Ethan Smith: belief

Transcript context

…I'm afraid of this world where everything is trained on AI, and AI is trained on AI and generating AI, and just like nothing is trusted. And I love how it's interesting just how much of these incentives are driving this. If ChatGPT was finding this valuable, this is what people do and then just goes off the rails. So there's just some team there that is keeping this from happening. How do you think this evolves? If you were them, what would you do over the next few years to keep things high quality and not drive these perverse incentives? So I would identify what the perverse incentives might be, and AI-generated content is one of them. The second thing is I think that LLMs and search are going to converge. And so you're seeing that with Google Search where they're having LLM, AI Overviews. You're seeing that with LLMs where they're incorporating maps and shopping carousels, and it's converging on search. I think it'll converge on a single experience, so that's the first thing. Figure out what 2007 Ethan would do not to create spam and make sure that he doesn't do that, like AI-generated content or it's great content. That'd the second thing. And the third thing is there's all these other interesting features, use cases that LLMs can be great for. So LLMs could be great for remembering everything that you've ever asked. It could be good for personalizing stuff specifically to Lenny. One interesting use case that I think will eventually come would be, I say, "Plan a trip to San Francisco," and decisions are made for you without any intervention. I have this wonderful EA named Jen. And I say, "Jen, I'm going to Miami. Please, just do everything for me," and she does everything for me. She knows me, she knows my preferences, she knows that I want a ocean view and I want a restaurant with music. She does all of that and I don't have to intervene. AI can essentially do that eventually, and that would do that because it would deeply understand you. It would remember everything about you. It would have context, it would have a reasoning, and then it would be able to make all these decisions without your intervention, which would be autonomous agents. So I think that that's also another very interesting place for someone like me to optimize for as well. Yeah. I was just going to say, just imagine not even being told this is what you're choosing. Like, "Oh, and go check out, subscribe to the best newsletter out there." And if you're out there, the good things will happen. Wow, what a wild world. Is there anything else that we haven't covered that you think would be helpful to folks that are trying to get better at this stuff? Try to take the first steps down this road of AEO?…

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