We Are the Adolescents, Not Our Technology

In Carl Sagan’s novel Contact, the astronomer Ellie Arroway finally meets the alien civilization she has spent her life seeking. When given the chance to ask anything, she chooses this: “How did you do it? How did you survive this technological adolescence without destroying yourself?”

It’s a question worth sitting with. Sagan imagined that any civilization capable of reaching across the stars must have once stood where we stand now. They too must have invented weapons before wisdom, discovered power before developing the institutions to contain it. And somehow, they made it through.

Dario Amodei, CEO of Anthropic, borrows Sagan’s framing in his recent 20,000-word essay, “The Adolescence of Technology.” In it, he warns that humanity faces a brief, dangerous window in which artificial intelligence could either transform our civilization or destroy it. The essay catalogs five categories of existential risk, from AI systems that might pursue misaligned goals to bad actors using AI for mass destruction. It is sober, detailed, and at times genuinely frightening.

Most commentary on the essay has focused on the technology. Will AI really become a “country of geniuses in a datacenter”? Are Amodei’s timelines credible? Is this warning sincere, or is it, as one critic put it, “market capture dressed as moral urgency”?

These are reasonable questions. But I want to suggest a different reading. What if the essay’s deepest insight isn’t about technology at all?

What if we are the adolescents?

---

The Maturity Gap

Developmental psychologists have long understood that adolescents are strange creatures. They can think like adults but don’t yet feel like them. Research by Laurence Steinberg and colleagues across eleven countries found that cognitive capacity reaches adult levels around age 16. But psychosocial maturity, meaning the capacity for self-restraint, especially in emotionally charged situations, develops years later.

Steinberg has a vivid way of describing this: it’s like engaging a powerful engine before the braking system is in place. The teenage driver has reflexes and intelligence. What they lack is judgment. They can calculate speed but not consequences. They understand risk in the abstract but feel invincible in the particular.

Now consider humanity’s situation with AI. We have, in a remarkably short time, built systems of extraordinary cognitive power. Claude, GPT-4, and their successors can write, reason, code, and analyze at levels that would have seemed like science fiction a decade ago. Our engine is roaring.

But where are the brakes? Our institutions, our regulatory frameworks, our ethical norms, our collective wisdom, these develop on a different timescale entirely. Legislatures move in years while capabilities move in months. International coordination lags while companies race ahead. We debate the proper oversight of AI systems that are already obsolete by the time consensus forms.

This isn’t a flaw in the technology. It’s a flaw in us.

We have always struggled to match our power to our wisdom. Every generation has faced some version of this gap. But previous technologies, from fire to the printing press to nuclear weapons, advanced slowly enough that our institutions could, however imperfectly, adapt. What makes AI different is not just its power but its pace. We are being handed godlike tools on an adolescent timeline.

---

Philosophers Who Saw This Coming

Amodei is not the first to notice this asymmetry between power and wisdom. The nuclear age forced an earlier generation of thinkers to confront it directly.

Günther Anders, the German philosopher, spent his career grappling with what he called “Promethean shame”: the recognition that our creations had outgrown our capacity to comprehend them. “The more our technological power grows,” he wrote, “the smaller we become.” Anders wasn’t anti-technology. He was worried about the growing distance between what we could make and what we could understand. The atomic bomb was, for him, the ultimate symbol of this gap: a power so vast that no individual mind could grasp its full implications.

Hans Jonas, another philosopher shaped by the nuclear age, used the phrase “Prometheus unbound” to describe humanity’s situation. We had freed the Titan who brought fire to humans, and now we wielded godlike power without godlike wisdom. For Jonas, this demanded an entirely new ethics, one focused not on intentions but on consequences we could barely foresee.

The nuclear age was our first adolescent crisis. We survived it, but barely, and largely through luck. The Cuban Missile Crisis came down to individual judgment calls in moments of chaos. We built safeguards afterward, but the fundamental gap remained. We never became wise enough for our weapons. We just managed not to use them.

AI is the second crisis. The question is: what did we learn from the first?

---

Reading Amodei Generously

Amodei occupies an unusual position. As CEO of Anthropic, he runs one of the companies building the very technology he warns about. This gives his critics ammunition. Fortune notes that his predictions haven’t always been accurate. Others point to what they see as a conflict of interest: Anthropic benefits from both the excitement and the fear around AI.

These criticisms deserve consideration. But the steel-man version of Amodei’s position is also worth taking seriously. His proximity to frontier AI development gives him something most commentators lack: firsthand knowledge of what these systems can do and how quickly they’re advancing. When he says he feels “the clock ticking down,” he’s not speaking from abstract concern but from watching capability curves climb.

What strikes me most in the essay, though, isn’t the risk catalog. It’s the recurring emphasis on character over rules. Amodei’s approach to AI safety, what he calls “Constitutional AI,” doesn’t try to anticipate every possible failure mode and write a rule against it. That would be impossible. Instead, it tries to instill values and identity at a deeper level.

“Claude’s constitution,” he writes, “presents a vision for what a robustly good Claude is like; the rest of our training process aims to reinforce the message that Claude lives up to this vision.”

