Nvidia suffered a record $600 billion market cap loss, the largest single-day drop in U.S. history.

This dramatic event highlights a crucial truth: even AI couldn't predict its own potential disruption. Why?

Because AI isn’t a fortune teller—it’s a learned history professor.

AI systems have absorbed humanity's recorded knowledge – our documents, audio, and video – but they inherit our limitations in predicting the future.

The core challenge lies in our inability to grasp exponential growth. Tim Urban illustrated this brilliantly with a simple visual illustration.

Look at the cartoon on the left, when we stand on an exponential curve and look backward to predict the future, we miss the crucial detail that we're actually pressed against a nearly vertical line moving forward.

To understand this limitation, let's visit America in 1776. George Washington's world had no electricity. Messages travelled only as far as someone could shout. Horses were the fastest way to move. Even hot showers were an impossible dream.

The exponential progress from 1776 to 2025 would have been unimaginable 250 years ago.

We – both humans and current AI models – tend to think linearly.

Look at the cartoon on the right, we look back 20 years to predict the next 20, or we extrapolate current trends forward. Both approaches fall far short of reality when dealing with exponential growth.

Today's AI models, despite their impressive capabilities, are essentially supercharged aggregations of human history and intelligence.

Until they can transcend human limitations in understanding exponential growth, they'll share our blindspot in predicting dramatic technological shifts.

That is it for today, take it easy until next time.

Image credit – Tim Urban’s blog “Wait but why”.

My blogs are “Notes to Self”—a way to bookmark inspirations, learning, and random ponderings.

Read all my “Notes to Self” at view all blogs.