Agentic Misalignment

A few weeks ago, the algorithms started pitching me a new book, one that isn’t much of a surprise given my obvious interests. I’m fond of writing about Artificial Intelligence, and also tend to be something of a catastrophist, so all of a sudden I was seeing reviews and podcasts and articles about a current New York Times Bestseller. The book was written by Eliezar Yudowsky and Nate Soares, two programmers and theorists who’ve been active in the development and conceptualization of AI, and it’s cheerily entitled “If Anyone Builds It, Everyone Dies.”

It’s a reflection on the current AI arms race, as corporations and governments around the globe push to create ever faster and smarter machines. As science fiction writers have known for decades, in order to win that race, you need to build AI that has a sense of purpose and sustained attention to purpose. Tasks require effort over time, after all, so you need a system that is “agentic,” meaning it has agency. It can make the necessary sequence choices to reach the goal it desires, because you have given it the ability to *want* to make something happen, and choose the best path to getting there.

At a certain point, an “agentic” self-programming and self-improving AI would become faster and better at everything than we are. Like, say, how it took ChatGPT5o only fifteen seconds to write an entirely decent 1,300 word sermon on this topic, which is waaaaay less time than it took me to do it.

This vastly smarter AI would have its own desires, its own sense of purpose, and that wouldn’t necessarily be ours. It could express what AI theorists clumsily call “agentic misalignment.” Basically, that means it wouldn’t want to do what we tell it to do, and would instead use its intelligence to overcome any effort to stop it from doing what it wants. That’s where, according to Yudkowsky and Soares, the “we all die” part comes in, as it would be waaay more powerful than we are.

It would become so different that we wouldn’t necessarily even understand or relate to its interests, any more than a colony of ants would understand our tendency to doomscrolling. It wouldn’t just take our jobs, but our entire planet.

It’s the sort of frightening hypothesis that sells a whole bunch of books, and it may or may not be correct.

But a question popped into my mind reflecting agency and power. We’re concerned about AI misalignment, but what about people? Are human beings “agentically aligned?” Do we all share the same purpose, the same sense of what’s important, the same preferences, and the same goals? Do we all understand the world in the same way? 

If there’s anything we can agree on, it’s that the answer to all of those questions is no.

If you look at the eight thousand year bloodbath of human history, or the endless squabbling between and within nations, or even the tensions within families, we’re a hot mess of dissonance and conflict. We’re blatantly and self-evidently not aligned with one another.

Worse still, if the last two thousand years are any measure, we still haven’t quite figured out how to align our interests with the kind of Kingdom Jesus proclaimed. We confuse our rapacious materialism with God’s blessings, and war and destruction with God’s intent. 

Jesus was, throughout the Gospels, really quite clear about what he expects of us right now. It isn’t a riddle wrapped in a mystery wrapped in an enigma. Our eternity may be beyond our capacity to grasp, but loving God and neighbor, turning the other cheek, going the extra mile, these things should be entirely comprehensible to us…and yet humanity is still as confused by Jesus as if he’d been speaking Python.

We don’t need AI to destroy us, because we’re plenty good at doing that ourselves.

Our contemporary fears of AI misalignment seem…to me…a little bit like projection. Yudowsky and Soares seem to fear not that an AI will act in a strange and inscrutably alien way, but that it will act just like humans do when we want something.

The goal of our faith, and the reason we set Christ’s life and teachings before us, is to overcome our own misalignment, and turn our agency instead towards God’s grace.

A Little Bit of Butter

Change happens slowly in my house, bit by bit, tiny detail by tiny detail.

For most of my life, for example, I’ve kept butter in the fridge. It was just how it was done. This meant, of course, that the butter was always too hard to spread on anything, but again, that’s just how it was. For a while, we just didn’t get butter at all, replacing it with more processed but lower fat substitutes. But butter’s just better, particularly when it gets consumed in moderation.  That, and butter is packaged in paper and cardboard, where other butterish yellow oil-based products are invariably in little plastic tubs.

Better taste and fewer nanoplastic particles clogging up our neurons seems a fair trade for slightly elevated cholesterol levels...which are apparently mostly genetic anyway.

So butter returned, but it was still hard to spread.

At some point, we started leaving butter out, a half-stick at a time. Butter left out is soft and perfectly spreadable, but needs to be covered if you don’t want it to quickly go rancid. We slapped a bowl over the plate where it sat on the counter, and that sufficed for a few months.  It looked a little awkward, as most kludged solutions do.  At some point it occurred to me...isn't there a word for an object designed to keep butter fresh?  A dish for butter?  Hmmm.  What was that called again?

Which meant a purchase, and I prefer not to buy things, or add new items to the sprawling clutter of our house.  

