Zero Billion, Inc.

Secret Master Plan

November 18, 2026

What is the end goal of all of this incredible AI progress?

Is it to make jobs easier? Is it to fire people and make companies cheaper to run? Is that what the data centers are for? Is that what superintelligence is for?

I don’t think so.

These systems improve every few months. They write, they draw, they design, and in many fields they now work at the level of a graduate student. If your work is one of the things they have learned to do, it is a strange thing to watch. And nobody has told you what you get in return.

Find a photograph of a kitchen from 1970 and hold it up against your own.

The stove is the same. The refrigerator, the table, the walls and the window are the same. The room costs more than it used to. The only real difference is that there is now a screen in it, and the screen is extraordinary.

For fifty years, nearly all of our progress has gone into the screen. An AI that answers your questions and does some of your computer work is more progress in the screen.

Some things did get cheap.

At 1957 prices, a megabyte of computer memory would cost almost $4 billion in today’s money. Now it costs about a tenth of a cent.

Then it reached the things we buy. A $10,000 TV in 1998 costs $164 today. A $10,000 computer costs $415.

What $10,000 in 1998 costs todayUS consumer prices. Each gridline down is ten times cheaper.
Got cheaperGot more expensive
$20k$10k$1k$10019982005201220192026Housing $23,500Food $21,700Computer $415TV $164
See the numbers
YearHousingFoodComputerTV
1998$10,000$10,000$10,000$10,000
1999$10,300$10,200$6,800$9,300
2000$10,600$10,400$5,300$8,400
2001$11,000$10,800$3,800$7,600
2002$11,400$11,000$2,800$6,800
2003$11,700$11,200$2,200$5,900
2004$12,000$11,600$2,000$5,100
2005$12,300$11,900$1,600$4,500
2006$12,700$12,100$1,400$3,800
2007$13,200$12,600$1,200$2,900
2008$13,500$13,300$1,100$2,400
2009$13,700$13,600$941$1,800
2010$13,600$13,700$873$1,300
2011$13,800$14,200$787$1,100
2012$14,100$14,500$712$917
2013$14,400$14,800$649$774
2014$14,900$15,100$601$665
2015$15,300$15,400$547$570
2016$15,800$15,400$508$461
2017$16,400$15,600$486$398
2018$16,900$15,800$467$332
2019$17,500$16,100$450$268
2020$17,900$16,600$438$231
2021$18,400$17,300$451$244
2022$19,500$19,000$446$218
2023$20,900$20,100$423$193
2024$22,000$20,600$409$181
2025$22,800$21,100$397$166
2026$23,500$21,700$415$164

Source: US Bureau of Labor Statistics, Consumer Price Index, annual averages, through October 2026. Each line is what $10,000 of that item in 1998 costs in later years. Housing is the shelter index. TV and computer prices are adjusted for quality.

But in 28 years our food and housing have just gotten a lot more expensive. $10,000 of food now costs $21,700. $10,000 of housing costs $23,500.

Housing and food never got on that curve.

Most people spend most of their waking lives at work, but a lot of them are not doing it because they love their work - they are doing it to cover rent and food.

Dario Amodei has written, in Machines of Loving Grace, about the diseases that superintelligence will cure, and the incredible economic ramifications across a range of things. I would add two additions to that list. I want housing and food to become too cheap to meter.

However, it cannot happen through a screen. Food has to be grown, stored and moved. A house has to be built, and building is labor and materials.

Manhattan holds about 1.8 billion square feet of buildings. By my estimate, raising that much takes several billion hours of work. 5 million humanoid robots working around the clock put in about 20 billion hours in six months, recreating Manhattan.

Imagine a world where Manhattan level cities are being created every year in the world. Rent and food that costs cents. And, eventually, children who find the idea of paying for food strange. This is where we are heading.

Humanoids are the backbone of this new world. But a humanoid is not created knowing how to do any of this. It has to learn from people first.

Waymo’s cars drove for more than ten years with a person behind the wheel before the empty seat. Tesla’s Robotaxi began with a person in the passenger seat. People still help these cars from far away when they are confused. And each time the cars come to a new city, it starts again: people drive the streets first, then people supervise, and only then does the car drive alone.

Every city needs a learning period. Robots will need one for every physical location.

First they have to learn to see and hear. That takes ordinary people recording ordinary life: cooking, repairing, talking, in their own homes and their own voices. There are about seven billion smartphones in the world, so most of the instruments are already in people’s pockets.

Then people will operate the first million robots by hand, from a distance, and the robots will learn from their hands.

Then the humanoids will create an abundance, in food, construction, material extraction etc.

This is an enormous amount of work, and it will go on for years. It is also an unusual kind of work, because its purpose is to end.

I think people who do that kind of work should be paid for it. This month we paid more than five thousand people for talking and for recording video. The companies building the models use it to make them better.

Today, the number of people who live in a world where housing and food cost almost nothing is zero. I think it will be a billion. We named the company after the distance between those two numbers, and our work is to make sure people are paid while it is crossed.

This is the future I am excited by. If you are too, join us.

Ansh Chopra
Founder, Zero Billion

Zero Billion, Inc. © 2026PrivacyLegal