Unlimited possibilities

There is a revolution underway, and everyone can feel it. The world of technology is changing. That’s not new. What’s new is this phenomenon that we keep coining artificial intelligence.

A computer program that gobbled up the information on how certain humans of the 21st century think about certain issues. Clearly there were patterns… Now, these patterns are all accessible and if you want to, you can make your own army of thinkers, thinking a bit like those humans that trained the machine.

Yes, you might not fully trust it, it’s good to be sceptical, but the program is improving and it is programmed to build trust - fast. Leaders always knew that trust can be built with three ingredients: reliability, competence and benevolence.

AI chat bots rank higher and higher for these three.

Reliability is an easy hurdle: Computer programmes were always reliable. They show up - all the time - giving you a response whenever you ask for it. Its output is non-deterministic, yes, but in a fairly narrow band.

Competence and benevolence are the trickier parts: Can it provide verifiable and helpful responses?

Over the course of last few months, the competence barrier is melting in more and more fields. Whereas the first version of ChatGPT confidently responded to a basic prompt for the result of 1000*10 with 10,002, a few years down the road, the OpenAI’s latest incarnation of ChatGPT solves and verifies unsolved math theorems like nobody’s business. Finding solutions that experts in these fields deemed unreachable for decades or even their lifetimes.

Last week, I asked Anthropic’s Claude to write me a video editing software, and it did. The video software was better than anything that I have ever used before, because it did exactly what I wanted - and more. I didn’t even have to create the cuts myself. I could tell Claude to do that for me. It finds and removes my pauses.

As a mathematician turned software engineer, it is only my practiced comfort using a terminal and my arrogant desire to teach machines what I want them to do, that gives me a head start ahead of everyone else.

Clearly it is only a matter of time, when more and more people realize that they, too, can build whatever they want. They don’t have to be geniuses or obsessive, they just need to have an idea and try it. LLMs and their harnesses are improving and building competences in many areas - definitely when it comes to writing software for computers.

As a human being, the way you are wired by evolution, by engaging with the bots, you will likely start trusting AI on the accounts of reliability and competence.

But what about benevolence? Does this computer programme running on resource-thirsty machines have your best interest in mind? Not some giant company’s, not a country’s government’s interest. Does it care about you, your community and the people, creatures and causes that are close to your heart?

Personally, I believe I have some leverage here to indulge in some distrust. Yes, the AI companies train their chatbots to tell me that they care about me and my development. But this is just the epiphany of very personalized advertisment. There is no human heart behind it, just a machine imitating human sales practices.

And whether its promises actually touch my heart, is up to me.

Well, at least when I am aware of it.

But am I? Do I know my values? Who am I? What am I here for? What do I stand for? What does my company stand for, and who does it serve?

Do I sometimes have the strength to stand back, avoid Claude and Co and think for myself - even if it is difficult?

Of course, it is not always easy: I think we all have experience by now, seeing that technology can become a crutch used by healthy people.

A crutch supports my mobility after I hurt my legs, but when I rely on it every day, my body forgets how to walk without it. I have to embrace the difficulty – possibly exposing myself to some painful steps on under-conditioned legs.

But there is a point, when the answer to the benevolence question shifts from: “The crutch is helping me, and has my best interest in mind.” to “The crutch does not.”

If I want to lead my life from a point of inner power, I won’t have a choice to slow down and examine what is good to me - in the moment.

Personally, I don’t want to be assimilated by an AI religion that allows for infinite productivity. I want to stay aware if the next prompt I write, serves me.

So, even as a practiced Software Engineer, I schedule a day off from AI, writing all my own emails, all my own slack messages, and even use the mouse sometimes – like a cave monkey.

With care,

Martin