TL;DR
- We are entering an economy where AI creates AI and sells to other AI. When my agent books a trip through a platform’s AI, which talks to a hotel’s AI, who is the customer – me or my agent?
- The history of humanity can be read as a history of what we stopped doing ourselves: first physical strength, then calculation, then memory. With AI, for the first time, we are delegating thinking.
- More and more people treat the digital version of reality as the official one – from the map on their phone to the filtered selfie they bring to a surgeon.
- If we delegate execution, judgment and choice, maybe the only thing left to humans is intent. That makes good intentions the most important thing in the economy.
- My personal answer: being human might mean sometimes… consciously choosing to be inefficient.
When did you last watch a film or a series that an algorithm didn’t recommend to you? Something you found yourself and trusted your own judgment that it was worth your time?
Many people can’t remember. And it’s not just about entertainment. Netflix has said that about 80% of the hours streamed on its platform come from recommendations. We make decision after decision based on algorithms, trends, ads and whatever we saw online last.
Which leads me to a question I keep coming back to: am I living my life, or is someone else programming it?
Who is the customer: me or my AI agent?
Economics lesson number one is supply and demand. People build businesses and sell to other people. Then people built software and sold it to people.
Now we are entering an era with no precedent: AI creates AI and sells to another AI.
Think about planning a holiday. You ask AI to find the best hotels, the cheapest flights, things to do, and to map a route. One step further, your AI agent books everything. It talks to a booking platform’s AI, which talks to the hotel’s AI. Three layers of machines talking to each other, so that one human action can happen.
So who is the customer here? Me, or my agent?
This is not science fiction. In January 2026, Google introduced the Universal Commerce Protocol, an open standard that lets AI agents create checkout sessions, and lets people buy directly inside AI Mode in Search and the Gemini app. At Google Marketing Live in May 2026, shopping went further in the same agentic direction.
As someone who has done SEO for years, I see the irony. SEO was about optimizing a site so a search engine would choose it. Now, with GEO, I use AI to optimize my site so that another AI likes it and recommends it to a third AI. Much of what we learned about marketing funnels assumes a human walking through the steps. Increasingly, the human doesn’t walk through them. Their agent does.
A history of what we stopped doing ourselves
We usually tell the story of technology as a story of what we achieved. But maybe the history of humanity is really a history of what we gradually stopped doing ourselves.
Thousands of years ago, a person hunted, built, made fire, cooked, guarded, remembered, gathered and planted. How many people around you can do all of that today? I can’t. Not because we’re worse, but because we don’t need to.
Then we started delegating:
- To other people. Society gave us professions – the blacksmith, the farmer, the doctor, even the shaman.
- Physical strength, to machines. Steam engines, motors and factories meant we no longer had to lift ten tons ourselves.
- Calculation, to computers.
- Memory, to the internet. Why memorize the multiplication table, if you can look it up?
When did you last use a paper map? We get in the car and follow the navigation, because it probably knows better. Most of the time it does. But it’s also where we first show signs of over-trusting machines. Researchers who studied 158 so-called “Death by GPS” incidents found cases where people followed navigation into danger, including 52 deaths, simply because they trusted the screen more than what was in front of them.
All of these delegations had one thing in common: we handed a machine a repeatable action. AI is different. For the first time, we are delegating intellectual work. We are delegating thinking. Analysis, writing, comparing, judging.
What becomes valuable when knowledge is free?
Some people call this a crisis of human expertise. If knowledge is almost free, and any answer is a question away, what is still worth something?
Many believe we are moving into a trust economy, where your most valuable asset is being trusted, being authentic and having a reputation that other people – not machines – believe in.
I think it goes further. I call it the synthetic economy: synthetic identity, synthetic trust, synthetic memory. If my AI agent decides whom to trust, then whoever can manipulate the information it reads can manipulate the decision.
When the screen becomes the official version of reality
There is a phenomenon you’ve probably noticed in yourself or in people around you: digital-first thinking. More and more people accept the digital representation as the real thing.
I recently heard about children who, asked whether a building in front of them exists, said no – because it wasn’t on the map on their phone.
Plastic surgeons have talked about this for years. Patients bring a filtered selfie and say, “I want to look like this.” In the American Academy of Facial Plastic and Reconstructive Surgery’s 2022 survey, 79% of facial plastic surgeons said the wish to look better in selfies keeps trending upward. When a doctor explains that cameras distort faces, some patients answer: “Yes, but what matters to me is how I look on camera.”
The more our eyes live on screens, the more authority the digital world gains. And that is frightening for personal judgment, and for the ability to live your own life. We first consume the representation of reality, and only then reality itself.
Living through an AI layer
Imagine experiencing the world through an AI layer that tells you: this is important, read this. This person suits you as a friend. This app says you’re highly compatible with this partner. This idea is good. This risk is too big, don’t do it.
For business, this is a superpower. Imagine simulating every decision: scenario A, B and C, with forecasts for each. At Webit a few days ago I heard Plamen Russev share that he and his family use digital twins for their health. As I understood it, they help simulate biological markers and longevity. That’s a good example.
But in my personal life? If I always take the shortest route, the cheapest flight, the restaurant with the most stars, the person who is “useful” to talk to – when do I stop optimizing and start living?
This is also why AI companions don’t seem as far-fetched to me as they sound. We already build parasocial relationships with people we’ve never met. Almost every job I’ve had came through people I knew only from Facebook or Instagram. Clients trust me with their teams because they know me digitally. And according to Common Sense Media, 72% of US teens have already used AI companions, and about a third of the users have chosen to discuss something serious with AI instead of a real person. If you haven’t seen Her, watch it.
If we delegate everything, what’s left of us?
The fashionable word in the US is “orchestrator”: the human moves from doing the work to setting goals and telling agents what to do.
But if we delegate thinking, goals, judgment, choice and execution, what remains of the human? Maybe only intent.
And if intent is all that remains, then the whole economy rests on one question: are our intentions good? Maybe the old cliché about the struggle between good and evil, told in so many religions and stories, turns out to be the only thing that really matters.
I see a version of this in my teaching. People who receive knowledge fully ready-made rarely use it. People who go through the process – who search, try and struggle – get results. Knowledge shouldn’t be a transaction. Experience needs some hardship in it. When we remove the hardship, what does experience even mean?
Why I choose to be inefficient
So here is where I end up. Maybe a human is the only being that can, and wants to, consciously choose the inefficient option. Just to live.
Not to optimize everything. To spend three hours drawing for no reason. To spend four hours in a park reading a book, and never post about it. A tree falls in the forest, nobody hears it, and it still matters.
There’s no single right answer here, and I’m not trying to scare anyone. I want to open the conversation: where do you draw the line and say, “This, I will never delegate to AI”? Regulation won’t keep up, so I believe each of us has to draw that line ourselves.
This article is based on my video (in Bulgarian) about the synthetic economy. For a neutral explainer of the trends behind it, read The Synthetic Economy, Explained on Egorithms.
About The Author
