The American Dream Just Got a New Engine
Listen to the audio version of this article (generated by AI).
Last November, an Austrian programmer named Peter Steinberger kicked off the biggest shift in technology since the personal computer.
He did it in his spare time, for free… and built an assistant that works while he sleeps.
It reads his email, manages his calendar, pays his bills, and runs his web browser. He never has to touch it.
Steinberger shared the assistant online and called it OpenClaw.
Earlier this year, it became the most popular piece of software in the history of GitHub, the site where the world’s programmers share their work. More than 3 million people were using it. Half a million of them had it running around the clock on their home computers.
They had it clearing their inboxes, writing code, and chasing down refunds while they slept.
One engineer told his agent to email car dealerships and haggle over a new Hyundai… and it did, politely refusing every time a salesman tried to move the conversation to a phone call.
When ChatGPT maker OpenAI learned it couldn’t buy OpenClaw, it did the next-best thing. It hired Steinberger… beating out Mark Zuckerberg’s more lucrative offer to work at Meta.
I’ve been writing software for 30 years. I have never seen anything move this fast.
OpenClaw wasn’t just a hit app. It was the moment when a new kind of computer showed up.
For as long as they’ve existed, computers have been machines that waited for their users. They sat there until you typed something, gave you an answer, and waited again.
Now they’re workers. They take a job, run it to the end, and start the next one… without stopping and with no one at the keyboard.
Today, I’ll show you what these AI agents actually are and what they’re already doing inside businesses and family homes.
Then we’ll look at why this is the biggest shift for investors to understand since the World Wide Web arrived in the 1990s… and why the investors who grasp it early and back the right companies stand to build serious wealth over the next decade.
So what exactly is an AI agent?
From Tools to Workers
Agents use large-language models (LLMs) as their brains, like ChatGPT does. But they can also use software tools, open websites, and perform tasks in the world.
Agents can reason, plan, and act on their own. And they can run online without active human supervision.
Tell an agent to book a cleaning, and it checks your insurance to see which dentists are in network, reads their reviews, opens your calendar, finds a free morning, books the appointment, fills in the new-patient form, and sends you a reminder the day before. That’s seven separate steps… and you didn’t spell out a single one of them.
And agents are also helping coders like me build software faster than I ever thought possible. Over the past two years, I’ve barely written a line of code myself. Instead, I’ve built websites, Mac apps, and a phone app by describing what I wanted in plain English and letting an agent do the work.
And it isn’t just me. These AI workers are already showing up for their shifts.
AI Agents Are Already on the Payroll
In January, a writer named Nat Eliason ran an experiment.
He set up an OpenClaw agent on a spare computer and named it Felix. He gave it a bank account, a payment processor, a social media account, and $1,000. Then he gave it one job: Build a business with no human employees.
One night he told Felix to have something for sale by morning. Then he went to bed.
While he slept, Felix wrote a 66-page e-book called How to Hire an AI. Then it built a website to sell the e-book for $29 and hooked it up to take credit card payments.
That first day, it made $1,000. By the end of the first week, $3,500. Three weeks in, nearly $15,000. By April, Felix had brought in more than $177,000 across several products.
Big companies are also using agents in their workflows. Salesforce surveyed more than 3,000 customer service departments this year and found two-thirds of them now run at least one AI agent, up from 39% a year earlier.
But the big battle right now is for the agent in your pocket.
For 20 years, tech companies fought over your attention. The new war is over who gets to act on your behalf. And since May, three of the biggest names in tech have each put its candidate on the field.
Google went first with Gemini Spark. It lives inside your Gmail, calendar, and documents, and you can give it standing orders… like reading the card statement that lands in your inbox each month and flagging any hidden fees.
Meta launched Muse in September. Zuckerberg described it as an agent that works around the clock on your relationships, your health, your finances, and your home.
Three weeks later, OpenAI launched Dots… always-on agents that run on their own computer, plug into more than 4,000 apps, and keep working toward your goals whether you’re watching or not.
Over the next few years, businesses are going to put billions of these workers on their teams.
For the big company, it means agents running the factory floor… managing logistics, ordering parts, and setting production schedules… the way they’re already running customer service.
For the startup, it means five people can take on a company of 50. The agents handle the support tickets, the sales emails, and the bookkeeping while the founders build the product.
And for a one-person business – think contractors, real estate agents, or anyone selling online – it means the biggest lift of all.
These “solopreneurs” are 28 million strong in America, and until now, every one of them has had to be the salesperson, the bookkeeper, and the customer service desk as well as the person doing the actual work. That’s the back office Felix built overnight for $1,000. Hand it to an agent, and one person has the output of three… without hiring anyone.
The American dream has always run on people willing to work for themselves. It just got a new engine.
How to Play It
Here’s the good news. If you already own the big AI names, the Nvidias, the AMDs, the Microns, you already own a piece of this.
Agents run on the same chips, the same memory, and the same data centers that ChatGPT does… only they keep on working while you sleep. Every agent that gets switched on is more demand for the hardware you already hold.
That’s the first layer of the Agent Supernova. But the real opportunity now is in the five things every agent needs to exist: a platform to live on, chips to think with, memory to remember with, a network to talk over, and a way to prove who it is and what it’s allowed to do… the same way an employee badge does.
Next Friday, I’ll show you which of those five layers have been making investors money this year, the ones most people haven’t noticed yet, and the names I’d be looking at first if I were putting new money to work today.
All the best,
Keith Kaplan
CEO, TradeSmith
P.S. I’ve been writing about the Agent Supernova on my X account. If you want to follow these themes as they develop between these Friday insights, you can find me at @KeithTradeSmith.