The Skip Podcast
00:03:16 · Dianne Penn
[00:03:16] S0: Thanks again, Nikhyl for having me. I'm Dianne. I lead our product team focused around research and labs here at Anthropic. This means one, working as product managers for our core foundation models. So everything from Claude two to now, the Claude 5.5 family that we just did our latest release, on Wednesday, to new model capabilities. So things like tool use, computer use, vision, reasoning, as well as on the lab side. This included initial incubations and, scaling of products like Claude Code, MCP Skills, and Claude Design. I think, Nikhyl, on your question around career, I definitely feel like I've had like a eclectic or pretty diverse set of experiences. I actually started out working as a bond trader in New York. Then started working in product management, at Credit Karma. First actually on the like strategy side before doing more product work. Then, as a PM on the Alexa AI team focused around model development. And that was how I really fell in love with AI and now working at Anthropic. I think something that really is kind of non obvious in a lot of those opportunities and roles is that I really just actually love figuring out the learning more than almost applying the specific like learning or field sometimes. And so I feel like there's people who really just like are really good at applying the same skill set over, over and over. And what I found was when I was in trading, the first like one to two years, like I got to like learn a lot about myself, how decisions get made, how to like stick with decisions, even if they don't work out perfectly. But it didn't feel really after a while, like rewarding in the sense of like, how was I contributing back? How was I making like new things like that are more creative or impactful? And so like, that's kind of how I found myself working in product management. It very much feels like there is users who have potentially problems or opportunities. We don't know how to frame that. And like initially, that's what attracted me to strategy. And in the strategy world, the hard part was actually, how do you actually make a strategy work well? And I think that was like one of the big learnings for me at both Credit Karma and Amazon and Anthropic, that the strategy is really important.
00:05:51 · Dianne Penn
We do need to have like a path of what we're trying to get to, but it's so much the execution and the details and the learning. And so in particular, I feel like between Credit Karma, Amazon, and Anthropic, like honing in on the like execution, the operational excellence, thinking about the every last detail has actually been one of the big things that I feel like has been both hard sometimes because there's so many details, like more details all the time, especially when you're building like larger systems and products, but have been really fun because the users are actually changing and are so different. I think the way that I kind of actually ended up, joining Amazon and Alexa was I feel like there was an era when folks talked about software was eating the world. And at Credit Karma, as you know, like the recommendations systems were very much becoming so core to how a user really interacted with their financial progress, their financial health. And it was very hard at that time for, I think most companies to actually be able to do foundational research or like R and D in terms of techniques. And I had a friend serendipitously reach out about this like first role as a PM with this, within this large, like, AI org at Alexa. And it didn't really feel obvious because Credit Karma was growing. We were just getting, I think, acquired by Intuit. But I just felt like that's opportunities to be like the first person working on this area and really reaccelerating my learning didn't come around very often. And even though it was... It felt like at the time, you know, LLMs wasn't even a term, we were starting to explore that this technology of like transformers and how you might replace like classic, like classifier systems for like natural language processing. And I think I just kind of fell in love with the root of the problem and the fact that like, it didn't mean we had the final solution or the end state for users and that this solution space was growing. And so I really felt like at the time it didn't feel obvious to go to Alexa, but looking back, I felt like that was one of those decisions that really helped my career. And then ultimately how I found myself in Anthropic as well.
00:08:22 · Nikhyl Singhal
[00:08:22] S1: I think that's really insightful. I think one thing that I think we all can agree on is that often careers make sense only in the rearview mirror. So like we look back and we're like, oh, let me tell you exactly what happened. And then I, we, I did this thing and it failed and it was so helpful, but at the time, you know, these are like earth crashing kind of decisions. I think what you're saying is you look for huge surface areas with high degrees of ambiguity, where the problem was sort of 1% solved versus say 95 solved. And yet those companies and opportunities often are theoretically high risk, not well known. And your point is, Hey, that's what kind of attracted me because as I was cracking the code, we were always solving new problems. And then the second thing I'm hearing is the detail point. It's like, well, the way to solve that wasn't just high level. It was a combination of being in the detail and being very strategic. So knowing the strategy, knowing there was a lot of directions you could go in and that you had to play a hands on role that felt like this combination of skill. Is that, is that a fair characterization?
00:09:31 · Dianne Penn
[00:09:31] S0: I think you're exactly right. Like in hindsight, it seems like all the different pieces hang together in the moment. I don't feel like I had a grand plan of this must be the way my career has to like work out. I think I very much try to look for what are things that really interest me if I had more time to look into. And then in particular, in the case of like Alexa and working on the AI team, I actually had friends who proactively were like, why would you leave Credit Karma? It's a startup. It was just getting acquired. You guys are going to grow to go to a large company. I actually ended up going almost like against some of my friend's recommendations because I felt like I personally wanted to learn how R and D, AI R and D worked. And so it was very much following kind of my own curiosity. And then I think on the second point around being in the details, this is something I think has been like really humbling to learn throughout my career, because it was kind of going from strategy into product management. And sometimes it feels like if you have the right strategy, you're going to have just exactly the right outcomes. And it just really made me appreciate back when, you know, we were working together, how hard every product decision actually was, not just the the initial plan, but the actual details to get to something valuable. And that is like 99% of what makes a great strategy actually work, that the details got worked out correctly.
00:11:03 · Nikhyl Singhal
[00:11:03] S1: Well, one thing that I think it just may... Just a quick pause, mean, I think there are a lot of people that are listening to this that do have roles that are more strategy oriented. And I would say Credit Karma had, as a company, unusually strong strategy, both personnel as well as there's a lot of thoughtful areas because there was a lot of surface area. How are you going to take on, you know, the auto industry or mortgages and other areas? I think the point that you're making though is very, very helpful. When you're a strategist today and you have strength and there's obviously huge IQ and opportunities to understand the market. But your point was that when you turned it into the execution and coupled it with that strategy, that's when you knew that it's more than just a plan. That's when the real learning took place. And that seemed like a big career unlock that frankly, lot of strategists don't do or not capable of doing. So one, it's an element of skill, but also it's an element of orientation, not to build bigger strategies at bigger companies, but actually go into the detail. And I think that's what I'm hearing as this sort of suggestion or career guidance for strategists.
00:12:11 · Dianne Penn
[00:12:11] S0: I think that's right for me personally. I think it's very much what is like the potentially the goal or what gives you your center of joy. And for me, like I always felt like when I was focusing more on strategy, there was something missing in terms of why I couldn't get to the right outcome necessarily. Or like, why doesn't the initial perfect plan work out perfectly? Right? Like why does the initial product pitch work? And what I have learned, I think both at Amazon and also now at Anthropic is that the 99% is after the write the initial plan. And it is just being in the details, figuring out every user flow, figuring out what people's feedback are from the model side, understanding, reading like model training transcripts, getting closer to details on user feedback. Like it's all of the other pieces that actually makes something successful, not just only the initial idea, which is important. But I think it just depends on kind of what is the impact you want to have on the world and like, what is your center of joy?