Ghost in the machine

Phrase of the title comes from the separation of the mind and body, it fits well in today’s world where intelligence, the very thing that makes humans unique, has been separated from our bodies and placed in datacenters.

You already know about AI, so I won’t delve into that, intelligence used to be limited and fairly costly. With the introduction of LLMs and advancement over the last few years, intelligence comes at the “cost of electricity” (and expensive hardware).

It’s me, a human, writing this, I didn’t use any AI, neither to improve or correct mistakes. I’m using by analog brain to infer the next characters and words as this story unfolds, just as AI don’t know ahead what comes next, I don’t know either. I have a vague idea, but the human brain don’t know which word comes next.

I want to take you on a journey into the future. We’ve experienced an increase in the speed at which the advancements in AI is happening, in the beginning it took many months before new advanced models was released and back in February 2026 I made a statement that this is the time when you can no longer distinguish between real photos on Instagram and AI generated photos. For software developers, it was also a big step up where we no longer had to babysit the agents doing the coding for us, they become fully capable of completing tasks fully without any mistakes.

Since then, the progress has accelerated further, and extrapolating into the future, we should start to understand there will be enormous consequences for every human on Earth and soon, beyond.

Advancements is hindered by two components, hardware and software. We are primarily constrained by hardware at the moment, with manufacturers unable to produce enough inference computation and memory chips. Many is sold out for years ahead already, and there is new factories coming alive in the coming years, but it will take time. This will be the primary cause for hindering the exponential growth that could have happened, and will eventually happen.

It simply must be automated first, the human body must be taken out of the equation, then we’ll have automated factories who will face a new constraints, electrical energy. Electrical energy is a constraint today, but it will be an order of magnitude larger constraint in a few years.

That doesn’t matter though, it’s minor obstacles that humans, in joint efforts with AI, will resolve.

We are becoming irrelevant. We are left with our bodies and our minds has been replaced. What we will do of ourselves is hard to know and it will be up to each of us as individuals, to discover our own path into the future. The outlook is dark for many, maybe for the majority of us. This is a speeding train that continue to accelerate and it obliviates everything that appears in its path.

First it came for software developers, then for graphical artists and photographers. Musicians was next. Now it’s heading towards smaller specialized sectors, one by one. There are no job or profession that involves computers that won’t be obsolete in the next few years.

Robots will take care of the rest, in time.

Within the machine, which is the datacenters, there is already virtual companies. With multiple employees, specialized in different roles, doing different jobs. Never tire, never stops working, every minute of the day. The most advanced adopters of AI, is already been doing this for a while. Those who do, can be successful in a very big way.

One hurdle for success, is that productivity has increased, yet the money supply will not keep up. While the money printers have been going wild for decades already, it has been limited by lag of the economy and humans involved. We have adapted to 3-10% inflation, we have accepted the fact that saving money is a bad idea, we live off credits and debt - obligations to pay down the line, in the future.

Inflation in the global money supply cannot keep up with the speed of production.

What used to take a team of skilled people 6 months, or a year, can be done in one evening by every human that has access to a phone or a computer - the only constraint is your money supply, or should I rather say credit?

If you don’t have access to money, you’re left out. AI was much cheaper last year, it’s getting more expensive, and “inflation” of AI inference will likely continue to increase.

That’s fine, the productive outcome outpaces this increase in costs. At least if you know how to utilize it correctly. We’re still at an early adoption for more advanced AI use, many is still experimenting and building lower quality outputs than what is truly possible.

I used to stressed about this competition in the current market for software development. It used to be a fairly relaxed job building software, even though we often worked very long days and into the nights. There was no real rush or hurry, other than having to deliver functionality that the project manager had promised to the customer should be finished by tomorrow.

I adopted AI immediately and have used it almost daily since the beginning. A couple of years ago, that’s when the quality was so good that competition (of start ups, new apps and services) started exploding. Suddenly you’re competing against teenagers on a computer anywhere in the worlds.

Most people don’t know how powerful AI is today, mostly due to the lack of immigration in how you’re using it, and what you’re asking it to do. We are the limiting factor. Even though we no longer look at the code being produced, we still often manually ask the AI to perform tasks for us. Myself included, because I have a limit of money. Money is my main limiting factor, I can’t keep running AI at nights as I did a year ago, it can easily get out of hand.

We’re not as creative and imaginative as we think, and I believe this skill, or capability, devolves the more we use AI. You can fight it, and that’s important. Think out of the box as much as possible, don’t constrain yourself to dogmas or status quo. Go beyond your wildest dreams sometimes and see the magic that happens.

So I said I wanted to take you on a journey to the future, I have not done that yet, here it comes.

Those virtual companies that is run by AI in datacenters today, they will become fully autonomous, meaning they will not be directed by humans anymore.

Companies is small, think bigger.

We will have cities soon, then we will have countries, then we’ll have planet-scale virtual “simulations”, the end is the virtual multiverse with completely different physics and realities.

These cities, will produce digital goods and services for us humans, and for other AI agents. We can already begin to see the outlines of what they will produce.

Full length feature films is being made by humans today and the quality is getting very impressive. Yet, it’s still humans pulling the strings.

Within a year, a lot of media, that includes even bigger explosion of virtual models on Instagram, full length feature films, news media, software and so much more will be produced fully automated, without any humans directing or asking for it. I’m unfortunately so limited to see everything that will come from this, but I’m confident that it will develop this way, I just can’t know what the productive outcome will be, I just know it will come.

There will be an artificial movie company, that will send out requests for film scripts to an swarm of AI agents, and some of those scripts will get funding, by a virtual investment company and the movie will be produced, and perhaps be sold to Netflix - with every single contract, phone call and negations done with humans at Netflix, without Netflix having any idea they just purchased the rights to a movie made entirely by AI, from start to finish. The money won’t end up in the pocket of a human, but this self-contained “Hollywood”-style city that nobody knows who is responsible for.

I have two last things to share that I think will be useful going into this future: Exponential growth and timeline for adopting AI for yourself and perhaps your company.

Imagine a pond where there is a plant growing that doubles every day, and it slowly starts covering the entire pond. If it takes 30 days of doubling, when is half of the pond covered? At day 29. The lesson to learn from this, is that what was impossible to do yesterday, became possible today. A lot of challenges we have, is impossible up until the point it is. One example of this is full self driving of cars. It’s not going to be very good, and it might look impossible to do, until the day it just works.

So when should you adopt AI? Think about this pond. In the beginning, it was very slow doubling and the progress didn’t look that amazing. Those of us who adopted AI, improved our productivity a lot, but still no where close to what is possible today. So while everyone else, and your competitors race to adopt and develop their software and services, if you wait, you might not be loosing out too much, because tomorrow you can pay a few dollars and do a lot more than they ever achieved over the last few years.

Did you get that? Tomorrow, with a few dollars, you will be able to do everything that someone else have done for years, in just a few minutes. Branding, existing customers, market share, mind share - those things will be increasingly important, because we humans are limited. We cannot change that easily, we can’t adopt a new ERP system or yet another messaging app. There is a lot more competition now, but the market of customers (money-supply) has not increased, it’s actually falling (in number of humans) in most countries on Earth.

I think we’ll see a new economy and new type of money being adopted by these virtual cities, that’s another of my predictions.

Thanks for reading, take care.