Vitalik: What We Need to Do Is Not Fight AI, But Create Sanctuaries

Just finished listening to Vitalik’s latest podcast on a16z — the amount of insight is overwhelming.

He founded Ethereum at 19, and now in his early 30s, he has transitioned from living on “autopilot” to actively being a “pilot.”

The core topic of this episode is the issue we’re most anxious about right now: AI is getting stronger — what should humans do?

Vitalik’s answer isn’t “fight AI,” but to create sanctuary technologies — technology that protects us without stripping away our privacy and agency.

Today, I’m breaking down the most counterintuitive insights, practical advice, and Ethereum’s new positioning from this episode.

The Biggest Risk of the AI Era Isn’t That AI Is Too Smart, But That Humans Are Too Passive

Vitalik states bluntly that the world today is less safe and less peaceful than 10–15 years ago.

Many people are chasing a kind of “safety”: handing everything over to “the uncle in the sky” — big companies, super AI, centralized systems — letting them make decisions, manage risks, and provide protection for us.

But the price of this safety is losing our privacy and our agency.

Vitalik calls this disempowering safety — safety that makes people increasingly powerless.

This is also where he redefines the mission of crypto/Ethereum. Ethereum’s significance isn’t to “fix the dollar” or patch up the existing financial system, but to create a new option. You are free to choose to use it or not.

That is what a true sanctuary is: safe, while preserving your sovereignty.

Sanctuary Technologies: Preserving Small Spaces of Freedom for Humans

Sanctuary Technologies is a term Vitalik coined himself. It’s about creating a safe little space where you can think, coordinate, and create freely, without being fully controlled by external forces.

It has several core characteristics: not totalizing (doesn’t try to rule the whole world); preserves privacy and agency; everyone can freely enter and exit — no coercion.

Ethereum is a classic example of sanctuary tech. It doesn’t try to fix the existing financial system; it offers you a parallel option. Use it if you want; don’t use it if you don’t.

This will become increasingly important in the AI era. Because as big companies and super AI grow more powerful, what humans truly need isn’t another “system that arranges everything for you,” but a space that lets you retain the freedom to choose.

From Autopilot to Active Pilot: Vitalik’s Personal Growth

Vitalik reflects that when he founded Ethereum at 19, he was largely on autopilot.

Many decisions were just going with the flow: dropping out, writing the white paper, getting his visa rejected by Ripple — which turned into a life turning point. Back then, he felt more like the world was pushing him along.

But now he’s increasingly aware: the world is changing too fast; no one is coming to save you. You have to be your own pilot.

He gives some very relatable examples. 10 years ago, not contacting friends for days was fine. Now, not replying to messages for a day causes anxiety. 10 years ago, you could actually get lost walking. Now, with phone navigation, cities have become a series of “teleport points.”

These changes remind us: every 5 to 10 years, the world “dies once and is reborn.” If you keep living by an old script, you’ll soon fall behind.

So what truly matters in the AI era is not passively waiting for technology to take you somewhere, but actively deciding how you will use technology.

The Stronger AI Gets, the More Humans Need to Keep “Manual Mode”

Vitalik emphasizes: Active learning is 10 times more effective than passive learning, even with the same amount of time spent.

Since childhood, he has forced himself to do many things manually — like not using a calculator in chemistry class, or not using navigation when walking. The goal isn’t to be anti-technology, but to keep his brain engaged.

The stronger AI becomes, the more we need to deliberately retain some “manual mode.”

Sometimes, intentionally don’t use AI to write code. Sometimes, intentionally don’t use navigation to walk. Sometimes, intentionally don’t use a chatbot to think for you.

This isn’t nostalgia or a rejection of efficiency. It’s to prevent brain atrophy and maintain your agency.

AI can help us do many things, but if all thinking, judgment, and exploration are outsourced, humans will slowly become passengers in the system. Vitalik’s reminder is: you can use AI, but don’t become completely dependent on it.

Practical Advice for Builders

Vitalik’s insights in this episode are very direct for ordinary builders.

First, force yourself to do things manually. Even if AI can help, do it yourself from time to time to ensure your brain doesn’t rust.

Second, active learning. Don’t just let AI give you answers. Derive, verify, and do it yourself.

Third, build sanctuary technologies. Whether you’re building open-source tools, decentralized protocols, or personal knowledge bases, prioritize one thing: does it help people retain sovereignty?

Fourth, don’t outsource all your brainpower. AI can handle execution, but strategy, direction, and values must be controlled by you.

Fifth, preserve serendipity. Attend offline events, talk to real people, and don’t let algorithmic recommendations discover everything for you.

All these points point to the same core: the AI era isn’t about using fewer tools, but about using tools more actively.

Ethereum’s New Positioning: Not Fixing the Old World, But Creating New Options

Vitalik is also clear about the positioning of crypto.

Crypto can’t fix all the problems of the dollar, and it doesn’t need to pretend it can. But it can create something new without those flaws.

Everyone is free to choose to use it or not.

That is where crypto’s greatest power lies: it doesn’t coerce you; it gives you a choice.

In an era where AI is becoming increasingly centralized, this will become more and more precious. Because when more and more systems try to make decisions for you, filter information for you, and judge risks for you, a parallel option that is non-coercive, freely accessible and exitable becomes very important.

The value of Ethereum/crypto isn’t “winning against the old world,” but giving you a new world you can freely choose.

The Most Counterintuitive Lines from This Episode

The biggest risk of the AI era isn’t AI replacing humans, but humans willingly turning themselves into passengers.

A sanctuary isn’t about making the whole world safe, but giving you a safe little space where you can still keep your freedom.

Active learning is 10 times more effective than passive learning, even with the same time investment.

The world dies and is reborn every 5 to 10 years. We have to pilot ourselves.

Inspiration for Ordinary People

The stronger AI gets, the more proactive humans need to be.

Don’t outsource all your thinking to models. Do more manual things, keep your brain engaged. Participate in building tools that preserve human sovereignty — whether open source, decentralized, or personal knowledge management systems.

Remember: technology ultimately serves humans, not replaces them.

Vitalik ended by saying that we humans are the brightest stars. AI can be powerful, but what truly drives the world forward are still active, agentic humans.

One-Sentence Summary

Vitalik’s 10 years of personal experience tell us: the AI era is not an era to lie down, but an era that needs humans to actively take the wheel.

Don’t outsource your brain to models. Do more manual things, build sanctuary technologies, preserve your privacy and agency.

My biggest takeaway from this episode: before, we feared AI taking our jobs. Now, I see AI is actually upgrading humans from “executors” to “designers.”

What’s truly scarce is never computing power, but humans willing to actively think and retain sovereignty.

Full transcript available for free on Podwise.