The AI thinking trap — Kelford Labs Daily

As a history nerd, but, let’s face it, as an “old books nerd,” I decided to finally read all six, 600-page volumes of The Decline and Fall of the Roman Empire last year.

(My copies are actually a theologian’s annotated commentary on Gibbon’s original 1700s text, from 1850.)

As you might know, a lot of it reads like a series of names and dates, and it can sometimes be difficult to understand exactly what’s happening when you get lost in the sea of facts and figures.

So I did what any nerd might find themselves doing these days: I built myself an AI-powered tool to help me understand it. I created an Apple Shortcut that allows me to photograph the pages I’m reading, extract the text, and then have it explained to me via the Intercom feature on my HomePod.

Basically, with a few taps on my phone, any passage in a book I’m reading that I don’t understand can be interpreted and explained to me a few seconds later.

Great, right?

Well...

The problem is, a few weeks into reading the books, and a few dozen usages of the tool later, I realized I was becoming more and more tempted to rely on the explanations.

I would find my eyes starting to drift over dense passages, and my hand reach for my phone almost unconsciously.

When my brain encountered something difficult, it reflexively reached for AI.

Using AI was actually reducing my ability to read and comprehend what I was reading. It was having the opposite effect of that which I’d intended.

So I decided I had to stop using it, or at least reserve it for passages that I had already attempted to understand but truly couldn’t.

And this has been my experience as a witness to other people’s AI use, too. Those who rely on it for every email, every document, everything they’re supposed to have read: I can see them losing their ability to think and work on their own.

Maybe that’s fine? Maybe that’s just the way of things?

But what if it isn’t? What if the value we provide to each other as entrepreneurs, colleagues, and partners is actually about our abilities and priorities, and not which LLM we’re paying for?

What if, if everyone else is using LLMs as much as we are, we lose any advantage it had once given us? What if we’re forced to compete on our own merits again?

Will we have already lost our ability to think?

So I say this with all the love in my heart but all of my concern too: I think AI should be used for activities that already feel like coasting downhill. AI can help us move through quick tasks even faster.

But when something takes work, takes thought, takes effort and takes care... Well, I think that’s the work we’re supposed to be doing.

Instead of merely summoning the ghosts of work someone else has already done, that the LLM has been trained on.

If we sell our thinking, we have to think for ourselves.

Otherwise, what are they buying?

Reply to this email to tell me what you think, or ask any questions!

Kelford Inc. shows communicators the way to always knowing what to say.