poetico.ai

"Do you remember that time like five years ago, when all the AI images were kind of yellow?", Kate asked.

I did not. Were there AI images already five years ago? Even though so much has happened since 2025, everything since then felt so similar and the passage of time seemed completely disturbed.

She continued without waiting for a response:

"It was because they had already fed almost all publicly accessible images into these things. And then they flooded the Internet!"

I don't even remember how she got from her own question about my novel to this topic, but she always finds a way to talk about the things she knows. Which is not a bad thing, I like it very much. She likes to talk about how bad technology and AI and LLMs and AGIs and all these letter combinations I don't really know have gotten, even though somehow this is how she earns money. But at least she earns money.

"Sorry for the tangent, but you know, this is not a single incident. This is how all of it works. It doesn't create anything. It doesn't understand anything. And that's why I think that actual writers like you will always have a place. So how _is_ your novel going?"

I was hoping she forgot.

"Well, the closer I get to finishing my first draft, the more I fear that this whole thing was just a waste of time."

I saw that she wanted to protest and say something, but I quickly continued talking.

"'Cause I saw a video recently where an AI author showed how he writes and how he has good ideas but just never felt comfortable enough executing them for a few hundred pages. Then he started using AI and now he publishes twice a week. And I was thinking that I should maybe at least try it once... To see if one could read the difference."

"Oh Harper...", she sighed. Pity and disappointment - I could feel it. "This is a cultural thing. Actual reading is not valued anymore, people skip the whole book with their little _Reado_ like they skip through movies. But when we come back from that - and I think at some point we will have to destroy most computers - no AI can write how the people actually felt during this time. And actual readers will always want to read actual writers." She always could say something that felt universally true.

On my way back home I realized that I actually kind of understood what she was rambling about. It was when I saw the (AI-written) headline in the subway. And when I read the article about some disaster (usually a flood caused by a glacier or a river, somehow each time a little closer), it all clicked.

It was not only that the AI couldn't feel things and only guessed based on what millions of humans wrote before about how people who lost everything had to feel. It was that since everyone started using chatbots there has been much less human written text. Much much less. And now the texts it spat out - like the images five years before - were yellowing. And this wouldn't be fixed, because it couldn't be. They would just put a blue filter over it again and hope that it would look kind of normal. Almost all the texts one reads nowadays feel completely similar and the few different ones feel similarly different. Even the articles that still had names below them adapted to the writing "style" of the ones without names. I thought of something Kate said in one of her last rants: "It's one big blur iterating on itself. It's a snake eating its own tail but telling you that it will be a super-snake next year and somehow now we don't have jobs or friends or cinemas."

I still thought about this when I arrived at my flat. This was unusual because I had always been someone who just went on with it. I never thought about AI much. I thought it was kind of cool a few years ago when we were told that it would cure literally all diseases and in the meantime, it can replace your partner and write your e-mails. And while it has helped in the medical field (as far as I understand it, it has not cured all diseases but sped up the processes), most of the young doctors got through med school only with the help of a chatbot and also most people don't have the money to go to the doctor and also I kind of miss writing e-mails. AI had become so omnipresent that thinking about a world without it was like thinking about a world without smartphones and cars. And maybe I was just getting old, but a world without AIs and smartphones and cars sounded really nice.

When I got off my phone and decided to go to sleep, another one of Kate's sentences got stuck in my head. "An actual reader will always want to read actual writers." I didn't know what to do with that thought. I didn't want people in a few decades to maybe, possibly read about how we felt in the boring 30's. I wanted the people now to wake up and realize that another way of life is still possible. And I wanted to sleep.

It was 03:14 when my phone told me that I should go to sleep. It was 04:37 when I realized what I wanted to do.

It all came at once and I struggled to write down everything that I could think of. It had been many years since I felt inspired and the feeling overwhelmed me. I needed to sort my thoughts and I needed to start a business with the stupidest name I could think of. Something like scriveAI or humanGPT or writo or poetico. The trend of every company ending in AI or GPT was probably 2 years old now, but the lowercase and the nonsense words were essential. I still decided to go with humanGPT, but quickly found out that a product with that name already existed. So poetico it was. When I looked up my lamp was already glowing, indicating that a few levels above the sun was shining through the windows. I got up.

I called Lem. I remembered that they used to do some graphic design before every logo and website was created by these hellish machines. Now I needed some good old human-made advertisement. We met up half an hour later and I started to explain:

"So you know how everything you read everywhere is the same AI-written slop?"

"Yeah."

"But what it cannot write about is how we feel about that slop because it writes a million times more than we do. But what if we..."

I felt a grin appearing on my face.

"What if we create a company that made a _technological breakthrough in the research of artificial intelligence_ and that has created a _brand new model_ that writes texts completely human-like and takes the modern human experience into consideration."

I paused dramatically.

"But actually all the texts are written by us writers, who lost their jobs to those machines."

Lem's anticipation was replaced by skepticism.

"Okay, so it's really funny, yes, but I don't know.

It feels strange."

"Strange?"

"Well, we want others to see that we are different from the machines. I'm not sure if disguising us as one helps that."

"But we'll reveal it at some point. We will make a big marketing push saying that this AI gets through the Turing Test every time or something like that. A chatbot can only write from something that has already been written. We will just write about what the people think and feel, what they wouldn't write down."

My long-abandoned horrible marketing background was showing and it was working on Lem.

"Okay, that's not just funny, I think this could actually do some good."

"Every AI company is so immensely overvalued. If we just get a small piece of that pie, lots of our friends could write again for a living."

I was overselling it and we both knew it. But any dream these days would do.

"I don't know if we can write as fast as a chatbot."

Fair point.

"Yes, but we just make ourselves the subcontractor. We don't have a chatbot, we interact with _the AI_ directly and return our results. And we can write fast. It doesn't have to be good, nothing is good anymore, it just has to be human."

I think that convinced them and in the afternoon we were already calling lots of other writers.

When I got home after that I directly wanted to try and set up everything to register our brand and business, but for some reason, I took one of many empty notebooks out, got a pen, and continued with my novel.