I read Asimov's Foundation again. It hits

I just finished rereading Foundation for the third time. First time I was sixteen. Second time in college. Third time, this week, sitting on my couch at midnight with a cup of coffee that went cold two hours ago.

It’s a different book every time.

At sixteen, Foundation was an adventure story. Hari Seldon was this impossibly cool mathematician who figured out how to predict the future using something called psychohistory, and then he set up a secret Foundation on the edge of the galaxy to preserve human knowledge through the coming dark age. It was about heroes and plans and outsmarting empires. I loved it the way a teenager loves anything with a secret plan.

In college, I read it as a political parable. The Foundation wins not through military strength but through economic and cultural influence. Each crisis resolves through cleverness, not force. I wrote a paper about it for a political science class. Got a B+. The professor said my thesis was “interesting but underdeveloped.” Fair.

This time, though. This time it scared me.

Psychohistory in 2018

Here’s the premise of psychohistory, Asimov’s fictional science: if you have a large enough population and enough data about their behavior, you can predict the future. Not individual actions (that’s impossible, even in the book), but the broad strokes. The rise and fall of civilizations. The flow of power. The shape of centuries.

Asimov wrote this in 1951. It was pure science fiction. Nobody had anything like the data or computing power to attempt something like psychohistory.

I’m reading it in 2018, and I can’t stop thinking about Google.

Not Google specifically. Google is just the most visible example of something bigger. We’re generating more data about human behavior than any civilization in history. Every click, every search, every purchase, every movement tracked by the phone in your pocket. Somewhere, there are models that predict what you’ll buy tomorrow, who you’ll vote for, whether you’ll quit your job.

These models aren’t psychohistory. They’re crude. They’re often wrong. They predict individual behavior, which Asimov explicitly said was impossible. But the trajectory… the trajectory is clear. More data. Better models. Longer time horizons.

What happens when the models get good enough to predict not just what you’ll buy, but what a million people will do? Ten million? A country? A civilization?

The part that scares me

In Foundation, Seldon is the hero. He uses psychohistory to save humanity. He shortens a 30,000-year dark age to a single millennium. He’s benevolent.

But Asimov was smart enough to plant the question: what if someone uses it for something else?

The Second Foundation exists in secret. They manipulate events, steer civilizations, adjust the Plan when it drifts. They do it with good intentions (mostly). But they do it without consent. Nobody voted for the Second Foundation. Nobody even knows they exist.

I keep thinking about the companies that model human behavior at scale. They don’t have Seldon’s math (not yet, maybe not ever). But they have something Seldon didn’t: real-time feedback loops. They can predict what you’ll do, show you something designed to make you do it, and then measure whether you did it, all in milliseconds. That’s not psychohistory. It’s something else. Something Asimov didn’t imagine because even he wasn’t that pessimistic.

Or maybe he was. The Second Foundation manipulates everyone for their own good. That’s exactly what the recommendation algorithms claim to do. “We’re showing you things you’ll like.” Sure. But who decides what I should like?

The Mule

There’s a character in the second book called the Mule. He’s a mutant who can manipulate emotions. He breaks the Seldon Plan because he’s an individual who can’t be predicted by statistical models. He’s a black swan. An outlier so extreme that all the math in the galaxy couldn’t account for him.

I think about the Mule when I think about social media. The models are getting good at predicting aggregate behavior. They can tell you which candidate will win an election (usually). They can tell you which products will sell. They can model population-level trends with increasing accuracy.

But then a single person, a whistleblower, an activist, a demagogue, someone with a phone and a message that resonates, can upend everything. One tweet goes viral and a company’s stock drops 20%. One video changes a national conversation overnight. One leak reshapes an election.

The models can predict the masses, but they can’t predict the Mule. And I wonder if that tension, between the predictability of crowds and the chaos of individuals, is something Asimov understood better than anyone.

Cambridge Analytica

I should say this out loud. I’ve been dancing around it.

This year, 2018, we learned that Cambridge Analytica harvested data from millions of Facebook users to build psychological profiles and target political advertising. Whether it worked (the debate is ongoing) almost doesn’t matter. What matters is that someone tried. Someone looked at the data we generate and thought: I can use this to predict and influence how people vote.

That’s psychohistory. Not the elegant mathematical kind that Seldon practiced. The crude, commercial, ethically bankrupt kind. But the impulse is the same: model human behavior, predict outcomes, steer the results.

Asimov imagined this taking place across a galaxy over thousands of years. We’re doing it across a social network in an election cycle.

Why I keep rereading it

I think Foundation is one of those books that changes as you change. At sixteen, I didn’t have enough context to see the warning in it. In college, I had enough political awareness to see the power dynamics but not enough technical knowledge to see the data angle. Now, in 2018, working in a world saturated with predictive models and behavioral data, the book reads like a cautionary tale wearing an adventure story’s clothes.

I’m probably reading too much into it. Asimov was a storyteller, not a prophet. He was trying to write a good yarn about the fall of a galactic empire, not predict the rise of surveillance capitalism.

But that’s what good science fiction does, isn’t it? It asks a question so clearly that every generation finds a new answer in it.

The question Foundation asks is this: if the future could be predicted, who should get to know?

In 1951, that was a fun thought experiment.

In 2018, I’m not sure it’s theoretical anymore.

My coffee is cold again. I think I’m going to read the second book.