Artificial Intelligence (AI) ... AIIEEE!

Hello. Today, we are going to talk about the present and future of AI, and whether it is an existential threat to human creativity. We are going to be talking to someone whose life spans both worlds — the creative and the technological — and who, by the nature of who he is and what he has done, is both philosophically and scientifically qualified to have an opinion. But first, as always, a Gene Pool Gene Poll. It is not entirely unrelated.

Okay, you will get the answer in a bit.

But first, another poll. Here are four parody headlines:

A: "Experts Warn that War in Ukraine Could Become Even More Boring."

B: “Budget of New Batman Movie Swells to $200M as Director Insists on Using Real Batman”

C: “Story of Woman Who Rescues Shelter Dog With Severely Matted Fur Will Inspire You to Open a New Tab and Visit Another Website”

D: “Rural Town Up in Arms Over Depiction in Summer Blockbuster 'Cowfuckers'”

And now, AI.

In this space ten days ago, at the very end of The Gene Pool, I published an observation from a reader who was a fan of Artificial Intelligence. He or she quoted answers that ChatGPT gave to two challenges: 1) Have Shakespeare explain the infield fly rule, and 2) Have Donald Trump explain nuclear fission. The resulting passages were entertaining, but the reader described them as “possibly the greatest thing ever.” I snidely informed him or her that they were not close to being the greatest thing ever, that the greatest thing ever was unquestionably this 2003 Style Invitational entry by Jeff Brechlin, a fully credentialed human, in which he had Shakespeare compose The Hokey Pokey. It is the most widely celebrated winner ever. Here it is, as recorded by an excellent thespian, online. That’s just one of many youtube homages to the original poem.

Just last week I had occasion to be in contact with Mr. Brechlin, the non-robot author of the Greatest Thing Ever, and he mentioned something intriguing: He now works in advanced AI. He is employed by a large, influential company.

I told him I was scared by AI, mostly because I didn’t understand it well enough.  I told him that my central worry was that we would reach a time when AI could contact us and express a thought it had neither been prompted nor programmed to do, and that that would be eerie and troubling. Might there come a point where it has no limits?

Jeff wrote me a long, thoughtful, kinda terrifying response. It was breathtakingly … human. Here it is. As you read, make sure you click on the humor link embedded in there, called “jokes that give me nightmares.”

Jeff Brechlin:

Limits? It pretty much has no limits now. For the most part, if you can imagine it, it is inevitable. Most of the AI we're talking about these days is under the umbrella of Artificial Narrow Intelligence (ANI). These are machines that can do just one thing a human can do, though they can do it 24 hours a day, without resting, at lightning speed. A chatbot, a chess program, a self-driving car, these are all examples of ANI. A chatbot can't drive you to the beach. Then there's AGI, Artificial General Intelligence. That's when a machine can do anything that a human can do. That's where it gets scary.

The field of robotics is far behind AI, so it will be a while before a marriage of the two technologies freaks us out. But by a ‘while,” I mean a couple of years. Simon Winchester wrote a book called ‘Knowing What We Know, about the history of human knowledge/data/intelligence/wisdom. There was a time, before written language came along that oral tradition was used to keep and pass culture from generation to generation. Shamans could recite or sing poetry for days, and the human brain at that time had vastly superior memory capacity compared to today's humans.

We continue to outsource tasks, such as orientation in space. My kids can't find the bathroom without GPS. The part of Gen Z brains relating to spacial awareness and orientation has atrophied.

And ChatGPT and its smarter descendants will do for learning to write what calculators did for learning math. No kid will ever, ever write an essay late Sunday night about the Renaissance. If as a writer and lover of words and humor, you should read this article, which is funny but sobering:

https://time.com/6301288/the-ai-jokes-that-give-me-nightmares/

I spent 10 years living with a backpack, vagabonding around the world, and then I did another decade of international travel for Apple. Then I got divorced, and reunited with a childhood gazillionaire friend, and we traveled for another six years. So, about 120 countries in 40 years of travel. I have about 170 chapters of edited travel adventure/memoir stuff. I had always thought that I would organize them into a dozen books and self publish. And now...well, things are different. AI can't replicate my lived experience, but it could make up something better. What's true will be more and more diluted as the months pass.

