A misty village woodshop: a robot arm holds an oak plank while an older Korean man planes it by hand and a teenage apprentice watches.
A village pottery workshop in watercolour mist: a clay 3D printer prints a tall vase while a Ghanaian potter smooths a bowl by hand.
Philip Deng The Future
Is Not
For Profit
AI, Capitalism and the Economy We Could Build Instead

A book of essays · in progress

The Future Is Not For Profit

The future is not for profit. The future is for us.

Standalone essays on AI, capitalism and the economy we could build instead. I'm giving it away free: read it online, download it, or listen to it in my voice.

Free when it's done. I'll only use your email to send you the book.

Why this book

Nobody Has a Plan

Tech doesn't have a plan. Politicians don't have a plan. No one has a clear vision for how the world should look, how the economy should change. Everyone knows that's a problem.

AI is going to become the engine of the next global economy. Hooking up autonomous digital super-intelligences to this version of capitalism is not going to work.

We are spending all of our time marveling at the engine. Almost no one is talking about the roads.

That's why I'm writing this book.

It's also why I'm giving it away. I'm not doing this to make money. There will be a way to chip in if you want to, but the point is getting these ideas out and getting the conversation going.

Inside

The essays

  1. Things Going RightWhat the worst AI scenario actually looks like
  2. A Little Grape JuiceSeeing AI for what it is, without losing the wonder
  3. GodisgoodCapitalism has an alignment problem, and AI is about to break it
  4. Faster HorsesWhat it will take to design the economy after AI

Each essay stands on its own. More are on the way.

At my pace so far, it's done March 2, 2027. Cheer me on →

Read a sample

A ceramic solar lantern, glowing, beside a hand-crank radio on a folded wool blanket, in loose watercolour.

Things Going Right

What the worst AI scenario actually looks like

The world is waking up to the extremely serious threats posed by the arrival of digital superintelligence. Deepfakes so sophisticated we lose touch with reality. Super-tuned, micro-targeted propaganda and mass surveillance. AI super-hackers capable of scrambling critical infrastructure like banking systems or the power grid. Genetically engineered bioweapons built to spread fast and kill. Swarms of autonomous killer drones. Computer systems so smart, so capable, and so rapidly improving that we lose the ability to understand, let alone control them.1

There are smart, serious people working on every one of these problems, and I’m grateful for their dedication. A couple of these scenarios genuinely terrify me. But to me, the one that seems most obviously threatening is the potential for widespread, rapid, and permanent job loss caused by AI. It’s on the experts’ list too, filed under “systemic risks,” alongside the concentration of power and the erosion of privacy.2 But rising unemployment doesn’t make for much of a movie. Nobody is hunted by a drone or poisoned by a lab-grown virus. People just stop getting hired, and then they stop getting paid. So why does it scare me more than all the rest?

One thing stands out from the rest. All those other problems are ways in which AI could go wrong, while displacing human labor is actually the plan, it’s things going right. Unlike the other scenarios where people are working on ways to prevent something from happening, this is actually the outcome we’re aiming for, which makes it vastly more likely for a version of it to come true.

Just how hard are we trying to bring about the era of machine labor? The world is on track to spend $2.5 trillion on AI this year alone.3 By 2030, Bain estimates, the industry will need about $2 trillion in new revenue every year just to pay for the computing power it’s building.4 Where does $2 trillion a year come from? The investors funding all of this aren’t shy about it. At Sequoia Capital’s AI conference last year, partner Pat Grady laid out where AI products are headed, and the firm summed it up as “from tools into copilots and ultimately autopilots, shifting from software budgets into labor budgets.”5 A labor budget is, of course, what a company spends on people. Even Bain’s math counts on companies using AI to cut costs in sales, marketing, customer support and research, and most of those costs are people.4 So the more money goes in, the harder everyone has to push to make AI do the work people are paid to do today.

Leave your email and the rest of the essay opens right here.

Be the first to read it

The Future Is Not For Profit

It's free. Leave your email and I'll send it to you the day it's done. Every name on this list is a reason to finish it.