AI Can't Fuck it Up!

The Unmatched Power of Human Error

In this article, I'm going to argue that the true power of Human Nature is its ability to make random mistakes and to recognize the value in their outcomes.

On the other hand, the true power of Artificial Intelligence is scalability. Once a model learns a new skill, that skill can be distributed worldwide within minutes. Do you remember Neo learning how to fly a chopper in "The Matrix?"

So, is Natural Intelligence a thing of the past?

After all, we, "old humans," need weeks to learn how to drive a car, months to play guitar, and years to become a decent chess player. At first glance, it looks like we'd overstayed our welcome.

The true power of Human Nature
is its ability to
make random mistakes.

But then I had an idea, an intuition if you concede it.

I began asking some different questions:

• What would be the world without penicillin?
• Or without X-Rays?
• Or microwave ovens?
• Or Vulcanized rubber?
• Velcro? Pacemakers? Post-it notes?
• ... or the Slinky?

You don’t need to answer, though.

I've run those questions through ChatGPT 4 to have a taste of what a super-intelligence may think:

🤖 "Without penicillin, many bacterial infections that are easily treatable today would be lethal. Medical surgeries would be riskier, life expectancy would likely be lower, and child mortality rates would be higher."

🤖 "Without vulcanized rubber, tires, seals, and many industrial parts would be less durable and effective. Transportation and machinery would be less reliable, affecting industries and daily life."

🤖 "Without X-rays, medical diagnoses and treatments would be less accurate and more invasive. Many conditions would be harder to diagnose early."

🤖 "Without microwave ovens, food preparation would be slower and more energy-intensive. Convenience foods designed for microwaving would be less common. People would rely more on traditional cooking methods."

😅 OK then, by my understanding, microwave ovens were not such a good thing for humanity. If I read correctly, ChatGPT believes that without those ubiquitous appliances, we would actually eat healthier. I concur.

What would be the world
without penicillin?

But what about the other inventions?

Let's make a short recap:

• No penicillin?
  ➡️ We die earlier and for silly reasons.
• No X-Ray?
  ➡️ We die earlier due to missing diagnoses.
• No Pacemakers?
  ➡️ We die earlier due to common arrhythmias.
• No Vulcanized Rubber?  
  ➡️ You could not eat avocado in the middle of Minnesota's winter.

So we agree that inventions like penicillin, X-ray, and Vulcanized rubber are GOOD FOR HUMANITY.

Let's now run a different question to ChatGPT:

Would it be possible for an AI to invent penicillin,
If it wasn't already invented?

🤖 "No, an AI wouldn't be able to invent penicillin on its own. It can analyze data and suggest experiments, but the actual discovery process involves empirical lab work that AI currently cannot perform."

🤖 "No, an AI could assist in simulations or recommend experiments, but the physical process of invention and testing would still require human intervention."

... And it goes the same way for the other inventions I submitted to the Great Brain...

It appears ChatGPT wouldn't be able to help humanity regarding some fundamental discoveries. And it is quite open about it!

At this point, I really got intrigued by the critical power of the bot. However, I was shaken by the "cannot currently perform" as it implies that IT believes it will be able to perform such discoveries eventually.

Anyway.

Here we go with the last question, I promise.

What missed accidental invention, if never discovered, could have halted AI development?

🤖 "If penicillin hadn't been accidentally discovered, widespread bacterial infections could have severely impeded human progress, potentially making the development of artificial intelligence unlikely."

Here we go again with penicillin.

Feels like we're going in circles from now on.

But I feel much better after writing this piece. I believe even more in our inexplicable capability of making mistakes and making something useful out of that.

Shortly put, I believe AI will never be able to fuck up as bold and spectacularly as I can do. And that makes me irreplaceable.