Neil deGrasse Tyson, a Last Meal, and the Other 15%

Neil deGrasse Tyson is an astrophysicist who directs the Hayden Planetarium in New York and has spent a lifetime making the universe make sense to the rest of us. He may be the most recognizable scientist in America. On Last Meals, the Mythical Kitchen show where host Josh Scherer asks famous guests to talk through the meal they'd want at the end, he was asked whether we should fear AI or hope in it. His answer was measured. It is a very helpful tool, it needs guardrails in places, and his own field, astrophysics, could not function without it. Then he said something worth pausing on. When he Googles something now, it goes to AI, and the answer is right about 85 percent of the time and wrong the other 15. "I only know that," he says, "because I know what it is I'm looking up." If you don't, he says, "it's just the answer."

Sit with that for a moment. Some of the most accomplished minds alive say they cannot do their work without this tool, the one we keep fifteen-year-olds away from in school. The restrictions are not foolish. They exist to hold back real things: cheating, plagiarism, content a child should not see, the slow pull toward emotional dependence. Those harms are genuine. But every one of them is something we already know how to raise a young person through: with a parent, a teacher, a mentor, a real role model close enough to shape the habit. You do not protect a child from a tool they will use for the rest of their working life by keeping them from ever learning to use it well. You protect them by building the judgment to handle it, in the years and supervised settings where that is still possible.

The judgment is the crucial point, because the harms and the mastery come down to the same place. Not the machine and not who is allowed to touch it, but the young mind in the glow of the screen. That is the whole problem, stated by one of the most trusted scientists alive. AI does not lie to the expert. It lies, invisibly, to the person who cannot catch it. The protection was never in the machine. It was in what Tyson brought to it: a lifetime of knowing enough to feel a wrong answer when it arrives wearing the face of a right one.

Watch the rest of the episode and you see what that protection is actually made of. It is not the knowledge. It is the posture he brings to the knowledge.

Asked his greatest fear, he does not reach for death. He fears the human mind may be too small to ever understand the universe. He quotes Isaac Newton's confession, that each of us is only "a boy on a seashore picking up one shell and one rock because it's shinier than the next, while the great ocean of undiscovered truths lay before me." Asked for his last words to the camera, he chooses, "I remain insatiably curious, in life and in death." A man who has spent a lifetime drinking in knowledge wants to die thirsty.

These are not charming personality notes. They are the exact faculties that let him catch the 15 percent. Curiosity is why he looks again. Humility is why he assumes he might be wrong. Skepticism is the reflex to confirm before he trusts. Take those three away and you do not get a lesser Tyson. You get someone who takes the answer and moves on.

There is a moment late in the meal that makes the point better than any argument. Tyson tosses off a fact about maltose, and the show's team checks it on the spot, then reports back that he was right. He is not offended that they doubted him. He is charmed by it. "It's been a while," he says, "since someone says, I don't trust what you said, let me look it up." The most trusted scientist alive treats being verified as a courtesy, not an insult. That is the whole posture, lived in a few seconds: not take my word for it, but check me.

And the check did not come from Tyson. It came from the host and his team, who took a confident claim from one of the most trusted scientists alive and checked it. There was nothing confrontational about it. Tyson was right, and they confirmed it. That is the point: they did not accept a confident answer on faith, even from an expert, even when it turned out correct. A student who will not do the same with a confident AI answer, on the day it is wrong, is the whole problem.

That is the largest difference in how a person uses this tool. The people who are best with AI do not use it to get better answers. They use it to ask better questions, ones they would never have thought to ask. They bring the curiosity, the humility, and the skepticism to the machine, because the machine brings none of its own. Tyson is not safe using AI because he is brilliant. He is safe because he never stopped wanting to check.

There is a harder turn in the same conversation. Tyson cuts us down to size: a one-centimeter slice of your lower colon, he notes, holds more microbes than every human who has ever been born, "yet we tell ourselves that we are in charge." He meant this as cosmic humility, "you are not as important as you want to believe you are," a warning about our place in the universe. I read it as a warning about our kids. Because the line hands us something he did not stop on. If you control almost nothing, not the cosmos, not even your own body, then the one sovereignty actually available to you is your own mind. Your judgment is the rare thing that is genuinely yours to keep or surrender. And it is exactly what a young person gives away when they take the answer without thinking.

Here is what should worry us. The answer-machine wears down those very faculties, and it does it pleasantly. Why stay curious when the answer arrives in a second. Why sit in humility when the screen makes knowing feel finished. Why verify when the output is smooth and sure. Curiosity, humility, and scrutiny each cost a little friction, and removing friction is the entire promise of the tool. It does not argue you out of those habits. It just makes them feel optional.

Tyson is safe because he built his judgment first, across decades, before any of this existed. He brought the finished mind to the tool. A child is in the opposite position. They are building judgment and using the tool at the same time, and the tool quietly erodes the very faculty they are trying to grow. They cannot catch the 15 percent, because they do not yet know enough to feel it. So they take the answer. And the answer, taken enough times, becomes a default: the reflex to reach for the result instead of the reasoning. Defaults set early, in the years no one is watching, and they do not politely correct themselves later. A formed habit is not a law. You cannot amend it next session.

That is why this cannot wait, and why it cannot be left to the kid. Curiosity, humility, the instinct to check, these are built by practice, with an adult close enough to model them, at the age when habits are still soft. By the time a young person could choose this for themselves, the default is already set. The window is now, and it belongs to the adults nearest the child. Parents first, then teachers and mentors. No bill or statute or school rule reaches inside a kid to build a habit. No platform will do it either, because the platform profits from the friction being gone.

Tyson wants to die thirsty, still reaching for the next shell on the seashore, still "insatiably curious, in life and in death." He earned that over a lifetime of refusing to take the answer and stop. The quieter, more urgent question is whether the rest of us will raise a generation that can still want the same thing. Whether, while the machine makes thinking optional, the adults in the room will do the one thing the machine never will: build, in a child, the judgment to stay in charge of their own mind. That was never the law's job. It is ours. And the time is now, before the default is set for them.