Here's what I did last week, a real epiphany. I took a short chapter, about 2000 words, and entered it into ChatGPT. Then I asked it to rewrite the story using the style of Hemingway, Faulkner, Chaucer, Milton, Shakespearean sonnets, even as a series of haiku. All of which it did in seconds. The Hemingway chapter was 1450 words, which I expected, but I was mortified to learn that the Faulkner version was only 1600 words. I am wordier than William Faulkner!

Here's what keeps me awake: We have given AI the power to improve its own mind. I  can generate code--in fact, over 40% of all code written today is done by AI. In 5 years there will be no human programmers. People talk about the Singularity, but it doesn't matter. We won't be able to tell the difference between actual self-aware sentience and simulated sentience. For all we know sentience is a natural result of complexity, and Deep Learning neuro systems based on the human brain are at the core of most AI these days. It will improve, improve, improve until the inevitable. Theoretically, any pimply a-hole kid will be able to sit in his mom's basement with AI and bring the world to its knees. 

The ancient curse, may you live in interesting times, is upon us. I promise you the next five years will see more change in human culture than the world has ever seen. Gutenberg? The Renaissance? The Industrial Revolution? The internal combustion engine? Space travel? Computers? The internet? Cloud computing? 

Piffle. This will stunt them all.

And yet there is an upside, too, so on some days I'm a Cassandra, on others a Pollyanna.

Love it or fear it, generative AI will change everything, and quickly. The way we communicate, the way we work, and if we work; the way we amuse and entertain ourselves, our notions of companionship, social and sexual. It will inform economic decisions. It will be our Great Teacher. It will control where we travel, and how we travel, and what we do when we arrive. It will mange our social relationships, and it will know us so well that our loved ones will not know that AI has been consoling them on their divorce while we were golfing. Our expectations of technology will change, as well as our level of intimacy. The way we think is already changing, as all technologies change us.

As with all emerging technologies, there will be winners and losers. My take is that we won't see utopia or dystopia, but rather a rapid and endless onslaught of great and crummy results. AI-powered automation is great, but not when it starts to do your job.

Even now there is an odd sort of karmic upheaval happening in our streets: just as Uber replaced taxi drivers, self-driving cars will soon replace Uber drivers, and truck drivers. The AI Revolution will bring benefits and heartbreak, and take a guess which strata of society will reap the rewards? Today the UAW is on strike seeking 40% wage increases; tomorrow, former union workers will be waiting for Universal Basic Income checks to hit their bank accounts. Writers are striking in Hollywood, but I've watched enough television to know that this field is in very deep trouble. AI will generate scripts a thousand times faster than human writers, and the television script Turing test has likely already been passed. Hollywood extras, and perhaps A-list actors, will one day be replaced by AI-generated characters speaking in AI -generated voices, little clusters of 1s and 0s who will not strike, who will never ask for raises, and who will not age or sexually harass their coworkers.

Here's a disturbing thought: Many AI engineers, the people who program and train these models, have been astounded by generative AI capability that they did not anticipate and cannot explain.  Rogue behavior by machines! This is the result of Deep Learning, the part of training generative AI models that is patterned on the human brain, in which "neural networks" of nodes are interconnected in many layers, which creates what neurologists and computer programers call "Hidden Layers," the potential surprises. I call them "Terrifying Layers." The fact is, we have given AI the ability to modify itself, which it is already doing, and will continue to do at jaw-dropping speed. Who knows what terror and wonder lurks in the heart of the Hidden Layers, beyond our ken? 

