Key Takeaways
- AI agents are ready to run profitable businesses today, but primarily in domains focused on scale and repetition, not innovation.
- Their sweet spot: managing the entire operational stack for high-volume, low-margin e-commerce ventures where human involvement makes the business unviable.
- Forget creative tasks like selling bespoke SVGs; agents are best for non-innovative activities, like taking on tasks from platforms such as Upwork or Task Rabbit.
- The "attention economy" already shows this pattern: AI-generated content is mass-produced and published to find viral success through sheer volume.
- The distinction is sharp: agents for making money now vs. agents for future research experiments. Focus on the former.
The Agent Profit Playbook: Scale the Mundane
Forget the sci-fi dreams of AI agents inventing new industries or out-thinking human strategists. Lukas Petersson and Axel Backlund from Andon Labs have a more grounded, and frankly, more actionable view: AI agents can make money right now, but you need to know where to point them. They distinguish sharply between using agents for research and using them to generate actual profit.
Swyx, a podcast host, cut to the chase, saying, “it's like this this like seriously we should do this to make money not as a research experiment.” This isn't about pushing the boundaries of what AI can do, but rather where it currently provides value that translates to dollars.
Where does that value lie? Petersson points to e-commerce. He suggests, “I think it can be done today, but you would do it in like e-commerce where it's like the probability of success is like really low no matter if a human or an agent does it, but like an agent could surely manage everything.” This is the core insight: agents don't improve your chances of success in a difficult market, but they slash the cost of managing that market. If humans fail 90% of the time in a niche, agents might fail 90% of the time too. The difference is the agent can run thousands of such ventures simultaneously, cheaply, and without complaint.
Automating the Arbitrage and Attention Games
So, what does this actually look like? It's not about agents building the next Apple. It's about them executing repetitive, non-innovative tasks at scale. Axel Backlund highlights arbitrage opportunities, asking, “I mean, there's also the other side of like have it just go on Upwork and let loose, you know?” This implies agents could take on countless small, discrete tasks that are too time-consuming for humans to manage at a profitable rate.
Conversely, Petersson warns against deploying agents in creative or high-value design work. He observed efforts where an agent “started like a design studio and like try to sell like SVGs for $100. Like it's just like it's not providing any value.” Agents today simply don't have the nuanced understanding or creative spark to deliver genuine value in such fields. They can generate, but generating doesn't equal providing value.
Swyx notes that this strategy is already playing out in a less direct, non-monetized economy: attention. He explains, “it's already happening in the nonmonetized economy which is the attention economy right so a lot of people are making AI videos and just posting them and like spamming 20 of them one of them works and then double down on that one.” The method is clear: generate volume, automate distribution, find what sticks, and scale it. This same principle, applied to a monetized context, is where agent-run businesses will first succeed.
What to Do With This
Identify a business process in your current venture or a new idea that is high-volume, repetitive, has a clear input/output, and doesn't require complex innovation or subjective value judgment. This week, task an agent with managing the entire operational flow of such a process, not just a small piece of it. Think about launching a micro e-commerce store in a niche with low human barriers to entry, or identifying a digital arbitrage opportunity that scales through sheer repetition and volume.