The Best Way to Achieve AGI Is to Invent It

**Pedro Domingos** (0:00)
Einstein took 10 years to come up with general activity, and that was just one equation. So, what makes you think we're going to solve AI in six months? I really do think we need fundamental new ideas. And what I see, unfortunately, is that there's more research than ever, but it's actually less diverse than ever, which is very unfortunate. There is a sprint, but it's a sprint to a local optimum. And there's nothing wrong with sprinting to the local optimum. So, like, there's a lot of mileage to be had, for example, in making transformers faster and more efficient. All of that is good. But at the end of the day, that is not going to get us to human level intelligence. I don't think, having said that, even if that's all we do, is make transformers better, that alone is already going to have an incredible impact. It's just that AI is a technology of an impact that people are not used to. So even just getting 10% to AGI is going to change the world dramatically.

**SPEAKER_2** (0:51)
Welcome again to the A16z AI Podcast. I'm Derek. You just hit play at another great discussion. This one featuring A16z General partner Martin Casado in discussion with noted machine learning researcher and now multiple times published author Pedro Domingos. Pedro is a professor emeritus at the University of Washington and has been involved in the world of machine learning since the 1990s. He's also the author of the popular 2015 book The Master Algorithm and recently published his first novel, 2040, A Silicon Valley Satire. And while the discussion touches on those books and the ideas of them, the focus really is on whether our current path of AI research is the right one to get us to the mountaintop, whether that's AGI, super intelligence, or whatever anyone thinks is the ultimate goal. Among other topics, Martín and Pedro assess and critique the continued applicability of scaling laws, the wisdom of building out massive AI data centers, and the importance of LLMs and the transformer architecture on the quest for the one true AI, or to use Pedro's term, the master algorithm. It's an insightful conversation, which is better heard from them than recapped by me. So here goes, starting with Pedro giving a very abbreviated version of his lengthy resume.
As a reminder, please note that the content here is for informational purposes only, should not be taken as legal, business, tax, or investment advice, or be used to evaluate any investment or security, and is not directed at any investors or potential investors in any a16z fund. For more details, please see a16z.com/disclosures.

**Pedro Domingos** (2:25)
I've been a machine learning researcher for 30 years since since my PhD in the 90s, so I'm kind of an AI old timer. I'm a professor at the University of Washington. I'm probably best known as the author of The Master Algorithm, which was a book popularizing machine learning that went bestseller. Most recently, I've written another book on, it's a satire of AI called 24A Silicon Valley Satire. And I'm continuing to do research and write and generally try to educate the public about AI.

**Martin Casado** (2:50)
So I noticed that you're emeritus. Can you describe what that means regarding your day-to-day?

**Pedro Domingos** (2:56)
Emeritus like, you know, who the Perla famous actually Turing Award winner said, is beautiful because you do everything you want and nothing that you don't.

**Martin Casado** (3:04)
That's cool. So what do you want to do, I guess?

**Pedro Domingos** (3:06)
Exactly. You know, my joke to people is that back when I was a professor, doing research was my hobby and now I can do it full time.

**Martin Casado** (3:12)
Nice. So you're doing research.

**Pedro Domingos** (3:14)
Oh, yeah, definitely.

**Martin Casado** (3:15)
What are you working on?

**Pedro Domingos** (3:16)
Solving AI, of course.

**Martin Casado** (3:19)
Is that going to come out when we have this conversation or is it a secret?

**Pedro Domingos** (3:23)
Let me just put it this way. I've solved it. I just haven't told you what it is yet.

**Martin Casado** (3:28)
That's right. And you're not going to tell us until?

**Pedro Domingos** (3:29)
I can't yet. I can't. I need to patent it. And you know, there's some legwork to do first. And prove that it works. But there's a manner of details.

**Martin Casado** (3:38)
So you've solved it. Okay. Everybody, you've heard it here. Pedro Domingos has solved AI. And by the way, solve AI, you mean like AGI?

**Pedro Domingos** (3:44)
No, I don't just mean AGI, I mean super intelligence.

**Martin Casado** (3:47)
Wow.

**Pedro Domingos** (3:48)
I see this all the time. And I love you because people don't believe me.