Why PMs are suddenly valuable again | Tulsee Doshi (Head of Product, Gemini)

SPEAKER_01: The model isn't predictable. SPEAKER_01: It's kind of like Snape's potions class. SPEAKER_01: You can put all the ingredients in, SPEAKER_01: but you also have to expect that there will be a little bit of magic, SPEAKER_01: and that magic is going to be surprising. SPEAKER_00: As you're hiring PMs these days, is there anything you look for more and more? SPEAKER_01: You're looking for PMs who understand the vibes of our models or other models. SPEAKER_01: I want PMs who have an opinion, who are playing with these models every day. SPEAKER_01: You ideally want to stress test the model on things that you are excited about. SPEAKER_00: A lot of people feel that Google lags behind the top two labs right now. SPEAKER_01: I actually think we do want to win the frontier conversation. SPEAKER_01: I don't think we should say that Google is not participating in the frontier. SPEAKER_00: You have this framework around how far ahead to build. SPEAKER_00: You call it a 60-20-20 rule. SPEAKER_01: Traditional startup wisdom. SPEAKER_01: You're just like, build an MVP that works. SPEAKER_01: But I think you want to actually build features that aren't necessarily always working today. SPEAKER_00: It feels like every AI assistant launching these days is taking all this juicy SPEAKER_00: Google data, my Gmail, Google Calendar, my Docs. SPEAKER_00: I'm curious just why Google hasn't launched something awesome yet. SPEAKER_01: The team is working on some amazing things, SPEAKER_01: but I think you are right that there is more work we need to do here on this front. SPEAKER_00: Tulsi, thank you so much for being here. SPEAKER_00: Welcome to the podcast. SPEAKER_01: Thanks. I'm excited to be here. SPEAKER_00: You are head of product for Google's Frontier models, which includes Gemini, which people know. SPEAKER_00: But there's also a bunch of other models that people may not be familiar with. SPEAKER_00: Share the other models that are under your umbrella. SPEAKER_01: Yeah, so we have a set of models we kind of call our GenMedia models. SPEAKER_01: These are our models that focus on video editing and generation, SPEAKER_01: image editing and generation, if you've used Nano Banana, Lyria, our music model. SPEAKER_01: We also have Gemma, our suite of open models. SPEAKER_01: And then one set of models that actually falls under our Gemini umbrella, SPEAKER_01: but I feel like we don't talk about as much as our audio models. SPEAKER_01: So dialogue, audio understanding, text to speech, translation. SPEAKER_01: All a part of the portfolio of how we build FredShare models. SPEAKER_00: Okay, so I want to spend a little time on models. SPEAKER_00: Everyone sees models launch. SPEAKER_00: There's like a model launch every week, it feels like. SPEAKER_00: You have a really unique lens in being able to see these models from research SPEAKER_00: to pre-training, post-training launch. SPEAKER_00: What's something that you think people maybe misunderstand SPEAKER_00: about how models are created or something that may surprise people SPEAKER_00: …