More Compute Power Doesn’t Produce AGI

2m read

·Nov 3, 2024

The artificial general intelligence crew gets this completely wrong too. Just add more compute power and you'll get intelligence when we don't really know what it is underneath that makes us creative and allows us to come up with good explanations.

People talk a lot about GPT-3, the text matching engine that AI put out, which is a very impressive piece of software. But they say, "Hey, I can use GPT-3 to generate great tweets." Well, that's because, first, as a human, you're selecting which tweets out of all the garbage that it generates are good. Second, it's using some combination of plagiarism and synonym matching and so on to come up with plausible sounding stuff.

But the easiest way to see that what's generating doesn't actually make any sense is just asking it to follow a question. Take a GPT-3 generated output and ask it why—why is that the case? Or make a prediction based on that and watch it completely fall apart because there's no underlying explanation. It's parroting; it's a brilliant Bayesian reasoning. It's reading from what it already sees out there, generated by humans on the web.

But it doesn't have an underlying model of reality that can explain the scene in terms of the unseen. I think that's critical. That is what humans do uniquely— that no other creature, no other computer, no other intelligence, biological or artificial, that we have ever encountered does. And not only do we do it uniquely, but if we were to meet an alien species that also had the power to generate these good explanations, there is no explanation that they could generate that we could not understand.

We are maximally capable of understanding. There is no concept out there that is possible in this physical reality that a human being, given sufficient time, resources, and education, could not understand.

Time on a number line example

He Grew Up on the Streets, Now He's Making Them a Better Place | Short Film Showcase

View All

Interest groups and lobbying | Political participation | US government and civics | Khan Academy

Let’s discuss interest groups. As you can see here, it is one of the three parts of the iron triangle that we first studied when we looked at the bureaucracy in the executive branch. The whole point of the iron triangle is to show how these different part…

How to Identify a Bull Shark | Raging Bull Shark

NARRATOR: Researchers have confirmed 100 bull shark attacks. But the real number may be much higher because the bull shark is so difficult to identify. When you first look at a bull shark, it doesn’t really jump out at you that it’s a bull shark. It just …

Snatoms! The Magnetic Molecular Modeling Kit

Imagine this is a hydrogen atom. This is another one, and then this is H2, the hydrogen molecule. You can break it apart, add an oxygen atom and make H2O—water. A carbon with four hydrogens, that is methane. And you can actually combine it with water to m…

How to Focus Intensely

In a world that is growing in distraction, the ability to focus is becoming increasingly rare. It’s a skill that, simultaneously, is becoming increasingly valuable. Its demand is rising while its supply is decreasing, to put it in economic terms. In this …

First Look at Jane | National Geographic

Louis Leakey sent me to Gombe because he believed that an understanding of chimpanzees in the wild would help him to better guess how our Stone-Age ancestors may have behaved. It had long been thought that we were the only creatures on earth that used and…

Debunking 3 myths about air pollution | Nat Geo Explores

(upbeat music) - [Narrator] Ever think of how many breaths of air you take in a day? It’s a lot, like 20 thousand, give or take a few. All day, all night, our bodies are at work bringing in the good (bell dings) and kicking out the bad (buzzer sounds). Bu…

More Compute Power Doesn’t Produce AGI

More Articles