Pete Huang

This week, Leopold released a series of essays that has generated a ton of buzz in the AI and broader tech community.

It has to do with the bigger question about AI.

Just how powerful is AI going to get and what are we going to do about it?

Leopold's general argument is this, that we are on our way to super intelligent AI, we are not on track to do this safely, we need the government involved, and China is going to be out to get us on this.

Oh, and by the way, Leopold is also starting a hedge fund to make investments based on the stuff he sees relating to AI.

That's sort of a separate thing.

Now, if you look at his actual writing, it's a pretty dense set of essays.

So today I want to walk you through what he's saying and the sort of back and forth that has happened since then.

First, let's talk about AI.

AI has always been a moving target of a definition.

I mean, at some point, even the autopilot systems on planes, that was called AI.

Even the programs that play chess against you on a computer, those were called AI and

All of this AI up until the last few years is very narrow.

Like that AI program that helps you fly a plane cannot also play chess.

They're all super specialized to what they were trained to do.

Recently, we're getting to AI that is more general in nature, right?

ChatGPT can think about marking strategy, but it can also tell you about repairing things in your home.

What we're improving on now is the reasoning capability.

We're currently on GPT-4, but a few years ago, the best thing available was GPT-2, and GPT-2 can't really compare to GPT-4.

It's just not nearly as capable at thinking about things and coming up with a well-reasoned answer.