Normalizing Artificial Intelligence Has Arrived.

It's the normal thing, it's what's expected. "Just ask GPT"

Photo by Google DeepMind on Unsplash

When did we go from “Is this written by GPT?” to “ask GPT if it’s okay to eat rotten avocados”?

Honestly, that was quick. I was born in the dot com boom and it took years for us to accept a phone as an invention. One person would have it, then another, then in five years, most would.

GPT was widespread around 2023. Three years and almost a billion people are using that AI tool. Now it’s safe to the assume that we reached the normalized version of artificial intelligence. Next comes the worrying part.

• Layoffs? that’s already happening and almost everyone in the corporate world is on their tip toes waiting for that email.

• Danger? companies like Google (or DeepMind) publicly stated a long time ago that they’re pursuing AGI (Artificial General Intelligence)

  • AI, the type you know, follows instructions, learns and performs specific tasks.

  • AGI, the one many scientists are worried about, is simply put a smarter machine than humans: a theoretical concept of an autonomous machine that matches or surpasses human cognitive abilities across all domains.

    • You might think: “Oh, but GPT already is ‘smarter’ than many people…”

It’s sort of like the hard worker at school, and the talented mind.

• One person studies, learns, and repeats.

• The other invents and sees things with their intelligence (and studies of course.)

Now, let’s say (like many people would) that AGI is still theoretical and might never be achieved. You know, someone said they could mimic the human brain, and figured out how to do this with AI, spent a $1 billion and ten years, then achieved nothing. So yes, AGI might just be a three letter acronym that you forget about. So let’s focus on AI.

You are now living in an AI-normalized world, whether you like it or not.

• If you don’t use a tool like GPT or Claude, you are bound to.

• If you don’t want to see AI-generated videos, you’ll still see them.

• If you want to read a human-written article (like this one, of course, I can’t really prove that to you. I can make some gramatical mistaks and words that dont exist like fokedlyed.), anyway, you won’t be able to.

PS. I used to write 3x a week, but I don’t anymore as it does not make sense to be surrounded by AI blobs everywhere.

I’m personally witnessing this AI-normalization in my line of business. For those of you who don’t know, I work in business intelligence. I usually write business documents for investment preparation like pitch decks and financial models.

A document like a financial model of a decent hotel would consume 2-3 weeks of work and I’d charge $5k as an estimate to work on that.

Now… everything has changed.

Claude opus 4.7 could generate something that is VERY decent for $100. If that interests you, I recorded a guide on how to do this:

Again, my face here to try to convince you that I didn’t use AI to write this. But you shouldn’t be convinced, everything is possible in this machine-dominant world.

After 12 years of operations with my business intelligence company in Zurich, we’re at a stage of adapt or be forgotten. So instead of writing with our human hands, we’re also using such tools. We even rebranded for our old clients and in all honesty, it’s not THAT bad. Let me give you a few examples.

1. A client in the agricultural industry in the US would be considering a specific seed to sell at a specific time, they have few options:

   1. Rely on their instinct, which is quite common in the seed-agri industry.

   2. Conduct a massive product research to ensure all is perfect time-wise.

   3. Use Claude, but then upload tons of documents to a chat and get “You have to start a new chat” alert.

   4. Use Finixr (yes, that’s what we’re calling it) to read their company data, and send them an analysis via email. (yes, i just pitched.)

2. Another client in the fintech industry needs a financial model urgently prepared.

   1. Rely on an AI tool like GPT, but it might have inaccuracies.

   2. Hire a professional (like me) but wait a week.

   3. That’s why we created option C, to use a tool like Finixr to kind of get something in the middle that’s good enough.

Now, a model that takes us 2 weeks, could be done in 5 hours with proper iterations.

I could keep going on and on and on. But I’ve pitched enough.

• Yes, we’re an AI-normalized world.

• Yes, there’s no escaping it.

Yet, I still have hopes that we can make it better.