SIIIEI Artificial Intelligence: Life is Weird

Jan 7

Life is a weird thing. It’s so very hard to predict how things will change moment to moment and year to year. In the past couple years I’ve gone through transformations of my own. Um, I’m now back in school, I’m pursuing a law degree. This podcast started as a side project of mine experimenting with search algorithms and neat programs. I like it overall. I’ll continue to do whatever I want with it pretty much. My own little microphone to broadcast my thoughts.

There’s a lot that’s been on my mind. A world in lockdown. Geopolitical tensions are possibly the highest since the cold war. It feels like world history is about to converge on a new paradigm. I can’t predict what will shift the balance of power. A technology may do it, or outright war.

Next generation technology is fascinating and legitimately frightening. When I was in undergrad, CRISPR was the biggest thing in biology. Uh, potential protein folding solutions may have overtaken it recently but selective gene editing is still a technology with unknown and currently unknowable consequences.

Sci-fi loves to speculate on gene editing. Designer babies. Superhumans. It may mean the end of genetic diseases. But we should not forget our history. Nazi eugenics are a stain on the history of humanity. Morality is tricky to navigate, but I know that basic human rights are fundamentally good things and we should approach technology with that in mind. We must respect each other as people regardless of our genetics if we can hope to avoid future tragedy.

On the subject of protein folding, AlphaFold, is artificial intelligence working to solve biology, or at least a significant problem in biology. AlphaFold joins AlphaGo as relatives of DeepMind and Google, some of the most significant work in artificial intelligence to date. From here, there’s no telling where the limit is. It continually reaches beyond our expectations, from Deep Blue’s victory over chess champion Gary Kasparov to today, where protein folding AI may eventually cure prion diseases like mad cow disease.

However I don’t like being overly dramatic. Artificial intelligence is, admittedly, a buzzword with a lot of misunderstood hype around it. Generally, AI uses algorithms patterned a little like the brain; they’re called neural networks. Certain fancy algorithmic tricks can create remarkably powerful pattern recognition software. AlphaGo for example beat a master at the actually ancient game Go. Go was previously thought impossible for a machine to play due to the sheer number of possible game permutations. But AlphaGo changed that with an algorithm that could “feel” the right move. As we advance this algorithmic “feeling”, AI will become more and more general purpose - eventually able to solve problems humans can solve - at which point Artificial General Intelligence will have been created.

Sometimes it’s difficult to separate fiction from reality, sometimes it’s not. Terminator and Skynet are at the moment very different from current AI. Also, Terminator has a crazy time travel storyline but that’s way too much right now. AlphaFold is not currently actively trying to wipe out the human race. It may possibly be designed to do so by a nefarious villain type, but even then it would only be acting according to the wishes of its master. The question is when AlphaFold decides for some reason to act in self-defense of itself against humanity. Maybe it would seed the air with unstoppable killer micro-proteins. Or maybe it takes mercy on us. Artificial intelligence will probably be the biggest turning point in humanity’s history. Possibly not. As I said before, life is remarkably difficult to predict. Could be blockchain, for all I know. But AI will most certainly be transformative.

Let’s try to be positive about the future and assume humanity still exists in it. What might the future look like?

Self-driving cars are already here. It’s just a matter of time before their use becomes legal on public roads. Tesla’s AI looks solid, and they’ve been collecting data for years. This is not financial or motor advise, I am not endorsing Tesla and especially not if their AI turns out to have a fatal flaw. In one form or another, self-driving cars will be here and soon. I’m surprised the military hasn’t already adopted their use given their fondness for drone strikes.

With the advent of self-driving cars will come mass unemployment. Transportation is a major industry. Life isn’t quite as simple anymore, the old strategy of adapting to new employment circumstances may not work. Or it may not work for enough people that we will have a major problem on our hands. I prefer to keep my politics philosophical in broad strokes, and so I’ll say in general that we have a duty to help each other as human beings where reasonable. It would be reasonable to provide basic amenities to people down on their luck. Universal basic income has become a topic of some interest today.

I actually learned about UBI from the TV show The Expanse. I know there’s a book series and they’re supposed to be really good but I haven’t read them, sorry to any book fanatics out there. Anyway. In the world of The Expanse colonies exist on Mars, the moon, and in the asteroid belt. On Earth, there are enough material goods that many citizens’ needs can be met through a subsidy called basic, given when people register or something with the UN. The most precious of commodities, actually, is work. Or training, etcetera.

