Sci-Fi Shapes the Future

I read a lot of science-fiction growing up.

And in a strange way, these futuristic novels prepared me for today’s world.

I devoured books by Robert Heinlein, Frank Herbert, Arthur C. Clarke, Dan Simmons, Neal Stephenson, Michael Crichton, and others.

So when the internet started taking off in the early 1990s, I already had a good idea about just how big (and crazy) it would become.

By that time I’d already read a dozen books about cyberspace, a term coined by William Gibson in his 1984 novel Neuromancer.

In Snow Crash, released in 1992, Neal Stephenson predicted social networks, internet addiction, virtual reality, the gig economy, hacking wars, and much more. The book was a huge hit in Silicon Valley, and inspired countless tech entrepreneurs.

Science-fiction has a long track record of not just predicting, but shaping the future.

The inventor of the modern submarine, Simon Lake, was directly inspired by Jules Verne’s 1870 book 20,000 Leagues Under the Sea. He incorporated ideas from the book, including ballast tanks and a periscope.

Igor Sikorsky, inventor of the modern helicopter, was inspired by Jules Verne’s 1886 novel Clipper of the Clouds, which featured a multi-rotor flying machine.

In Fahrenheit 451, Ray Bradbury described tiny “thimble radios” which fit inside a person’s ears. Sound familiar?

And when Motorola engineer Martin Cooper created the first mobile phone, he credited the communicator from Star Trek for inspiring him.

Sci-Fi on AI

Sci-fi has a strange ability to shape the future. That much is clear.

So what does fiction have to say about AI? It’s a classic theme of the genre.

From 2001: A Space Odyssey to Blade Runner, many sci-fi stories involve artificial intelligence.

It’s a dark force in many fictional worlds. In one of my favorite books, Dune, released by Frank Herbert in 1965, any form of machine thinking is prohibited. In Herbert’s universe, AI nearly destroyed humanity.

My favorite line from the book is:

“Once men turned their thinking over to machines in the hope that this would set them free. But that only permitted other men with machines to enslave them.”

–Frank Herbert, Dune

As a result, in the Dune universe, one commandment stands above all others: “Thou shalt not make a machine in the likeness of a human mind.”

It’s crazy to think Herbert wrote that in 1965. This aspect of the book has been on my mind lately. It’s a deep one. And relevant to the present.

Recent advances in agentic AI are straight out of a novel. The progress lately has been disturbing. It appears that we’ve entered an accelerating self-improvement cycle, where each new AI model is better than the last, and it is used to build the next. Bigger, faster, better, more disruptive. Rinse, repeat.

Agents can already build complex software with a single command. And hackers are having a field day with all the new goodies they have in their toolkits.

And it’s not just coding. Soon, agents will dominate all knowledge-work. Legal, medicine, accounting, writing, marketing, mathematics, physics, and just about everything else.

I can see the dangers Dune warned about clearly.

Too Much Power

The companies that win this sector will have unprecedented power. Because in the near future, AI will be used to do… everything.

And governments are already attempting to shape AI to their will. Chinese models won’t talk about Tiananmen Square, for example. And I’ve had American models get squirrely when I try to explore certain subjects, like the Nord Stream pipeline explosions. Every AI model has its own unique biases. Some are intentional, while others are accidental.

Already, special interest groups are publishing vast quantities of content designed to be vacuumed up and incorporated into future AI models.

These are small things for now, but soon they won’t be. The race to control the future of tech is on, and there are only two credible players. The U.S. and China.

We’re going to see some wild developments over the next few years. Chinese AI models, which undercut U.S. prices by ~90%, will likely be banned at some point. China has already banned our leading models.

Agentic AI models have also proven to be extremely capable hackers. This means that in the future, you’ll probably need a digital ID to access top-tier agents. More surveillance incoming.

All sorts of things, both beautiful and terrible, are coming soon.

It’s a sci-fi world, we’re just living in it.