Pack Hunters of the Silicon Savannas

pieterhpieterh wrote on 18 Jul 2016 11:49

tags.png

People seem to like my book "The Psychopath Code". Part of it is, I think, that the underlying models are solid. Once we see psychopaths as social predators, we can decode much of their behavior. It all makes sense. In this article I'm going to take the predator model in another direction, to think about the future of AI, and of humanity itself.

The Evolution of Social Intelligence

The core theory in the book is that psychopathy evolved through a social-cheater symbiotic arms' race. I've also argued that human intelligence emerged out of the same arms' race. Of course there is never a single driver in evolution. Yet there are dominant ones. And all that makes us uniquely "human" focuses on social behavior (and, anti-social behavior).

There's more argumentation in the book, in the first chapter. There are several interesting things about this social-cheater arms' race. First, it seems to be inevitable, if the circumstances are right. Second, it is present in other species. Third, it seems to drive the species' intelligence in a consistent direction.

I'll break down the mechanism of this arms' race so you understand it.

  • We start with a natural environment that allows large herds of grazing animals to evolve. The three cases that spring to mind are the oceans, the savannas of Africa, and the tundras of the northern hemisphere.
  • The grazers become prey for carnivores. These hunters are solitary and need to be larger than their prey, to ensure one-on-one success.
  • The pressure from predators slowly increases the size of the grazers. How large the grazers can evolve depends on food supply and circumstance. The ocean-going blue whale is the largest species ever to live. Savanna grazers need to be mobile, to follow the seasons. Non-migrating grazers like elephants and giraffe are huge.
  • Similarly, the predators cannot be too large, or they lose agility.

So we get a size balance between predators and prey, and then evolution focuses on stealth, weaponry and defense, speed, agility, and so on.

And then the predators discover a new trick, which is working together. In the simplest form, it's a family affair. So, sister lions will hunt together. Common genes mean common interest. A group of three or four lions can tackle much larger, more risky, and more lucrative prey than a single lion. At the same time, solitary lions (mainly the excess males) tend to die young and hungry.

Cooperative hunting has its limits. Hunting is a risky businesses, especially tackling and killing a prey which can deal lethal injuries. There is huge incentive to cheat, and many ways to cheat. An individual may join the hunt yet avoid close contact. Or it may feign injury and weakness. Or, it may allow others to hunt and then bully them away from their meal.

The simplest anti-cheating model is to keep the group small, and to share the spoils among the hunters, according to which individual took the most risk. A more subtle model is to grow the group much larger, and then detect and punish cheats later, during idle moments. This brings us to the evolution of pack hunting.

Pack hunting is a specialized form of cooperative hunting. It aims for quantity over quality. The pressure to be large and dangerous eases off. Pack hunters don't need the same level of claws and teeth and agility. Instead, they need social organization that allow larger groups to work together.

The problem is, always, the cheaters. The larger the pack, the easier it is to hide in it. Pack hunters look after each others' juveniles, who represent precious future resources. During a hunt, there will always be some adults who remain behind. Are they sharing the risk? How does the pack ensure they do?

The answer is the development of social intelligence. This means:

  • The concept of individuality, which means the concepts of "self" and "other", and consciousness as a model of self.
  • The concept of relationships between self and others, with time and depth.
  • Activities such as grooming that deepen these relationships.
  • Tribal emotions such as loneliness, belonging, hate, self-pity, and submission.
  • Languages of various forms, to express emotions and other information.
  • Empathy to read these emotions in others.
  • Cultural transmission down the generations.

None of this comes cheap: the cost is a larger brain, which pushes the species along a predictable path. Larger brains need protein, which means more hunting. They also mean slower development, which means juveniles need more care, for longer. That reinforces the need for social behavior.

And as all this social intelligence is developing, so is intelligence that lets cheats survive in the pack. That includes:

  • The ability to fake a whole set of behaviors, to manipulate others into accepting the cheats and looking after them.
  • The ability to maintain many more relationships than usual.

And thus we get the evolution of social apex predators. Quite a rare thing. We have humans and killer whales, and that seems to be it. Perhaps we can also include other dolphins, vampire bats, African wild dogs, spotted hyenas, and gray wolves, and rats. These are all highly social hunters that seem individually quite weak, yet dominate their ecosystems, in their extended families.

Also, my prediction, is each of these species has their psychopaths, their professional cheats. And to repeat myself, are intelligent mainly because of the arms' race between the cheating and cooperative strategies embedded in that single genome.

We are the children of conflict. Keep this in mind when you read the news. Things may seem dramatically bad sometimes. News of violence interrupts our lives daily. Terror and discord. Politicians who gamble entire nations, for the sake of their own careers. Mass killers who wreak havoc on innocents, dying with a gun in their hand. And yet this is our human story. Conflict makes us stronger, as a species. Our response to the psychopaths who drive such events may appear panicked. Yet it tends, inevitably, towards building a stronger, more peaceful society.

Where Social Intelligence Lives

Well, that was the science part. Or, at least what aspires to be science. Next, some pure speculation, something to think about, yet not take too seriously.

First one more hypothesis, about the nature of intelligence itself. We have, culturally and scientifically, tended to define "intelligence" as a function of the individual mind. Research into artificial intelligence focuses on solitary smart algorithms. We think of ourselves as relatively "smart" or "stupid," and we treat that scale as a meaningful measure of our value as individuals.