If you’ve ever raised a teenager, this might sound familiar. You can’t give rules for every situation. There are too many. Instead, you try to cultivate character: honesty, kindness, good judgment, the capacity to reason through novel situations. The rules are scaffolding. The character is what remains when the scaffolding comes down.

This is not a complete solution, and Amodei doesn’t claim it is. But it points toward something important. The adolescent problem isn’t solved by better surveillance or stricter constraints. It’s solved by growing up.

---

The Personal Question

Most commentary on Amodei’s essay is addressed to policymakers, researchers, and institutions. What should Congress do? What should AI companies do? What should the international community do?

These are important questions. But they leave something out. Most of us are not policymakers. We’re ordinary people living through an extraordinary transition. What does this essay mean for us?

If humanity is in adolescence, then each of us is experiencing that adolescence in our own lives. We encounter AI daily now. It writes our emails, summarizes our documents, answers our questions, generates our images. It shapes what we see, read, and believe. The abstraction of “humanity’s relationship with AI” shows up in the concrete of a thousand small decisions.

Here are three orientations I find useful:

Cultivate discernment, not just information. The “engine before brakes” problem applies individually as well as collectively. We now have access to more information than any generation in history, but our capacity to evaluate it hasn’t kept pace. Learning to distinguish signal from noise, expertise from confidence, evidence from assertion, this is the personal equivalent of building institutional brakes.

Embrace appropriate uncertainty. Amodei’s essay is careful to acknowledge what he doesn’t know. Timelines could be faster or slower. Risks could be greater or smaller. The temptation is to resolve this uncertainty into either denial (“AI risk is overblown”) or doom (“catastrophe is inevitable”). Both are forms of cognitive comfort. The mature response is to hold uncertainty while still acting. We don’t know exactly what’s coming, but we know enough to prepare.

Focus on character, not just rules. Just as Amodei’s Constitutional AI tries to instill values rather than enumerate prohibitions, we might do the same in our own lives. What kind of person do you want to be in relation to these tools? Not just what should you do with AI, but who should you become? This question has no simple answer, but asking it matters more than any particular rule.

The Stoics understood something relevant here. We don’t control events. We control our responses. The future of AI will be shaped by forces far larger than any individual. But how we face that future, with fear or steadiness, with paralysis or purpose, that remains ours.

---

The Case for Mature Hope

One of the strange features of Amodei’s essay is its tone. Despite cataloging existential risks, it isn’t despairing. Despite working at the center of the industry he warns about, he doesn’t come across as paralyzed by contradiction. There’s something almost calm in the writing.

I think this reflects a distinction worth making: the difference between optimism and hope.

Optimism is a prediction. It says things will work out. It’s a bet on the future, and like any bet, it can be right or wrong. Optimism without evidence is just wishful thinking.

Hope is different. Hope is a choice about how to act in the face of uncertainty. It says: we don’t know what will happen, but we can affect what will happen, so we should act as if our actions matter. Hope isn’t naive. It’s compatible with clear-eyed assessment of danger. What makes it hope rather than optimism is that it’s grounded in agency rather than prediction.

Humanity has survived previous adolescent crises. We survived the nuclear age, not because the odds were in our favor but because specific people made specific choices at critical moments. Vasili Arkhipov, a Soviet naval officer, refused to authorize a nuclear torpedo during the Cuban Missile Crisis when his fellow officers wanted to fire. That single act of judgment may have prevented World War III. It wasn’t optimism that saved us. It was the accumulated effect of human choices under pressure.

AI presents a new version of this challenge. The choices ahead of us are different in their particulars but similar in their structure. Will we make the right calls? No one knows. But the possibility remains open, and that possibility is enough to ground mature hope.

The adolescent who survives becomes the adult. The civilization that makes it through becomes the one that can answer Sagan’s question. That’s the work ahead.

---

An Invitation, Not a Prophecy

Let me return, finally, to where we started. Ellie Arroway’s question was never really about aliens. It was about us. Sagan knew that any civilization asking that question is already in its adolescence, already uncertain whether it will make it through.

Amodei’s essay is not prophecy. It doesn’t tell us what will happen. It tells us what could happen, and invites us to consider how we might respond. The risks he describes are real, but they’re not fate. The future is not written.

What the essay does, if we read it right, is shift our attention from the technology to ourselves. Yes, AI is powerful and rapidly advancing. Yes, the systems we’re building have capabilities that outstrip our institutions’ ability to govern them. But that gap is not a property of the technology. It’s a property of us. We are the ones whose wisdom lags behind our power. We are the ones with the powerful engine and the underdeveloped brakes.

The good news, if we can call it that, is that this means the solution is also in us. Not in some abstract “humanity” but in the choices each of us makes: how we use these tools, how we govern them, what character we bring to the transition.

Sagan’s alien civilization had no answer to give. They could only say: we made it through. The answer isn’t a formula. It’s a story, and it has to be lived.

We are the ones growing up now. The question is what kind of adults we’ll become.