Speaker unverified: belief

Transcript context

…I think we’re not saying one actually obvious thing that we’re not realizing, that’s a gigantic thing that’s hard to measure, which is making all of human knowledge accessible… …To the entire world. One of the things that I think is hard to articulate, but there’s just a huge difference between Google Search and an LLM. I feel like I can basically ask an LLM anything and get an answer, and it’s doing less and less hallucination. And that means understanding my own life, figuring out a career trajectory, figuring out how to solve the problems all around me, learning about anything through human history. I feel like nobody’s really talking about that because they just immediately take it for granted that it’s awesome. That’s why everybody’s using it—it’s because you get answers for stuff, and think about the impact of that across time. This is not just in the United States; this is all across the world. Kids throughout the world being able to learn these ideas—the impact that has across time is probably where the real GDP growth will be. It won’t be like a leap. It’ll be that that’s how we get to Mars, that’s how we build these things, that’s how we have a million new OpenAIs, all the kind of innovation that happens from there. And that’s just this quiet force that permeates everything, right? Human knowledge. I do agree with you, and in a sense it makes knowledge more accessible, but it also depends on what the topic is. For something like math, you can ask it questions and it answers, but if you want to learn a topic from scratch—we talked about this earlier—I think the sweet spot is still math textbooks where someone laid it out linearly. That is a proven strategy to learn a topic, and it makes sense if you start from zero to get information-dense text to soak it up, but then you use the LLM to make infinite exercises. If you have problems in a certain area or have questions about things you are uncertain about, you ask it to generate example problems, you solve them, and then maybe you need more background knowledge and you ask it to generate that. But it won’t give you anything that is not in the textbook. It’s just packaging it differently, if that makes sense. But then there are things where it also adds value in a more timely sense, where there is no good alternative besides a human doing it on the fly. For example, if you’re planning to go to Disneyland and you try to figure out which tickets to buy for which park when, well, there is no textbook on that. There is no information-dense resource on that. There’s only the sparse internet, and then there is a lot of value in the LLM. You just ask it. You have the constraints on traveling on these specific days, you want to go to certain places, and you ask it to figure out what you need, when and from where… …What it costs and stuff like that. It is a very customized, on-the-fly package. Personalization is essentially like— …pulling information from the sparse internet, the non-information-dense thing where there’s no better version that exists. You make it from scratch almost. And if it does exist, it’s full of—speaking of Disney World—ad slop. Like any city in the world, if you ask “what are the top 10 things to do?” An LLM is just way better to ask… …Than anything on the internet.…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.