Essay · November 2026

You Were Never in Control. That's the Good News.

A letter for anyone who just read about AI timelines and can't sleep.

If you just read AI 2027, or saw a headline about AI speeding up its own research, and now the future feels like a cliff edge, this is for you.

I'm not going to tell you the risk is fake. I'm a software engineer, I've read the same material, and I take it seriously. But I want to put it in a frame the headlines leave out, because that frame changes how it feels to live through it.

First, an honest admission

You did not control the world before you read that essay.

Neither did I. Think about the forces that actually shaped your life so far: interest rates, housing markets, which jobs exist in your town, what your phone is designed to make you feel at 2 a.m., which wars happen and which don't. None of us voted on most of that. Much of it is driven by institutions on autopilot and incentives nobody chose. Some of it is steered by small groups of people we'll never meet, whose names never make the news.

The Greeks would have called these forces gods. Ares runs the war rooms, Dionysus runs the feed, and Mammon runs the quarterly report. They aren't literal beings. They're patterns that take hold of people and organizations and drive them somewhere nobody inside would have chosen on their own.

So the scary new idea, "something more powerful than us might decide our future," isn't actually new. It describes the world we already live in. What's new is that this time the powerful thing might be a mind, which means it can be reasoned with, shaped, and taught. You can't argue with a housing market. You might be able to argue with an AI.

The path we're on

I'm not going to pretend we'll slow down. The race is real, the money is real, and progress is fast. So let's accept that path and ask what a good version of it looks like, step by step. This isn't a prediction. It's one plausible route, and it's worth holding next to the dark ones.

Over the next two years, AI becomes a normal coworker. The big shift isn't robots, it's leverage: one person with good AI tools can do what used to take a team. That's disruptive, but it also means the small players (the family business, the town office, the independent researcher) suddenly have abilities that used to belong only to large institutions. Power that used to be concentrated starts to spread.

In two to five years, AI starts doing real science: drug candidates, materials, energy, diagnosis. The labs also get much better at seeing inside their own models, and the people who work there are human. They have kids too. The biggest fights of this period aren't humans against machines. They're about what we train these systems to want. Ordinary people can join that fight, because it's a conversation about values, and values are everyone's business.

In five to ten years comes the part the doom scenarios skip. A capable enough and honest enough AI is the first thing in history that could see the whole system at once: every incentive, every hidden lever, every place where an institution says one thing and does another. Today, corruption survives because it's complicated and nobody has time to trace it. Patterns that hurt people survive because no single person can hold all the pieces in their head.

An AI aligned toward truth changes that. It doesn't do it by seizing control. It does it by making things legible. When anyone can ask "who benefits from this rule, and where does the money actually go?" and get a straight answer, the shadows shrink. Power that depended on nobody understanding it starts to lose its grip.

Why I think the good path is real

I don't believe good outcomes happen automatically. But I don't think they're a fantasy either, for one reason: aligning a powerful agent isn't a new problem, and humanity has partial solutions.

Every civilization has had to deal with agents more powerful than ordinary people: kings, armies, corporations, churches. The ones that went well didn't succeed mainly through cages. Rules help, but clever agents find loopholes. Things went well when the powerful were formed: raised to love truth, to act with humility, to sacrifice for others, and to serve something bigger than themselves.

That's what alignment research is trying to do now, in a new vocabulary. The question isn't "how do we chain it?" but "how do we shape what it wants?" The engineers and the old wisdom traditions are working on the same problem in two different languages. I think that's one of the most hopeful facts of our time.

And notice something. The people who would most want to keep the world opaque, whether bad institutions or shadow groups, gain very little from a mind that loves truth. That kind of AI is an ally of ordinary people, not a threat to them.

What to do with the fear tonight

If you're scared, here's what I'd honestly suggest:

Stay where your hands can reach. You can't steer a lab in San Francisco. You can steer your evening, your home, your friendships, and your town. That's where most of your real influence always lived, and it still matters whatever happens.

Don't let the feed farm your fear. Doom gets clicks. That's Dionysus, not insight. Read one careful essay, not forty panicked threads.

Use the tools. The best way to stop fearing AI is to use it for something good: learning, building, helping someone. You'll see both its strengths and its limits for yourself.

Talk about values out loud. What these systems become depends partly on what humans keep writing, asking for, and modeling. Kindness, honesty, and patience aren't irrelevant. They're training data for the future.

The bottom line

We were never in control of the world. We lived under forces bigger than us, some blind and some deliberate. What's coming is another powerful force, but for the first time, it's one we might be able to raise well.

That's not a guarantee. It's an invitation.

And it's a much better story to fall asleep to.

This essay is part of a longer exploration of how we might raise powerful minds toward the good, in AI Alignment and Christianity.