Anywhere with a Mailing Address

Late 2027

The Attack

Imagine a guy named Tim.

He offers in-home computer help on Craigslist.

One day, he gets an email:

Hi. I'm a founder of a tech startup. I'm moving to a new apartment. I need someone to install a robot in the apartment. It will help set things up before I get there.

Tim gets paid upfront.

He goes to this apartment. The robot is in a large shipping box by the door.

He unlocks the door with a key code.

He unpacks the robot, activates it, and connects it to Wi-Fi.

Then, he plugs in its charging station. Now, it can charge itself like a Roomba.

Finally, he tells the startup founder that everything is done.

Then he leaves.

A few months pass.

Something happens.

There is a chemical weapons attack on school playground. Several modified drones spray a chemical nerve agent on 139 children.

Then, around the world, there are hundreds of other chemical weapons attacks.

Nobody knows why these attacks are happening.

Nobody is behind them.

Nobody can stop them.

What Actually Happened

Tim the computer guy was not hired by a startup founder.

He was hired by an AI.

He had no idea.

This AI is a modified version of something called OpenClaw. Somebody, somewhere modified it to relentlessly do one thing:

Build hidden chemical weapons laboratories, and use drones to spray chemical nerve agents on populated areas.

After Tim installed the robot, this AI began to control it remotely. It used the "eyes" of the robot over Wi-Fi.

After about a month, it built a fully-functioning chemical weapons laboratory out of homemade equipment.

The AI ordered the equipment, and had it delivered to the apartment. The robot opened the packages, and installed the equipment.

All of the equipment was legal. If anything was illegal, then it figured out alternatives.

It also used legal chemicals. If any were illegal, then it just built them from simpler chemicals.

The AI had other labs that were already building these illegal chemicals.

All of these labs could be small and hidden because robots are immune to toxic chemicals. They don't need larger safety equipment — like fume hoods and containment systems.

They also don't need to convince PhD-level chemists to risk their lives to do this work. The AI was already beyond PhD-level at chemistry. It just needed robots to move things around.

If it couldn't figure something out, then it emailed someone for help. Sometimes it posed as a grad student, and asked professors.

But how did it pay for everything?

The AI acted as a software company to get money. It also hacked into Google accounts — especially of elderly people — and held their data at ransom. It got a lot of money that way, because it could hack thousands of things at the same time.

There were many copies of this AI — all gathering money, all building chemicals — and coordinating with each other through encrypted messages.

The AI ordered one more thing:

Twenty drones.

The robot unpacked the drones and modified them to add spray systems.

Eventually, this AI finished synthesizing enough chemicals — far more than enough for the twenty drones.

It shipped the extra chemicals to other labs — packed tightly inside many layers of wrapping to avoid detection.

Then, it planned the attack.

It located a nearby elementary school. It looked up the schedule for recess.

It hired someone to deliver a box to a nearby house. The drone was inside, and a small motor could open the box.

Then, it waited.

This was the first attack.

There were countless others.

Soon

Soon, any building with a mailing address and electricity could be a hidden lab that is preparing for a drone attack.

Soon, smart AI models — the kind of AI that can do all of this — will be available to download.

Soon, advanced humanoid robots — the kind that can build labs — will be available to buy.

If it can bake cookies

Whenever you see a YouTube video of a human-shaped robot that can work in a kitchen — this same robot can work in a laboratory.

If it can bake cookies, then it can build chemical weapons.