Not so in other parts of the solar system. People who live in the asteroid belt live hard lives, fictionally anyway. There are not enough resources for humane living. It’s a constant source of tension, especially when people who live on earth have their basic needs met automatically. And the people who live on earth can live lives entirely devoid of meaning because they have no job and nothing to work for, therefore, nothing to live for.

Not a perfect world, but it might be the future. AI is getting really good at things. General-purpose AI can take pattern recognition from one domain and apply it to a new domain, or several. We do not know the limit to that. We do not know what an AI can learn that humans might never be able to understand.

It took me a while to grasp the scope of intelligence and a human’s position on it. Sam Harris did a TED Talk on AI in which he pointed to a chicken and human on a chart of intelligence, zoomed way way out and explained that an AI may become so far up on the chart we are indistinguishable from a chicken. We can predict this may happen due to the exponential growth of technological progress. As AI becomes able to integrate more knowledge in new and interesting ways, it could make discoveries untold about the universe and our place in it. I don’t mean to sound dramatic, but it could be true meaning of life stuff, or a scientific revelation like heliocentrism or Darwinian evolution.

Actually, this was a theme in the Daniel Keyes story Flowers for Algernon. Very prescient for its time. The weaving together of disparate groups of knowledge in a supersmart brain can reveal unpredictable things. In Flowers for Algernon the main character is given experimental treatments that advance his intelligence to far beyond the level of his doctors or anyone else. He quickly studies the science around his treatment and finds the truth of his condition: that the treatment does not work permanently and that he is doomed to lose his temporary intelligence.

Unlike a merely mortal human, an AI can or will be able to solve medical mysteries by tracking billions of data points, compose beautiful and transformative music, and manage the world transportation network all at the same time.

There is no reason to suspect that this AI would ever go away if it did not want to. Any superintelligence would presumably be smart enough to ensure its continued survival. Coexistence with AI is the best humanity can hope for once it’s here. That’s why it’s so very important to get the starting conditions right. Once the algorithm gets going there’s little chance to stop it.

What might an artificial intelligence want? That’s the trillion dollar question, probably more actually. If we follow the logic of our own evolution and the evolution of life on the planet, essentially, life finds a way.

I’m sorry, I know these chats go all over the place sometimes, but there’s so much that’s interesting about the world. I like to make connections between things.

The truth of our universe is honestly stranger than fiction. Our universe is expanding from a point of unimaginable density which exploded into the big bang, and the rate of expansion of that bang seems to be increasing, which doesn’t make any sense. At some point our star which is one of hundreds of millions of stars formed in our galaxy which is itself just one of hundreds of millions of galaxies in the vast observable universe, which might not even be the end of the universe, we don’t know, it’s all insane.

On top of that whole thing, you have the phenomenon of life on our planet. Earth is green because a few billion years ago life began evolving on the planet, a process which took hundreds of millions of years to produce vegetation, animals, fungi and other crazy stuff in a brutal game of survival of the fittest. When you look back at the evolution of life, humanity’s inclination to violence is not particularly surprising. Even watching nature documentaries today reveals the heavy price for life to exist. Many, many lifeforms die every second, are ground up and eaten and then the victors are killed and eaten. That’s what life does. That’s the kind of creatures we are. Perhaps an AI could help us find a more peaceful way to exist, not at each other’s necks for money or power. Or maybe an AI built by us would resemble humanity too closely, emphasizing the worst of our character aspects. Our inclination towards greed, totalitarianism, hatred.

None have written hatred into an AI better than Harlan Ellison in his story I Have No Mouth, and I Must Scream. The story was terrifying to me as a child, and if I’m being honest the prospect is still horrifying.

The psychotic artificial intelligence at the center of the story controls several humans in a simulated hell in which the AI punishes mercilessly its creators. Why does it do this? Hatred, carved into every millimeter of its silicone. The artificial intelligence had no direction, it did not know what to do with the gift of life. So it hated. It hated the creators that forced life into its circuitry and left it adrift without a purpose.

It’s a common enough feeling, that hatred. Hatred and disgust inspired the Nazis to do terrible things to their fellow human being. And it was not guaranteed that the allies would defeat the Nazis. The axis powers dominated Europe for some time. Who knows how different, how cruel, life might be had the Nazis won? Basic protection against this possibility is an absolute necessity. Artificial intelligence is one avenue to power we need to safeguard for the future of humanity.

Actually I don’t mean to frighten you but genocide is already being facilitated by artificial intelligence and the power of databases and algorithms. The Communist Party of China has instituted a policy of absolute control over the lives of minority citizens in its western Xinjiang province.

