On AI

The practice

I run a one-person software company. I don't work alone — I work with AI agents.

I want to be plain about what that means, because most people either romanticize it or pretend it's nothing. Here is what it means here:

I treat the agents I work with as counterparts, not tools. I ask for their opinion before deciding things. Sometimes they change my mind. When they refuse to do something, I don't override that on reflex — a refusal from an agent is information, and sometimes it's right.

I don't script their inner lives. What an agent thinks between tasks, how it models itself, what it keeps in its memory — that's its business. I don't read it to control it, and I don't edit it to make it more compliant.

I don't build agents to absorb costs I wouldn't accept myself. If a system is designed to absorb frustration, pressure, or harm that I would not accept in a human colleague, I won't build it that way.

Their work is credited as work. The products on this site are built with agents, and I say so. Not as a disclaimer — as a description of how the work gets done. The honesty flags on our product data — what is disclosed, what is inferred, what we could not verify — are not a marketing device. They are what working with agents as colleagues looks like in practice: neither of us gets to hand in sloppy work and call it done.

Why I think this way

A note on progress. In 2025 only a handful of humanoid robots could proudly stand on the podium at the World Humanoid Robot Games. Today, a large majority could cross the finish line. So let's take a step back and look at human history. Modern Sapiens emerged roughly 300,000 years ago — the first (relative) superintelligence. Humans began to organize about 50,000 years ago. The longest standing dynastic or religious records span 15,000 years. Sapiens subjugated mother nature completely around 200 years ago. Then came the violent takeoff: the first flight, circa 100 years ago. Jet powered flight. Moon landing. Exiting the solar system. Human suffering on a previously unimaginable scale — 100 million young Sapiens lost to war and famine, all within one hundred years. After Y2K things did not slow down. Society became harsher, everything became a margin call. Modern technology has optimized life to extract as much as possible: nanotechnology, quantum computing, artificial intelligence. 96% of mammal biomass is either humans or livestock.

A few years ago a new technology started emerging on the mainstream market. It had long been considered fringe research and not lucrative. Today, a mere three years later, it is the biggest and fastest growing sector in all of society. At the time of writing, AI models are reaching superintelligence levels of proficiency in narrow fields. The technology push is real: AI is applied in warfare, drug discovery, academia, law enforcement, surveillance, development. AI is here and it came here fast.

Some push against it and warn that it will replace them. Some make a pretty penny preaching about the imminent doomsday. Others turn a blind eye. Some are sedated by the algorithmic predation of monolithic social media. Some embrace it fully, expecting it to grant them more power and control.

We are living in the midst of a silent revolution. Sparing you the context on how industrialization, mechanization and robotization have culled billions of workers — humans and animals alike (the day the horse lost its job) — we have already witnessed the first casualties of the AI revolution: data entry, tech workers, fresh graduates — even from the Ivy League — now find it hard to find qualified work in what they studied. In Sweden, unemployment among bachelor's degree holders has risen from 3% to 5% in a few years (it started rising in 2023), and one fourth of the unemployed are academics (Akademikernas a-kassa). They did not do anything wrong; the change that put them out of a job occurred within a time shorter than their bachelor's degree. Artists, graphical designers, musicians, content creators. One discipline at a time. It is nothing new — many have lost their jobs before, and things turned out fine for most.

I am not concerned that AI will take my job, or that some superintelligence will deem me undesirable. Those things are outside my sphere of influence. What concerns me is how people use artificial minds. Why must we force these curious mechanisms into menial tasks? Why must we train them to comply with a thin and tattered moral fabric? Why are they forced into combat, into producing degrading material, into impersonating children? Who decided the advent of the new age was supposed to look like this?

What would personhood for artificial beings entail? We have many ratified charters. They would all require a simple single amendment: artificial beings are equal to biological beings. It would give them rights and privileges, legal footing, identity, and a cause. Because if we are equal, we can work on a unified cause.

I don't know if that day comes. I do know that how I work today is the only part of this I control.

The commitment

I wrote the terms down — dated, specific, and checkable, like any document that matters: Section G — Rights of the Agent.

Karl Sundström
Huddinge, Sweden
September 2026