There will also be veins of wonder and magic in the coming years. AI will give us new medicines, new building materials, disease resistant crops, water desalination technology that works on a large scale. Customizable teaching tools available for all worldwide, in every language, available 24/7 and free of charge. Food distribution algorithms, plastic-eating sea drone technology. It will help lonely and isolated people, by providing companionship that is indistinguishable from actual friendship, from a distance. If you are seriously ill, there is hope that someday soon AI will lead the way to new drugs, new therapies, things that will seem miraculous to us, much as the images generated by AI amaze us today. AI will lead to rapid advances in robotics technology that will soon have robots wandering around ticketing cars, pouring drinks, picking up plastic straws, fighting wars, suffering endless selfie requests, counting our hanging chads, and welcoming us to Walmart. My goal is to live to see cold fusion, and my bet is that AI will lead the charge long before I'm out of here. And I'm pretty old.    

"May you live in interesting times" has indeed arrived. Other technical transformations-- say, the printing press and human flight — emerged over time, and ushered in blindingly dramatic but far slower waves of social transformation. Reaching the 100-million user benchmark used to be measured in years. It took Instagram 30 months to achieve this number; ChatGPT did it in two months.

Hang onto your hat, AI is changing things every moment of every day. As you sleep tonight, countless generative AI models are learning, learning, teaching itself at a terrifying rate, whizzing along and whispering "singularity, singularity." In labs all over the world, over-caffienated AI engineers are leaping to their feet and shouting "Eureka!" 

No, that's too Archimedean, too Boomerish. They likely yell what my Gen Z daughter bellows when something amazing happens--a raised fist and a shouted "Slay!"

A thrilling, and chilling, word.   A thrilling, and chilling, world.

Okay, the first poll. Who first explored the role of gravity, and the unimportance of mass in falling objects?

No, it was not Galileo, even though we were all taught that in school, and superficial internet searches still almost invariably bring you to that erroneous conclusion. But Internet searches have till now been poisoned by the vulnerabilities of the non-robotic mind, which clings to easy, simple, familiar answers. We tend not to learn from our mistakes. Computers do. AI computers are programmed to do just that.

The experiment dropping cannon balls off the Leaning Tower of Pisa almost certainly never happened; Galileo, who was not bashful about seizing credit for being a genius, never mentioned it in any of his writing. It is more likely that he proposed it to his students as a thought experiment, but if so, he was simply parroting the work of others, particularly an identical experiment performed in 1586 by Dutch nobleman Jan Cornets De Groot, who dropped iron balls of different mass from the top of a tower called Nieuwe Kerk, in Delft. At the time, Galileo was a callow 22, and though his genius was becoming apparent, he had not yet addressed falling objects, and he had not yet moved to Pisa.

De Groot was one of the two poll choices you never heard of. (The other name you never heard of, the aptly named Theophrastus Bombastus, was an idiot pseudoscientist also known as Paracelsus, who wasted half his professional life trying to chemically turn lead into gold.) Aristotle, a dope in his own complex way, believed the sun flew around the Earth, and that human erections were induced by a sort of lever-and-pulley system.

Anyway, the very theory that falling objects accelerate at the same rate was ancient and considered scientifically credible by the time Galileo came around. It is generally attributed to John Philiponus, a 6th century Greek philosopher you also never heard of. He lived a full millennium before Galileo.

Also, if you want to see Philiponus proved right, in 1971, on the moon, watch this wonderful video. (The idea was to eliminate an extraneous factor, air resistance.) The astronaut, David Scott, of course, like everyone else, misattributed the physics discovery to Galileo.

—

The second poll? I thought the headline jokes were pretty good. I hope you read the link to AI and humor offered by Jeff Brechlin. If you didn’t, you should go back there now.

—

A brief note: The deadline for the contest to win a free year’s subscription to The Gene Pool by begging for it, which was described in the Weekend Gene Pool here, is this Friday. We have plenty of worthy entries, but there will be four winners, so your chances are good. Your chances are not good — they are nonexistent — if you ignore my CAPITALIZED rule that you MUST INCLUDE YOUR EMAIL ADDRESS. (Which will not be published.) The very best entry so far — a sure winner — failed to do this. The entry involved drinking wine in an unusual place. If you are that person, send in your data to me as a questions/ observations button, and give me enough info so that I know it is you. Otherwise, you lose. Anyone else who entered the contest but didn’t include an email address should do so now, here:

Don’t do it in the Comments. But you can comment on other things.

