The Alarming Truth About Superintelligent AI: Roman Yampolsky on Human Survival (Replay)

Roman Yampolsky (0:01)

Queen Carvania stood haloed by the morning sun. An army hung on her every word.

Tom Bilyeu (0:08)

My champions, I have sold my chariot on Carvana to a lovely SUV, an inexplicably queenly offer. They're even coming to the castle to collect it. Tonight, we feast.

An offer you can feast on. Sell your car today on Carvana. Pick up fees may apply.

Tom Bilyeu (0:30)

Let's talk about the moment your side hustle becomes a real business. Now, I see way too many founders stay stuck running their business like it's a hobby. And then they wonder why nothing ever feels real. It doesn't feel real because it's not. I've helped thousands of entrepreneurs build real businesses through Impact Theory University. And the first step in the path to building a legit business is registering it. And Tailor Brands will help you do just that in a few clicks. An LLC, a Limited Liability Company, is an official business structure. It's inexpensive to form and very simple to maintain. It's how you open a business bank account. It's how customers know they're dealing with a real company. And Tailor Brands doesn't stop at the filing. Their dashboard tracks what comes next based on what your business actually needs. So EIN number, registered agent, licenses and permits, bookkeeping, business banking, I tell every founder the same thing. Just make it official. Click the link in the show notes to check out Tailor Brands and get started on your business today. This is a paid advertisement. In 2023, nearly half of all AI researchers said advanced AI carries at least a 10% chance of causing human extinction. And yet we're speeding up, not slowing down. My guest today, Dr. Roman Yampolsky, is one of the leading voices in AI safety. And when I asked him for the odds that super intelligence wipes out humanity, he said it's high.

Once AI becomes smarter than humans in every domain, we will not be able to control it. In today's episode, we talk about the shocking timeline AGI is on, why super intelligence may be much closer than people think, and why the survival of our species could come down to decisions being made right now.

If you want to understand the most important technological threat in human history, as well as our biggest opportunity, this is the one episode you can not miss. So without further ado, I bring you Dr. Roman Yampolsky.

Where's ChatGBT at right now? Do you consider ChatGBT to be artificial general intelligence? I doubt you'd call it super intelligence, but would you classify it as that or do you still think we're a ways away from something that would qualify?

Roman Yampolsky (2:52)

So that's a great question. If you ask someone maybe 20 years ago and told them about the systems we have today, they would probably think we have full AGI.

We probably don't have complete generality. We have it across many domains, but there are still things it's not very good at. It doesn't have permanent memory. It doesn't have ability to learn additional things well after it's already been pre-trained and deployed. It can do a certain degree of learning, but it's still limited. It doesn't have same capabilities as humans do throughout lifetimes. But we're getting closer and closer to where those gaps are closed, and it's starting to be productive in domains which are really interesting and important. Science, math, engineering, where it starts to make novel contributions, and now top scholars are relying more and more on it in their research. I think we're getting close to full-blown AGI. Maybe we're at like 50 percent, but it's hard to judge for sure. Just how many different sub-domains exist is the deciding factor.

Tom Bilyeu (3:59)

Okay. So one idea that you put forward that's very interesting is like, hey, I'm an engineer. I love AI, but I would like you to keep it very narrow, please. What are the things about general AI that become problematic, that aren't problematic in narrow AI?

Roman Yampolsky (4:16)

So a whole bunch of them. One is testing. How do you test a system capable of performing in every domain?

There is no edge cases. Typically, if I'm developing something narrow, very narrow system, I'm just playing tic-tac-toe. I can test if it's making the legal move. I can test zero, I can test 100, I can test all these weird special cases and know if it's behaving as expected. With generality, it's capable of creative output in many domains. I don't know what to expect. I don't know what the right answers are. I don't know how to test it. I can test it for a specific thing. If I find a bug, I fix it. I can tell you, I found a problem and it's been resolved, but I cannot guarantee that there are no problems remaining.