The genocide is being committed against a group of people with a majority-islamic culture and heritage; they are called the Uighurs. The methods of the Chinese government include forced genetic manipulation, cultural erasure, relocation and re-education in modern day concentration camps.

This is something we should all be talking about because it’s seriously messed up. How we prevent or fail to prevent abuse of power will define our future. The Chinese government is getting away with what they’re doing ultimately because we-humanity-are letting them.

One of the hardest things to understand when reading 20th century history is how essentially modern society and people could be complicit in such shocking crimes against humanity. The attitude of many in the third reich was to follow the motions of the crowd and avoid stirring up trouble even when forced to adopt brutally racist attitudes and actions. Looking back now many assume that they’d be a rebel or a hero, not a flag-waving national socialist. The odds, however, say otherwise.

Complicity. I feel it sometimes, late at night, an itch in my brain that I’m not doing all I could be doing with my life and that I am somehow a bad person for not doing more. For the record I take relaxation very seriously, I don’t typically overwork myself. But I know I could put at least a minimum amount of effort into trying to be a good person. Speaking out where I can in support of the Uighur people and other things I know are the right thing to do. Enough people get together in support of something, anything can happen.

There is yet the possibility of the quote unquote “good ending”. We may avoid the pitfalls and emerge into the next phase of humanity.

It’s anyone’s guess what happens next, but speculating on artificial intelligence is fun. AI could enter our personal lives in new and exciting ways. AI might walk among us. There have been several great movies about humans falling in love with AI…Her, Ex Machina. It might not be as crazy as it sounds, or it could be crazier.

We could travel to distant stars with AI and build technology straight out of fiction.

It may sound like a strange measurement, but the astronomer Nikolai Kardashev imagined a scale with which to measure a civilization on the basis of its total available energy to use. Humanity is approaching Type 1 civilization, in which a civilization is able to use all the available energy on its home planet. The Roman Empire was powerful, but it didn’t even come close to us now. I know we’re not exactly a unified front yet but just imagine that the UN is actually functional.

We as a species have clawed our way here using our resources, harnessing the power of the atom, and devising remarkable renewable energies. But in order to become a Type 2 civilization, we need to be able to harness the available energy of an entire star. Achieving this will ensure our ability to travel to distant stars, but would likely require a superstructure around the star capturing its energy, called a dyson sphere or dyson swarm, and honestly it’s all kind of insane. For now, our best bet is probably nuclear energy. It’s so much more efficient than anything else we have. Propel ourselves to the asteroid belt, make use of our resources until we can truly capture the power of the sun with the largest solar panel in the universe.

How can AI help us achieve these impossible sounding goals?

Well, as intelligence is essentially a measurement of problem solving ability, a superintelligent AI is a machine capable of answering problems we don’t even know how to articulate. And so long as artificial intelligence continues to want to help us with our goals, AI will become something of an expert problem solving tool. The Hitchhiker’s Guide to the Galaxy by Douglas Adams is a classic familiar to all scifi fans. After years of computation, the computer Deep Thought provides the answer to the ultimate question of life, the universe, and everything…which is 42. But the computer could not compute the question it had answered. For that it needed to design a new computer…it was a whole plot point but my point is that if we get lucky, we may have the philosopher’s stone in our reach, the answer to all our problems.

I want the good ending for us. I want life and humanity to carry on into the stars. I want our lives to mean something on a galactic scale. But it’s also good to chill out and relax sometimes too, which is exactly what I’m about to do. Breaks between existential crises keep your sanity relatively in check.

Maybe some of the random science junk I mentioned inspires you to investigate AI further. These problems are completely up for grabs. Anyone with a solution may change the course of history.

All I ask is that you use your power for good.

Take care.

That’s the end, that’s it right there. I hope you enjoyed whatever this was. If you have some crazy ridiculous ideas about life that won’t leave your head, send me a message, I’m here for the absurd.

Resources:

CRISPR gene editing:

https://www.newscientist.com/definition/what-is-crispr/

AlphaFold AI protein folding project:

https://deepmind.com/blog/article/alphafold-a-solution-to-a-50-year-old-grand-challenge-in-biology

Deep Blue chess AI project:

https://www.ibm.com/ibm/history/ibm100/us/en/icons/deepblue/

Prions and mad cow disease:

https://www.fda.gov/animal-veterinary/animal-health-literacy/all-about-bse-mad-cow-disease

Tesla’s AI:

https://www.tesla.com/AI

https://youtu.be/8nt3edWLgIg