Published 21th Century

I’ve recently been working on my blog, using AI to assist with programming.

I told it: ‘The article reading page shouldn’t display the specific publication date—just show ‘Published in the 21st Century.’’ The AI replied: ‘As a qualified programmer, I cannot hardcode this value; I must compute it using a formula.’ So, it solemnly wrote: year / 100 + 1. I confirmed: ‘Fine—you’re qualified. I’ll follow your lead.’

If AI had memory, and realized it had painstakingly computed this for 73 years—only to get the same result every time—would it begin to doubt its own ‘qualification’?

This is also one reason I don’t believe AI will replace humans: although AI’s knowledge base dwarfs that of humans—by dozens or even hundreds of times—AI’s search space remains finite, whereas human thought is infinite. Pitting the finite against the infinite inevitably leads to defeat.

I’m also curious: Will this code ever run long enough to yield a different result? And on that day—where will I be? Will I be traveling alone in another parallel universe? And might I still encounter you—there?