Essay

Generative AI Optimises for the Statistical Average

Convenience removes the need to practise small habits. AI may be the next step in that quiet trade-off.

Originally published on Substack2 min read
AIThinkingHabits

My mom used to say: “Anything in excess is poison.”

I didn’t think much of it when I was younger.
But over time, I started noticing something.

I used to remember phone numbers.
I used to do quick mental math.
I used to sit with problems longer before reaching for help.

Then smartphones arrived.
They didn’t make me less intelligent.
They simply removed the need to practice those small habits.

Now I see a similar shift happening again.

At a workshop recently, a speaker said something that stayed with me:
“Generative AI optimizes for the statistical average.”

Whether taken literally or not, the idea made me pause.
Because LLMs are incredibly powerful.
They can retrieve, summarize, connect, and generate at a speed no human can match.

But every powerful tool changes what we practice.
Calculators changed arithmetic.
GPS changed navigation.
Search engines changed memorization.

So the question becomes:
What will LLMs change in us?

The more I use them, the more I notice a subtle shift.
I reach for answers faster than I sit with questions.
I sometimes generate before I think through the problem.

And it makes me ask myself:
Am I using AI to extend my thinking…
or to avoid it?

Because the value was never just the answer.
It was the struggle required to reach it.

I don’t worry about AI becoming smarter than me.
I worry about becoming too dependent on something that isn’t thinking at all.


#AI #CognitiveThinking #ThroughMyQuietLens #SmartPhones #LLMs #StatiscialAverage #AverageThinking

Originally published on Substack. Republished here as part of the Through My Quiet Lens archive.
Continue exploring

The Editorial Page →

Equality or Equity? →