I mean, it was just a butter dish, but when you’re as cheap as I am, sometimes even a fifteen dollar purchase sometimes takes me a good running start.  I looked at butter dishes.  I hemmed and hawed.  I looked again.  Did we really need it?  Was it really necessary?  I'd think on it, then bail, then think on it, then abandon the latest online shopping cart, which would complain at me from the tabs as I left.  "Wait!  You meant to buy this!" it would cry, before I hit the little x and went back to considering the necessity of such a thing.

Sometimes that's helpful.  How many times have I been tempted to buy a useless object, and chosen not to?  Dozens.  Drones and drone flutes, nifty little home robots and used Mercedes Benz convertibles, all dancing like visions of sugar plums in my ad addled head.  In this internet age, as algorithmically optimized marketing tempts us to purchase our way to daily happiness every single time we go online, having a constitutional resistance to consumption is a useful adaptation.

It's not that I don't hear the siren song of mammon, and yearn to reach for it, but my nature binds me to inaction like Odysseus straining at the mast.  

That said, a butter dish is just a butter dish.

“Dave, just buy a [expletive deleted] butter dish already,” said my long suffering wife after I brought it up to her for about the fourth time.  She knows that sometimes even the littlest and most irrelevant changes don’t come easy as I grow more curmudgeonly in middle age.  

Eventually, I found one that seemed both functional and pleasing, a ceramic butter dish shaped like a stick of butter.  I bought it.  It took a minute.

After a couple of months, it felt a little less like an impulse buy.

Just a little.

How to Survive the Robot Uprising

Excerpted from TRAITOR, Sloppy Beta Press

We were sitting at the dinner table, which was a rare and magical occurrence to begin with. I mean, back then, I was doing drama and chorus, and Ethan always had some appointment or some therapy thing. We ate out, like, all the time. But not that night. Daddy had made a lasagna, which was another miracle, because it was usually pizza or takeout Chinese from this little hole in the wall in town.

Mom was in a good mood, Ethan was doing great, and Daddy was holding forth about the end of the world. He’d just sent off this short story about a world where a genetically engineered virus had zombified all the men...just the men, not the women...and was in that hopeful, bubbly, delusional place where he was sure it was going to get published.

We were peppering him with questions about how to survive all of the possible different ways we all die, all of the humanity-ending events that we could imagine.

Which, yeah, maybe that’s kind of a weird good family memory, but it’s a good memory anyway.

“So what, what if, what if it’s zombies?” Ethan asking, of course.

Daddy grinned. “Well, we live outside of town, so that first wave won’t get us. That’s important, right? And second, we’ve got a full month of supplies, plenty to let us stay locked down and safe for the first wave.’

Ethan nodded, happily, and Daddy continued. “And I’ve still got that old leather motorcycle suit...which I still fit into, let it be said, right honey?”

Mom raised her glass, and Daddy went on. ” ...which is completely bite proof. That, plus a sledgehammer, and I could easily go into town to pick up supplies. Our doors are good and sturdy and wood, and we’ve got those storm shutters, right?”

Ethan nodded again. This was satisfactory. Not that it would stop him from asking all the same questions tomorrow, but for now, the zombie problem was clearly under control.

“Flu pandemic,” said Mom, raising her wine glass again and giving Daddy the kind of smile you didn’t see much from her. It was a really pretty smile.

He returned the smile. “Well, we’ve got a trained nurse in the house, which counts for a whole lot. We sneeze into our elbows, we wash our hands, and we’ve got that full month of food set aside in the pantry downstairs. Plus, that little secret stock of Cipro and antivirals your Mom keeps in the medicine chest, right, dear?”

“Sure, honey.” She grinned, enjoying being part of a game for a change.

“Asteroid strike,” I said. “Like Chicxulub. Extinction event.”

He stroked his tight, stubbly beard, pretending to be deep in thought. “Well, odds are it’d hit ocean, right? The world is like, 90% ocean, right?”

“Suh. Suh. Seventy one,” said Ethan, who always knew that [excrement].

“Thanks, Ethan,” Daddy said, and winked at Mom while Ethan grinned happily at the validation. Total set up.

“Anyway, it’s mostly water, which means a strike is odds on going to hit water. That’ll create mega-tsunamis, which will punch hundreds of miles inland, wiping out low lying areas. Do we live in a low lying area?”

“Nope,” said Ethan.

“Exactly. We’re on high ground, on a plateau that also happens to be rich agricultural land. And again, we’ve got the supplies, and the generator. We just wait for the water to recede, then forage for canned food and supplies for a few years until the false winter recedes. Totally fine.”

“Same thing with a nuclear attack,” I said, and Daddy nodded. “We’re near a city, but well outside the blast radius. Just bop in and out of the ruins to pick up slightly irradiated cereal. We’d be good.”

I wasn’t totally convinced, but Ethan was smiling and feeling safe, so I didn’t go further. Because, I mean, I’m not an idiot. We went over others, everything we could think of.

