not real-life vampires. We’ve already talked about those. We’re talking about fictional vampires here.) It turns out that Mary Shelley and Lord Byron were vacationing around the time of the “Year without summer.” Instead of the bright and sunny days they expected, it rained endlessly. Apparently, when you’re in the 19th-century one-percent, you don’t starve during a famine. Rather, the thing to do when your vacation is ruined by a supervolcano is to have a competition: who can write the scariest horror story.

Shelly wrote Frankenstein and Lord Byron wrote The Fragment. The latter would later inspire Stoker’s Dracula. (Vampires and volcanos, check.) Shelly is said to have won the competition. Incidentally, even their offspring could establish quite the 19th-century brain trust. Shelly’s daughter would become the noted feminist author Wollstonecraft. And Lord Byron’s daughter is none other than Countess Ada Lovelace, first steampunk computer programmer and extremely popular with our Instagram following. (Vampires, volcanos, early feminist writers, and steam-powered early computers, all together in the same room, check.)

Back to the superintelligence

So that brings us back to the super intelligence. Is it really 40,000 years old?

We started off by suggesting that convergent evolution would dictate the design of any true super intelligence. Readers were right to assume that any intelligence would somehow need to be like us. (Specifically, the goal evaluating system would need to be sophisticated to exhibit any kind of truly intelligent behavior.)

Putting together an nanobot army with just one hour of Internet access does not seem thermodynamically feasible. (The laws of physics apply even to superintelligences.) Nanobots need to be constructed according to quantum mechanical principles. These are extremely computationally difficult, requiring massive amounts of computer time. (And, as we’ve already noted, computation, of any kind, requires a lot of energy due to thermodynamics.) The super intelligence can’t run the computations itself, of course. (As a newly evolved being, it’s using all of its computational resources to power its own intelligence.) It needs to find massive computer power on the web in an hour without anyone noticing the sudden energy drains & heat emissions. (Is that fan on your Mac going off your browser? Or is it the super intelligence running the quantum mechanical calculations for a new nanobot army)? This is unlikely to happen without anyone noticing. (We are, after all, already looking for this kind of activity on our networks.)

Perhaps if such killer nanobots already existed, it would be easier. Industrial control systems aren’t supposed to be connected to Internet (don’t connect your robot nanobot army to the net, please!) But let’s say someone got sloppy, and our super intelligent hand-writing robot hacked into a nanobot control system. Since the killer nanobots already exist, things get much easier. (Although, in the case, since humans invented the technology, presumably they would be monitoring for abuses. Surely someone would be preparing countermeasures for an attack by the killer nanobots just invented?)

Where is all the energy coming from to let the nanobots and the superintelligence wipe out humans?

One final problem: with the killer nanobots wiping out humanity, where is all the energy coming from to power this emergent intelligence? Remember, high intelligence (of any kind) requires a lot of energy. (And technologies like dissipation of heat. Did we mention that humans are extremely good at getting of rid of excess body heat, one reason our ancestors tried to run down other animals to exhaustion? Our sweat mechanism is extremely highly evolved. But we digress.)

Article continues next page.

1 2 3