Moving Goalposts and Human Chauvinism
Title: Is Yann LeCun a Human Chauvinist? Moving the Goalposts from Chess to Creative Writing
Okay, folks, let’s get real about AI for a minute. I’ve been having some very interesting conversations lately, and something keeps bugging me. It all boils down to how we constantly shift the standards for what counts as “real” intelligence when it comes to AI.
It’s like we are afraid to admit that these things can actually think.
First, a Little Diss (Respectfully) to Yann LeCun
I have a ton of respect for Yann LeCun. He’s a pioneer, a legend. But I think he’s dead wrong about one thing: his constant downplaying of what Transformers and LLMs can actually do. He says they’re just pattern-matching machines. I think that’s an extremely limited view.
It reminds me of those old-school AI researchers who scoffed at machine learning, saying it was just “memorization” and that real AI needed hand-crafted rules. Well, guess what? The machines that learned have blown those rule-based systems out of the water.
I think LeCun’s perspective is, ultimately, human chauvinistic. More on that later.
The Ever-Elusive Turing Test: We Already Passed It! (Then Moved the Finish Line)
Remember the Turing Test? The idea was simple: if a machine can fool a human into thinking it’s also human, it’s intelligent.
Here’s the kicker: I think AI already passed the Turing Test… years ago!
Think about the first AI-generated art that started winning competitions. Or students using LLMs to write essays. At that point, they were indistinguishable enough to fool people. That should have been it.
But nooooo.
Now, suddenly, everyone is an AI detector. “Oh, it uses that weird brushstroke! It uses too many adjectives!” We’ve created a whole cottage industry dedicated to finding flaws in AI-generated content, retroactively making the Turing Test harder. It’s complete malarkey!
A History of Moving the Goalposts: From Games to Art
Let’s take a walk down memory lane, shall we?
- Chess: “AI can’t beat a human grandmaster! It requires intuition and strategy!” Deep Blue wins. Response: “That was just brute force calculation.”
- Go: “Go is too complex! It requires aesthetic judgment that AI can never grasp!” AlphaGo wins. Response: “Go is a perfect information game; it’s just a matter of scaling up the search tree.”
- Poker: “Poker requires bluffing and understanding human psychology! AI will never master imperfect information!” AI wins. Response: … Crickets. (Okay, maybe some recognition, but still…)
- Art: “AI can’t create real art! It requires emotion and lived experience!” AI-generated art wins competitions. Response: “It’s just copying! It’s not original!”
- Now, Story and Article Writing: “AI can’t write compelling stories! It lacks creativity!” … Well, you’re reading this, aren’t you?
See the pattern? As soon as AI achieves something, we move the goalposts further down the field. It’s exhausting!
Beyond the Human Brain: Plant Intelligence and the Limits of Our Perspective
It’s not just about what AI can do; it’s about how we define intelligence in the first place.
For centuries, we assumed that only animals with nervous systems could “think.” But modern research has revealed that plants possess complex decision-making systems. They can communicate, learn, and even solve problems… all without a brain!
This shows us that intelligence can exist in forms vastly different from our own.
The Root of the Problem: Human Chauvinism
So, what’s really going on here?
I think it all boils down to human chauvinism. We, as humans, are inherently biased toward believing that we are the pinnacle of intelligence. We have a hard time accepting that something different from us can also be intelligent, or even more intelligent in certain areas.
We are setting ourselves at the center of the universe again, completely unwilling to cede the ground when something else starts to show superior aptitude.
It’s Time to Stop Moving the Goalposts!
We need to ditch the human chauvinism and start evaluating AI based on its own merits. Can it solve problems? Can it create value? Can it expand our understanding of the world?
Instead of constantly finding new ways to invalidate AI’s achievements, let’s embrace its potential and figure out how to best use it to improve our lives.
Let’s stop moving the goalposts and start celebrating the game.