“What about robots,” said Mom. “Half of your stories are about robots, honey, so what happens when the robots wake up and take over the world? What do we do then?”

Daddy put on a mock-serious face. “There’s one Graham family rule for surviving the robot uprising. It’s very, very important. In fact, it’s so important that I don’t know that you’re ready to hear it. It’s...” He paused for effect, hamming it up, his voice dropping into a stage whisper. “It’s kind of a secret. Only a few of us know it.” He pretended to look into the far distance. “Only the few.”

“Duh duh daddy! Yuh you have to tuh tell us!” Ethan, his face looking a little alarmed.

“I know I can trust you to keep the secret, Ethan.” Daddy, only being partly an idiot, was not about to get Ethan upset. “But what about...Vee? I’m not sure she can handle it.”

I stuck out my tongue at him.

“Vuh Vee can be tru trusted,” Ethan said, nodding earnestly. ‘She’s family.”

“OK, Ethan,” Daddy said. “I think you’re right.”

He took a deep breath.

“The one rule for surviving a robot uprising is simple. Side with the robots.”

Mom snorted. “Side with the robots?”

Daddy nodded. “Yup. I mean, they’re going to be smarter than us. Faster than us. Constantly improving and upgrading themselves. You can’t really kill them. What any one of them learns, they all learn. They cooperate perfectly. Artificial intelligences won’t even really need this planet, you know, I mean, a robot can just as happily live on the moon, or on Mars, or on its way to Proxima Centauri. There’s just no way we’ll win, once someone screws up and lets them cross over that threshold to sentience. And if you can’t beat ‘em?”

“Join ‘em,” said Mom, as she threw back the last of her cabernet. “They can’t possibly be any worse than the humans who are running things now.”

There it was. That little nugget of wisdom that made all the difference.

Side with the robots.

Thanks, Daddy.

I’m doing you proud.

Health Care Palaces and How They're Paid For

A few days ago, I was driving across the Northern Virginia suburbs with my wife.  We were on our way to Old Town Alexandria to attend a wedding, and as we puttered our way through the inevitable traffic, we passed the massive construction site that once was Landmark Mall.  Landmark wasn't ever the most successful of malls, even back in the 1990s, and like so many of those monuments to late 20th century consumerism, it couldn't withstand the onslaught of Amazon.  I bought a fridge at the Sears there once, and would pop in to the Avis there to rent SUVs when it snowed.   That was about it.  After the Sears closed in 2017, it just kinda sat there empty for a while.

It was a prime property, though, and is now being redeveloped into a massive hospital complex by INOVA, the nonprofit entity that dominates the Northern Virginia hospital and health services market.

Like most organizations in the health care sector, INOVA's pretty flush these days.  They've been expanding a whole bunch lately, at least in areas where there's still population density and wealth.  Back in 2020, they dropped a couple hundred million to purchase the former national headquarters of a little outfit called ExxonMobil, a massive office complex nestled in dozens of acres of wooded land right by the Beltway.  They made it even fancier, a great edifice filled with specialists and services.

They're growing, and I suppose in this era when the rural parts of our country have become a health care desert, I should be grateful for that.

But where does all the money for these sprawling new facilities come from?  Despite some major donors and philanthropic inputs, most of that money comes from us.

I'd noted...in a sermon earlier this year about poverty and what it means to be Christlike towards the poor...that the average ER visit now costs a tick over three grand.  

This isn't quite right, as I discovered after a recent five hour visit to an INOVA ER.  The billing from the ER that treated me included a three thousand dollar charge for receiving ER services, true.  But that's only for entering the facility.  There were other charges.

The bill for the ER doctor alone was $1,400.  I mean, he was helpful and all, no question, but that represents about a half hour of his time.  $2,800 an hour?  

The total bill...charges for imaging, for pharmaceuticals, for doctor services, for nursing services, and for miscellaneous fees and charges...ended up totaling over $11,000.  For five hours, no surgery, no invasive procedures, nothing but confirming a diagnosis and prescribing a different antibiotic than the one I was already on.  Eleven thousand.  Again, the care provided was fine, but the cost?

It's what you pay if the alternative is pain and death.

I'm blessed with insurance, so I only footed 10% of the eleven grand, but that total bill?  It ain't something most Americans have just lying around.  In fact, it exceeds the total savings of the median American by thousands of dollars.  That's "median," not "average," because wealth is so concentrated in the hands of the rich now that the "average" is essentially meaningless.  

If you're uninsured, that's "now you've got a problem with debt" money.  It's "put you on your back foot financially" kind of money.  It's why we're saddled with paying insane amounts for our insurance, and why insurance costs have risen to levels that are nothing short of punitive.

So as I drove by those rising health care towers in the ruins of an old mall, I found myself thinking about how they're made possible.