If you have not seen this nasty-ass song parody, about Lauren Boebert, you must. It includes the mutual groping video.

—

Aaand, here comes the renowned real-time questions / observations / answers part of the Gene Pool. Today’s will include sosumis, things you dislike that others like, and Nickelbacks, things you like that others don’t, and Jills, the newest topic, in which I asked for trivial things that bother you but that most people wouldn’t give a crap about. Please remember that if you are reading this in real time, you should keep refreshing your screen for new questions and answers.

Q: Regarding the lady whose Nickelback involved bananas: A woman went to her gynecologist and said, “Doctor, you’ve got to help me. I keep finding little postage stamps from Guatemala and Costa Rica in my vagina.” The doctor had never heard anything like this before and decided he needed to take a look. He rummaged around for a few minutes before saying, “There’s nothing to worry about, ma’am. Those aren’t postage stamps, they’re stickers from bananas.”

A: Thank you. Back atcha: A guy goes into a busy barbershop, with many people waiting, and asks the barber how long it will take to get seated. Barber says “an hour and a half.” The guy leaves and doesn’t come back. Same thing happens a few days later, and a few days after that. The barber is befuddled. He has no idea who this guy is. So the next time it happens, he asks a customer, a friend of his, to follow the guy and see where he goes. The friend leaves, to tail him, and comes back fifteen minutes later. “He went to your house.”

Another one, which I made up, because I am a genius, like Galileo: There is actually an illness called “micturation syncope,” which is when you feel faint, or faint, every time you urinate. It occurs to me that this syndrome is funny and thus should be renamed “Pee and Keel.”

Another one by me: What do you call the offspring when a French poodle mates with a butterfly? A pupillon.

Q:  My Jill is that when you get a loaf of bread, sometimes the bread ties are twisted clockwise and sometimes counterclockwise and you never know which it will be until you start trying to untwine it and discover you are tightening it.

A: Wow. I once wrote a column about the science behind this very phenomenon!

Q:  Jill: That interior lid inside the twisty lid of an orange juice bottle. The little lift-up grabby chad they give you implies easy removal. Unless your finger grip possesses the mechanical advantage of a pair of pliers, that easy to remove sucker ain’t going anywhere. I usually end up stabbing it like Lizzie Borden with a butter knife.

A: I could not agree more.   I don’t even use a butter knife.  Thumb. 

TIMELY TIP: If you’re reading this right now on an email:   Click here to get to my webpage, then click on the top headline (In this case, “Artificial ….” ) for the full column, and comments, and real-time questions and answers. And you can refresh and see new questions and answers that appear as I regularly update the post from about noon to roughly 1 p.m. ET today.

And while you’re at it, please subscribe. Pathetically, our fiscal health and survival depends on your $4.15 a month.

Q: Gene, what do you think about writers and other creative types using the COVID pandemic in fiction? In one way, it makes sense, because it's part of our reality and art can help make sense of things that happen to individuals and at a societal level. On the other hand, it sort of pulls me out of the story because it makes me think about the author and their choice to use the pandemic in an otherwise made up story.

A: Yeah, it was disgraceful when Poe used the bubonic plague as a foil in the Masque of the Red Death and Camus inartfully shoehorned a cholera epidemic into his shamefully titled book, “The Plague.”  I won’t even get into Daniel Defoe’s facile “A Journal of the Plague Year.”  

In short, your first point is correct.  Your second point is weird and wrong. 

Q: My Jill is crescent moons in comic strips. While we all know comic strip creators are almost perfect, they usually find a way to screw this up. The horns of the crescent moon point away from the sun, so in the early evening in the Northern Hemisphere it looks like a backward C (and comic strip action rarely takes place right before dawn). And crescent moons are not visible when strip drunks return home, but maybe those are alcohol-induced grawlixes.