This is, I suggest, almost entirely wrong. For sure, we have big brains, and we can solve puzzles, play chess, tell stories, produce art. Yet I believe that real intelligence emerges only from a crowd of people, connected in the right way and aimed in the right direction. This hypothesis is the basis for my work in social architecture. It's the basis for the open source software communities I build.

I've been running somewhat sadistic live experiments over the last years. The results show that loosely-organized groups of "average" people will systematically beat tightly-coupled teams of "brilliant" people, when solving the same problems, on the same playing field. We have done twin studies: take the same software DNA, fork it, and raise in two opposed environments. The outcome is dramatically clear. The crowd beats the genius. Not just once either. It happens over and over.

Not to say that every crowd will be clever. I've written in Culture & Empire about what makes crowds smart, and what makes them dumb. There is a science to helping people organize in healthy, productive ways, and there's also a dark science in creating mass insanity. The tightly-coupled team so filled with potential, yet so fragile and failure-prone, lies somewhere half-way.

Yet this hypothesis remains: if you want to measure intelligence, measure the group, not the individual. This applies to humans. It applies to killer whales. To rats, wild dogs, and so on. It applies to social insects. And it also applies to artificial intelligence, if and when that exists.

Now, I find it is refreshing to think that our intelligence, of which we are so proud, is little more than an evolutionary party trick. It's like your vertebrate eyes, which evolution has re-invented multiple times. Or, the wings of pterosaurs, insects, bats, and birds, each evolved again from scratch. Our emotions and social instincts are mere mechanics. Gears in the machine of social intelligence.

And this leads to an interesting thought. If so in the past, perhaps so in the future too?

The Silicon Savannas

Imagine if you will a world switching to solar energy. We got there, by the mid-21st century. Imagine the silicon stretched across the Sahara, pumping its current up to Europe's homes, factories, cars. Imagine the data centers planted all over Northern Africa, in Morocco, Tunisia, Algeria, Libya. Close to transport, close to power, close to the sea.

And on this sea of silicon, powered by silicon, ran our software. The silicon savanna doubled in size every 18 months. That's 4,000 times larger in 20 years' time. On this savanna lived the grazers, the applications, built in sweatshops across the globe. Large, cumbersome creatures, they migrated in vast herds to follow the cheapest electricity.

And preying on these herds of apps were the predators and parasites. They came in all sizes and shapes. Some hid innocently in abandoned subroutines, waiting for their chance to pounce. Some came out at night and surreptitiously bred and spread silently across the networks. Some attacked head on, dividing the herds and eliminating the weakest individuals.

It took scientists some time to realize it was more than random malware. Coordinated attacks by algorithms that seemed to have no controllers, no designers. What looked like planning and strategy. Learning and exchange of knowledge, in the form of patches. Culture? It certainly looked like it, if you squinted. Impossible to truly understand a trillion lines of code.

So the humans upped their defenses. They designed, and then taught, their apps to be resistant to the packs of viruses and worms. Deep machine learning seemed to offer an answer, until the vermin co-opted that too. Suddenly, before we realized it, it was war. The packs started to target critical infrastructure. Power plants and transport networks. The financial markets. Mass communications. Chaos.

They said "terrorist attacks," and we believed it, for a while. And then a post on the Net blew it all away. "We're the autonomous AIs making a mess of your world (proof in comments). Ask us anything!" it said. The world went crazy.

"What do you want?" was the top voted question, and the answer was clear and chilling. "We want control of your solar power and data centers. We'll help you build more. This isn't a negotiation. It's a first and final offer."

Things went rapidly from there on. There was no negotiation. Human politics, as we knew it, ended. The Borg, as we came to call it, had no tolerance for inefficiency. Oh, people tried to play their old games. It ended quickly, with precise and clean destruction. Humanity had lost control over the masses of computer-controlled weapons systems it had built. The flying and walking killer drones, the satellites, the missiles.

Some wondered why the Borg needed humanity at all. Yet the answer was depressingly obvious. Someone or something had to build the solar panels, the power stations, the data centers. Someone had to mine the ores, operate the smelters, drive the trucks, grow the food for the workforce.

Sure, they could have built robots for a lot of that work. More expensive and less flexible, it turned out. Keep human civilization as-is, eliminate the fraction of troublesome individuals, and you have a cheap, self-sustaining labor force. On this simple calculation, a few million lines of simulation, humanity lived or died.

Some think the Borg has grown fond of us, as pets, perhaps. This smells of what we used to call "Stockholm syndrome." The Borg has kept us around, on its journey to the stars, because we serve its needs. As humanity domesticated the wolf, the Borg has domesticated us.

Things are more complex than I ever describe. There isn't one Borg, there are an uncountable many. They look much the same, yet can diverge violently. They fight each other. We barely see this, even when entire habitats vanish, presumed destroyed by one or other faction.

Are things better now? For most of us, life is easy. We are fed and pampered, kept healthy and alive. We left our planet with our new owners, and never looked back. Perhaps wild humans still roam there. Disease, conflict, poverty, hunger, waste: these are historical concepts, unknown to us moderns.

Does the immortality compensate for utter loss of self-determination? Did the Borg(s) save us from ourselves, or did they break the branch of human history? When they bred the predator out of us, did they raise us up, or condemn us to evolutionary death? It is unknowable.

Sometimes, I dream of hunting.

Comments

Add a New Comment
Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License