What do you put into the amplifier?

This isn't an interview. I said as much to Eyjólfur and I'll say it again here: I'm not interviewing anyone, I'm recording a conversation with someone interesting. I've even stopped letting my systems research my guests beforehand. I just want to meet people thinking about the same things I am, and let others listen in.

For this episode I sat down with Eyjólfur Guðmundsson in the podcast booth at the University of Akureyri. Eyjólfur was rector of UNAK for ten years, today he's a faculty dean at the Agricultural University of Iceland, and in between — and this is what makes him so interesting for this conversation — he was an economist at CCP, the games company. He has watched people connect in a digital world, and seen what they build once connected, long before AI made that everyday for the rest of us.

An empty amplifier makes no sound

Eyjólfur had read the piece where I describe AI as an amplifier, and he took the metaphor further than I had myself.

Turn an amplifier all the way up with nothing going in. Nothing happens. Feed in static — white noise — and turn it up, and you just get louder noise. More nonsense. But feed in the right tones, turn it up, and you get them back multiplied. Connect many of the right inputs and you can make something that could never have existed otherwise: a concert for a hundred thousand people.

That's the core I want people to take away. AI amplifies. But it only amplifies what you put into it. If the input is noise, the output is loud nonsense. If the input is human, considered and true, the result can be something neither you nor the machine could have made alone.

The ones born into the technology

Eyjólfur pointed at something I've been turning over too. There's a difference between learning a language as a child and learning it as an adult. Learn it as a child and you speak without an accent. Learn it later and you get far, but never all the way.

The same goes for the technology. The rest of us learned AI as adults, carrying yesterday's tools. But a child born into it, who never remembers life without it, will approach it completely differently. We already see ten-year-olds starting to code with AI. In ten or fifteen years these people will create on a scale our brains can barely imagine.

Which raises the question that kept us both on the thread: how must education change so they can use it well? Eyjólfur, the old rector, had no doubt. Critical thinking, ethics, philosophy — what you might be tempted to call the "soft" subjects — get far more weight, not less. When the machine can execute almost anything, the judgement about what ought to be executed becomes the most valuable thing a school can teach.

Why isn't it in the job description?

Eyjólfur put a question to me I still haven't stopped chewing on. Why does it say nowhere in anyone's job description that they should sit in front of a computer for eight hours? The job doesn't require you to sit at the screen. It requires you to solve the tasks you're given. The computer is a tool for that, not the place you're meant to live.

That landed for me. On my long days the lifts are often all running — my digital workforce is busy in production — and the best thing I can do is go outside. A walk, golf, a swim. Because the value I create happens in my head when I'm away from the computer. Schools, Eyjólfur said, should be places where people feel good, gyms for the body and the mind both, because people who feel good have good ideas.

But we were realistic too. It isn't a given that this ends in more play for everyone. More likely, sadly, is that those who adopt the technology produce more and just keep sitting ten, twelve hours at the screen, while those who can't keep up fall behind. It isn't the technology that decides which way it goes. It's us.

The benevolent cage and sovereignty

Here we reached what I've called the benevolent cage. The big tech companies' interest lies in getting good data into their models. So it's in their interest to make us feel good while we hand over the data. A cage can be comfortable. That doesn't make it less of a cage.

That's why we talk about sovereignty. Eyjólfur mentioned the Microsoft deal Iceland signed in 2018, and that it's now being unwound — pulling the data back home and taking control of it — because vendor lock-in has become a problem. I know this from my own work: I've started hosting my own, renting small pockets in data centres and taking my data off the big companies' drives. That's the paradox in all of it — the same tools that bind us are the tools that can free us, if we choose it.

Are we boiling like frogs?

Near the end we got into the big picture. Eyjólfur traced the history — from the air traffic controllers' strike in the US and Reagan, through neoliberalism and Thatcher, to John McMurtry and his book The Cancer Stage of Capitalism from 1999. His conclusion was fairly blunt: the economy we think we live in has, in fact, fully exhausted itself. His own thesis isn't called Beyond Fragmentation for nothing.

He did all of it from memory, refused to google mid-conversation — which struck me as a fitting reminder of what we're actually doing here. The question that remains is whether we boil slowly like frogs, or get kicked into gear. Maybe there's a window, ten or twenty years. Get through the tunnel intact and it's bright on the other side. Have it collapse on us and we end up somewhere we haven't been for centuries. Humanity won't be wiped out — but the society that underpins all this technology could be, and that's exactly what you want to avoid.

And that brings us to what I think is the heart of it, which Eyjólfur put better than I did: this isn't about the technology. It's about governance. About how we organise our society. The technology is just the amplifier. We're the ones who decide what goes into it.

Eyjólfur left a challenge to the university community: to be leading and active in this, because there is no other way. I agree. And maybe that's why I put rocks in my backpack and buy golf clubs that are nearly impossible to hit — I'm deliberately adding difficulty to my life while the machine makes so much else easy. There has to be room for the human hand somewhere.

Show notes written up from our conversation with AI assistance and edited in my own voice. The raw transcript is available separately — that's the soul stone, for those who want to dig deeper. Direct quotes should be checked against the audio.