Lenny's Podcast: Product | Career | Growth
00:44:08 · Ian Silber
[00:44:08] S1: Yeah. And I think we're seeing what's been really fun design space for us is there's two things we think about. One is how can we help you express your intent better? We don't want you to have to think about being a perfect prompt engineer, but how do we even, one, teach you that, oh, this is something that the model can help me with. And then how do we make sure that we best understand what you want? And so we recently shipped some features where if you ask sort of an open-ended question, it can follow up with some follow-up questions that you can just tap through and answer versus having to type everything out. Or another good example is, so then once you actually ask it to do something, the chat transcript doesn't need to be purely text. We've seen these now evolve so far from what it was when ChatGPT just launched. It's no longer just a pure chatbot. It actually has many features within that. So a good example is what we call writing blocks. You ask it to write you a letter or an email and it now contains that email in sort of a block that you can directly manipulate and you can type with it and you can brainstorm and you can edit little bits and parts of it. You can copy just that. And that has made that interaction much better. And we're seeing the same thing with image generation and creating special tools for that kind of mode. And I think that's where UI is going to start to really continue to evolve is how do we make... If you come at this and you're a designer, maybe ChatGPT should actually really look different or have different affordances or you're a data scientist or you're using this to run your home, your personal life. Maybe it should really work or look differently. So understanding the context and giving you different affordances based on that is something that we're thinking a lot about.
00:46:08 · Lenny Rachitsky
[00:46:08] S0: In terms of the vision and future of where all this goes, I know this is always evolving, but is there kind of a sense of what the future of ChatGPT Codex is? I know there's this idea of a super app. You talked about voice, this personal assistant idea. What's a way you think about where this goes long-term?
00:46:25 · Ian Silber
[00:46:25] S1: I think that it will start to become more and more adaptive to all of the context of your life. I think proactivity is a big untapped part of the experience that we're really excited to think about. We're starting to see this with ChatGPT Work now where it actually has pretty good suggestions based on your calendar or your Slack or these different sort of inputs and external context. I think that's going to become more and more important or shift the way you use it. I think voice, we just had a step change in the model that just feels much more natural. And again, I can only imagine how that's going to feel in two, three years as it just gets better and better. I do love interacting with ChatGPT, especially Voice In. I want to just really quickly say something. And then I think its ability to go not just from asking questions, but actually doing things for you, that's all going to change. And hopefully all of this gets simpler as we do it too. When we think of the super app, I think eventually we really think of the idea of, especially for our broad consumer base, we do think that there should be this one universal input that you can ask that knows when it needs to answer really quickly or should be chatty or whether it should go and go off and do work for you. And so that's a big excitement that we can get to is how can we just continue to simplify the experience such that you shouldn't have to think about how this works. You shouldn't have to think about what model you're on or what sort of speed you're using. That goes back to the balancing the different kind of personas and user needs. Of course we want to expose that for the right people, but then we want to make sure that we can get a default experience that works great for most people. But yeah, I think it's going to be really proactive. I do think that we're going to get much richer input through voice and richer visual outputs that are more interactive. I think that you'll start to build more durable workflows with ChatGPT that you can use over and over. Right now, a lot of times it feels like you're coming and starting from scratch each time. And we want to really make this kind of the tool that you can use to help you with whatever sort of task you have.
00:49:04 · Lenny Rachitsky
[00:49:04] S0: This all sounds wonderful. I can't wait. Just zooming out a little bit, and then I want to talk a little bit to designers again. But you're in this epic role right now. I imagine there's a lot of surprises as you entered OpenAI and took on this role. What's something you wish you'd known? What's something you wish you could whisper into Ian of day one to give you some advice?
00:49:26 · Ian Silber
[00:49:26] S1: I mean, so many things. I think the first one is recognizing that we're really early and instead... I actually was pretty intimidated coming to OpenAI, if I'm being honest. I remember interviewing. Oh no, I got to... GPT-4 had recently come out and GPT. I was like, "What is all this stuff?" I was using it and I was using the stuff, but I didn't feel like I had a true understanding of it. And to be honest, I still have so much to learn. So that idea of you are early, we're figuring this out. That I think I would want to hear. And the other thing that I found, and for me, this is more of a personal thing, it's less about AI or anything, but I stepped into a really different role most of my career. I definitely managed before, but never at a scale that's growing so fast and then at a company that is so different that... the role was just every day was something new. And for a long time I felt like I'm kind of bad at this or I don't know what I'm doing. And then I went and talked to friends that had led really large design teams or were in a similar transition as me and they all felt exactly the same or the problems that I had were not unique. And so this is, I guess, more personal advice that it's like, I think at any stage of your career, there's that feeling of imposter syndrome and what am I doing? I don't know how to do this. Finding people you can talk to that are going through the similar things, you'll find that there's usually a lot of shared challenges and struggles and having that, whether it's a mentor or just a buddy that is impartial, out of it, was really helpful to me. And finding that community I think has been helpful. So yeah, so that and then understanding that this is all so new and evolving really quickly.
00:51:37 · Lenny Rachitsky, Ian Silber
[00:51:37] S0: That is extremely helpful to hear. Elena Verna had this really great post. I don't know if you know her. She's head of growth at Lovable right now. She had this great post called Stop the AI Confidence Theater, where she just kind of is like all these people just keep posting all these things they've done with AI and all the ways it's changed their life and it's just the killing it, figure it all out. And really it's very few people, very few of these things are actually that great and working and it's just like, stop. Stop pretending like this is actually, you figured it all out. [00:52:05] S1: I tghink it's much better to be humble about it and be like, wow, yeah, I don't know. This changing quickly and I'm just going to embrace it and see where it goes. [00:52:13] S0: Yeah. And that connects to what you said was a core component of people that are doing well right now. People that are curious and just kind of poking around and trying to try these things and try them again because maybe it didn't work now. It didn't work before, now it works now. [00:52:25] S1: 100%. Yeah. Every model, every new feature, it's like the people that are poking around at those edges I've found are just having such a blast doing it. [00:52:39] S0: And this connects to something else Kevin Weil, former CPO at OpenAI said on the podcast, which has always stuck with me, which is the model we have today is the worst the model will ever be. [00:52:48] S1: Exactly. I mean that goes back to when we... Yeah, I mean a new model comes out, what, every week? [00:52:59] S0: Every day.