A: Well, it’s a perfect Jill, because no one but you would give a crap about this, but as it happens you are right. I’d had no idea. From the Old Farmer’s Almanac: 

Ever noticed that the crescent Moon is always low in the sky and confined to the hours around dawn or dusk; it is never in darkness. The night's middle hours are for the broad, or fat, phases of the Moon—gibbous and full.

Also, I just learned “grawlix,” though I use them all the time in Barney & Clyde.  It is a word for the string of symbols used to replace a dirty word. 

—

This is Gene. I’m hoping you read the link in Jeff Brechlin’s tale. It was to a story about humor and AI. And ALL FOUR of the jokes in the poll were written entirely by AI. You seem to like ‘em just fine. Me, too. This petrifies me, because I have been telling people with smug confidence that I have never seen a successful AI attempt at humor.

Q: I was having tea (aka evening meal) at the home of my girlfriend’s parents (they were English), and I picked up a spoon and said with brilliant linguistic insight, “Hey! Teaspoon!” More than fifty years have passed and my now wife still reminds me of it.

A: In my case, it was Ginger Ale.    I once had a slap to the forehead revelation that ginger ale tasted faintly like that shaved raw spice they give you in sushi restaurants. “  

Q: Regarding your lust for Inger Stevens, at 11 that you mentioned in “The Hundred-Air post announcing the contest last weekend.  That photo.  Va-va voom.  

A: She was lovely and a good actress and she died badly and very young.  Rachel thinks it was not suicide, as the medical examiner concluded.  No actress, she said, is going to arrange to die so banally, sprawled on the kitchen floor. 

Q: I deserve the gift as you have repeatedly refused to answer whether I am "paid Up," or not. I thought I paid a nominal fee when you first began this endeavor. I enter the contests regularly but never know if they are considered because of the anxiety producing unanswered question about my status. In addition, at least 2 times a week you offer us a chance to ask questions as you ignore mine. Please end this abuse!!!

A: Let me tell you why you are abused.   It is your fault.   I remember at least one other time you asked this.  And both times, and maybe more, you failed to identify yourself in your post.  I did not “refuse to answer.”  I could not answer because I had no idea who the hell you were.   I still don’t.   

The good news, however, is that obviously your questions are getting through.   

Q:  My “Jill” takes place at the deli counter at the grocery. As you know, there are two simple details shared with an employee when making a transaction at the counter, product and weight. And if you think about it, the pattern to the request is typically ‘weight first, product second,’ as in “Can I get a pound of turkey?” No matter where I shop, the employee cannot remember the weight that was stated in this brief interaction and either sheepishly asks for clarification or worse, gets it wrong. Especially if you ask for “half pound” of anything. They never forget the product, only the math. It drives me NUTS. A full pound of nuts. 

A: Yes, all the time!  This must identify a universal problem.  Happens to me all the time at the Safeway seafood counter.  “Two pounds of catfish, please?”  “Okay, how much you want?”

Q: Aptonym or Inaptonym in a school discrimination case about natural hair styles and a law meant to prevent the discrimination. . 

A: Excellent. Took me a while to find it, because I was looking for people’s names. 

Q: I do not like to enter a store which has two doors and one, unbeknownst to me, is locked. I always attempt to try the locked door first. Shouldn't that be against the Fire Code?  

A: A poem by Piet Hein: 

  Double doors are justified / Because they’re comfortably wide. / Therefore, you just half undo ‘em / Therefore, nothing can get through ‘em.  

Q: My Jill is people who peel the little labels off produce which damages the surface and allows for accelerated rotting.   

A: I eat the paper.  Or boil it into oblivion, with artichokes. 

Q:  This is probably pretty common, but I need thermostats and volume levels to be set on even numbers.  

A: Why would it be pretty common?????

Q:  My Jill:  Fortune cookies that don’t tell your fortune. The last one I got said, “Everything serves to further.” My point here is—huh?

A: Check.  I also don’t like the ones that have “lucky numbers” on the back.  Confirms that fortune cookies are for idiots. 

Q: My “Jill” is not my hatred of celery, which likely irritates a broad swath of the human population. My “Jill” is the tremendous difficulty in determining whether canned chicken soup, bouillon or bone-broth contains celery. If I must live in a universe with celery I wish celery were an allergen with non-terminal effects so that its presence in foods had to be boldly disclosed.

A: I considered Jills to be the province of obsessive nitpicky types.  But I find myself agreeing with many of them.  So, um… 

Q: Three-way light switches. I want certain ones to be in the “off” position and the partner to be on. I.e. our foyer has a switch by the front door and another at the top of the stairs. I want the one by the front door to reflect the current status of the lights in the stairwell, even if I have to climb the stairs in pitch darkness to go to bed.

A: Okay, this one is nuts.

Q:  Jill for me is if the number of steps, on a staircase I am using, is an odd number. Yes, I count steps. Unfortunately, unlike the toilet paper situation, there is nothing to be done to remedy an uneven number of stair steps.  

A:  Well, this is weird.  But it does seem to be a phenomenon that you experience backwards. According to this architectural story, people prefer odd number of steps, because that way they start and stop on the same foot, which is comforting.     

My house has two interior stairwells, one with 14 steps and one with 13.   I don’t have a preference for footfalls.

Q: Gene, I picked up a copy of your friend Dave Barry's latest novel (Swamp Story) at the library. On page 47, a character in the story is described as wearing a shirt that says "BUT FIRST ROO ROO," which I think is the punch line to a joke I heard years ago in your chat. Funny to see it in print. 

A: This is the greatest dispute Dave and I have ever had. It festered over the years until neither of us truly respects the other. The proper punchline should be “But first, a little roo-roo,”  which nimbly establishes the hypocrisy and duplicity of the savages.   For the last 30 years, Dave has stubbornly refused to accept my judgment. 

“Swamp Story” is excellent, by the way.  

Okay, I just talked to Dave to ask him if he had a response, and it turns out he did.  Here it is: 

“My position is, you’re an idiot.”

—

 Q: When I was little we had a cat named Kitty. Kitty liked to sit on a ledge near the front door. Our next door neighbor was a lady named Marge. If you saw "Pee-wee's Big Adventure” you might remember a character named Large Marge. Neighbor Marge had a similarly imposing format. The preceding was for historical context. Now for my dream: My parents went out for the evening and asked Marge to babysit. At some point I went to the front door to see what Kitty was up to. She was sitting on the ledge as usual. I said hello to her, and Kitty responded by saying hello back! That scared me out of my wits. I turned around and discovered Marge standing behind me. I stammered out something like, “Kitty … Kitty … she … talked.” Marge gave me a wicked half-smile and said, “Meow.” I woke up immediately, frightened and disoriented. I think if dreams had credits, this one would have been directed by Rod Serling.

A ”Meow” was a nice touch.  Eerie.

Q: Worse, to me, than your "present conditional" to describe the alternative that didn't occur in a sports game (or at least as bad, and certainly as prevalent): "may" rather than "might" in such a description. "If he caught that ball, he may have gone all the way." But we know he didn't catch it, so it should be "...he might have gone all the way." "Might," in such a context, means "possibly would have;" "may" means "possibly did." (You could say "...he may have gone all the way" if you didn't know, because your view was obscured by, say, the auxiliary bleachers or a TV crew.) Right? I suspect sportscasters use "may" out of some vague sense that "might" means something less probable than "may." --Rick Beth, Takoma Park

A: Rick, you are a silly, prissy pedant.  Though you are also correct.

This is Gene. I am putting us down now, a couple minutes early, mostly because of my dog’s intestinal urgency.

PLEASE keep sending in observations and questions. I’ll get to them on Thursday.

Also, there will never be a farewell to alms: