How AI Is Actually Changing Product and Design

In this episode, I speak with Aoni Wang, product designer at Harvey, and Yiyang Hibner (Silicon & Spice with Yiyang), product manager at Airbnb, about how AI is changing the way they work day-to-day. This conversation is a bit more light-hearted than the usual conversations on AI Proem, but I think it is equally insightful because these two professionals work in the heart of Silicon Valley, and they give us a glimpse into how people are really adopting AI.

Both are already using AI as part of their regular workflows, from thinking through ideas and understanding complex systems to research, writing and building interactive prototypes.

We explore the tension between making execution faster and deciding what is actually worth building. As AI makes it easier to produce prototypes and generate work, Aoni and Yiyang discuss the risk of confusing speed of output with speed of progress. More work can be produced, but that also creates a greater need for filtering, judgment, and accountability. They also discuss why AI can be too agreeable, and why users still need to challenge its outputs, verify information and bring their own expertise to the process.

The conversation then looks at how AI could change the boundaries between product, design and engineering. Then they discuss the growing importance of skills such as systems thinking, prioritization, communication, taste and the ability to define what a good product or experience should look like. For Aoni, this includes thinking about visual language and coherent systems rather than simply producing individual screens or prototypes.

Finally, we get into the more practical questions around working with AI, including how teams maintain shared context when everyone has their own AI tools, how people decide which products are actually useful, and where human judgment remains difficult to replace. The conversation offers a view of AI adoption from inside the workflow rather than from the perspective of the technology itself, raising a broader question about the future of work: as more of the execution becomes easier, how will people decide where to focus their time, attention and expertise?

The AI Proem Podcast is part of the AI Proem newsletter, which has ~13k followers globally. To learn more about China AI, the business of AI, and how AI is impacting society, please check out the newsletter here and more insightful conversations here.

Chapters

* 00:00 The Evolution of Design Work with AI

* 05:14 AI’s Impact on Product Management

* 07:59 Efficiency vs. Overwhelm in AI Utilization

* 11:26 Navigating AI Slop and Quality Control

* 14:19 The Future of Roles in Design and Product Management

* 17:59 The Negative Impacts of AI on Work

* 21:22 The Changing Landscape of Collaboration with AI

* 28:03 Shifting Workflows: AI vs. Traditional Software

* 30:17 Creative Tools: Evaluating Effectiveness and Usability

* 33:38 The Importance of Human Skills in an AI-Driven World

* 36:49 AI’s Role in Decision Making and Workflow Management

* 40:57 Personal Use Cases: AI in Everyday Life

* 46:28 Differentiated Views: Perspectives on Work Culture

Transcript (AI-generated, for reference only)

Grace Shao (00:00)

Hi guys, thank you so much for joining today. Really really excited to have you on.

Yiyang Hibner (00:03)

Thanks for having us.

Aoni Wang (00:04)

Thanks for having me, Grace.

Grace Shao (00:04)

Yeah, so before getting started, I know you guys both have some disclaimers from the company you have to say. So go ahead, do your thing.

Yiyang Hibner (00:12)

All right, I’ll go first. I’m Yiyang, I’m based in Bay Area and currently I work at Airbnb. And opinions are my own. So you know, Airbnb is one of the companies I’ve been working at in the past few years. There are a lot of other tech companies, so like some experiences are not unique to the current company I’m at. Yeah.

Aoni Wang (00:30)

Hi, I’m Aoni. Similar to Yiyang, I’m based in the SF Bay. I now work at Harvey as a product designer. But today all views are my own and also coming from my own experiences.

Grace Shao (00:41)

Yeah, perfect. So once we get that out of the way, I would love to hear about the your own experiences a bit more. Could you guys just give us a little bit of color on what is it that you do? You know, what brought you here today? You know, compare the work that you are doing today to maybe even two years ago prior to the proliferation of like AI tools and, you know, AI agents.

Yiyang Hibner (01:03)

I’ll let Aoni go first.

Aoni Wang (01:05)

Yeah, yeah, sure, sure. Yeah, I could go first, yes. So yeah, I’m a product designer, and usually to people that means I’m designing software interfaces. So whether that is desktop software or mobile apps, and I would be designing the interface that people see, the interaction, collaborating very closely with product managers and engineers most of the time. For me, I have about 10 years of experience in this field working across different tech companies. Mostly working on mobile consumer type of products. And to us designers or product designers, Figma is our best friend. So if I were to think about two years ago, I’d be spending a lot of time in Figma. A lot of my design work actually happened in Figma. For example, the ideation, sketching out ideas before it actually becomes finalized or pixel perfect, sketching up mock-ups, interactions, flows and So from like rough sketches all the way to that polished mock-up that I can hand off to my team to build, all of that would happen in Figma. And now I would say where I spend my time and how I design have changed the most. So I guess it’s probably true for everyone. I’m chatting with Claude or ChatGPT a lot every day. And even before getting to Figma, I would do a lot of design thinking, sketching with AI agents. And Figma is only a small part of my design journey now. Or depending on the work, like it’s not all I use to design anymore, especially with coding agents. Now it is much easier and much faster to prototype, meaning that I can build something that looks functional, feasible, and realistic, and that my team can play with it, I can use it to communicate my ideas. In the past, that is usually a much longer process.

Grace Shao (03:01)

Yeah, I definitely want to double click on the Figma moat later because I think that’s something we talked about offline as well. It’s quite fascinating how back then Figma’s moat, like, its value is also the shareability of it. But now, you know, I’m surprised that you’re moving away from Figma in some of your workflow. But let’s talk to that a bit later. Yiyang, tell us about your work. Like what is it that you do? How are you using AI already in your day to day work and how has that really impacted your workflow compared to maybe two years ago?

Yiyang Hibner (03:30)

Yeah. So I because I’m a product manager, so I interact with a lot of different cross-functional partners. Good engineers, designers, sometimes legal partners, or like product operations. So I think the biggest change is to have I think AI does make me more organized. Because as a PM you live and breathe in documents, in presentations, pitch decks, one-pager, strategy vision docs. So a lot times I you have to wonder, okay, what’s the best way to convey my I ideas in a very efficient way, tailored to my audiences. I think whether Aoni mentioned Claude Code or other tools, like Gemini, you know, company has different tools. I think these help me to find my voice, especially during kind of presentation and influence across, that really helps. Another big change is that Unlike designers, like you know, people really use Figma a lot. I think product managers and people have different like specialties. Sometimes some people prefer one way of working, other people prefer different ways of working. And for me, having AI to kind of be a thought partner, whether it’s through user research or competitor analysis or industry kind of news, is sometimes you know, you can only read so many newsletters and what’s the best essence of all or what’s happening in the payments world, it could happen pretty fast. I think AI is a really good thought partner. And then having the ability to build a quick interactive prototype that is kind of 10x my previous efficiency, right? You know, instead of having to describe the requirements or the desired user flow in the table or in the documents or in you know, these pitch decks. Now I could have an interactive prototype send a link right to my engineers and get some feedback and they can add comments. I think that really saves time back and forth.

Grace Shao (05:14)

That’s really interesting. So I’ve actually never worked obviously in a product manager role. Explain to me what exactly that you’re using it on. Cause like you were saying it increases efficiency, but what exactly? Like you’re saying it makes your a demo more interactive. What are you making?

Yiyang Hibner (05:29)

Yeah. Yeah. I think made made the demo more interactive and then kinda similar to let’s say Figma, you have a prototype. Now I have a prototype you can click through and then also kind of explain. One one example is let’s say if I have a screen to say, hey, here’s how it’s interactive. And then also I have kind of a little annotation on the side saying, Hey, why we’re doing this way. Or I could also have like option A versus option B sometimes when I share a prototype with others. So you understand, hey, that’s a thinking process behind. Why we prefer one way versus another. And then I would say as a product manager, the biggest question is like what to build and then who to build it for. Right. You gotta make sure kind of like I don’t want to say I’m the captain, but a lot of times the product manager helped to guide the team to say, hey, this is direction we should take. And then we partner with designer to say, hey, how can we really get into the weeds versus this is the other direction we want to take. Right. So having AI really help us to make sure is this the right way, how deep do we want to grow? Hopefully that made sense. Yeah.

Grace Shao (06:31)

Yeah. So you one thing you mentioned that was quite interesting was that you said the relationship between the developers and yourself have changed. Walk us through how so.

Yiyang Hibner (06:41)

I think with AI, a lot of like silly questions from PMs, you know, I don’t I don’t need to bother people. One one typical workflow, right? If I want to figure out how the current system works, especially when you first join a company, like, how does this system work? How does it interact with with another system? How does it interact with another system? Then usually I would have a one-on-one with the engineer and can you help me walk through the previous architecture? Especially I work in the payments world. It’s pretty complicated, right? You know, how one system does this thing, not that thing. But with AI, you know, I could ask AI to partner with me anytime, right? And then the questions now I ask engineers are the ones I really cannot figure out on myself, on my own, right? So that helps to reduce the barrier, right? I don’t feel intimidated by all the new system that I’m not familiar with. And now I can like, you know, chip away at it one at a time to say, okay, this is how this is supposed to work and then I understand it better. Or, you know, I could have a mind map. Or a lot of people use NotebookLM. You have a podcast. So I’m I could listen to let’s say a feature that’s released, you know, recently, they have the let’s say if they have a super long PRD, but now in the format that I enjoy consuming that I can really understand what other people are building as well.

Aoni Wang (07:59)

I think that’s definitely a great question because what really is common is that execution is cheap. Everyone can build things, like Yiyang mentioned, like AI has empowered everyone to figure things out on their own, to build a functional prototype on their own. And it’s it is very easy to get overwhelmed, but it also feels great to be empowered. To me, it’s more around where focus and energy has shifted. And to me it’s more now that everyone can come with a prototype or everyone can build a prototype, to me the design question is more how do I define a good experience? How do I define a cohesive experience? How do I empower the team to build these prototypes that are near ready rather than just an idea? And another thing is I think as a designer, besides making these judgment calls, taking in all the ideas from team, there’s also the work of setting up your own AI environment in some way. So setting up the context that your AI agent needs to help you make decisions at work, and also kind of like how you collaborate with AI, how you tell AI what is good design, what kind of style. Or visual language you work with. These are things that take effort as well, but also these are things that compound. So it is worth spending some effort there so that your work start to compound.

Yiyang Hibner (09:29)

Yeah, I think so your question is whether it makes us more efficient or adds more work. I do think both can be true at the same time. Reminds me of Shopify, I think Shopify CEO, Tobi, he was advocating for AI, you know, the headcount you had to justify why cannot use AI and stuff. Recently he said in a interview or podcast, he was saying the AI grenade is a real thing at Shopify now, that people just Have AI generated email responses or documents throw at each other that actually you have to figure out it’s more work to figure out, what’s the right thing? Is this AI slop or this actually has some true value to it? So I think it makes us more efficient because on the surface, people to handle more projects, you know, they can respond immediately if it’s their agent or themselves responding. But on the other hand, when you take out more work, maybe the quality of the work could be diminished diminished if you are not careful about it, right? Somebody said that in the past, before AI slop, there’s human slop, right? There are also humans who don’t care about the output of their work. They just like put something together or throw it to others. So now it just the magnifying glass made it the problem way worse. So I do think it’s more work to come through. You know, kind of like you had to filter out the dirt versus the pearl, right? But at the same time, it does make the tedious work a little more enjoyable, right? You know, like all the status updates, like tell me all the tickets assigned to me or my team, like I want to see, you know, instead of you have to go through all these little details, like you manually update the spreadsheet, that kind of thing. So that part makes us more efficient.

Grace Shao (11:10)

That’s a very sensible way of looking at it, I think. It like you’ combing through information that’s really tedious and repetitive is h easier, but actually we have to deal with a lot of the slop. So does that turn you guys off? Or does it does it cause frustration amongst your peers that there’s a lot of slop going around?

Yiyang Hibner (11:26)

I think sometimes you can tell people are using AI, you know, in the real world or at work, right? But at the same time, it’s almost like the default is everybody uses AI. But how you use it, do you own your output? That makes a difference, right? And then I try to say, hey, you know, every document, the presentation I produce, I have to fine-tune it, right? You know, at as if there’s intern representing you. Do you trust the intern’s work and you gotta double check the work? Or like you’re you’re like agent manager now, right? Can you trust what the agent has to say versus we just let the AI run wild in representing you? But if you care about your brand enough, you have to own the output.

Aoni Wang (12:05)

Yeah, I resonate yeah, I resonate

Grace Shao (12:05)

Yeah, I see you nodding a lot. Yeah.

Aoni Wang (12:07)

That s I res I resonate with that very strongly because I feel like even though AI is taking on a lot of the tedious work, taking on a lot of the execution work, essentially it is you or me. It is the person that is going to communicate about the work and going to stand behind the work and the intentionality has to come through the work. When you’re presenting something, like yiyang said, when you’re presenting something from AI in a room with people, they’re gonna ask, how did you make this decision? You have to be able to explain your thinking. Like, the output of AI is just a vehicle to show your thinking. You cannot give an answer like, I don’t know, Claude just created it for me. That is kind of AI slop.

Grace Shao (12:52)

Right. Do you think you guys in the Bay Area would be more AI-pilled but maybe peers working in other tech firms that’s not in the Bay Area?

Yiyang Hibner (13:02)

I do think so. Also the word AI-pilled, I first heard it from a different podcast. Somebody interviewed a head of Claude Code, you know, Cat Woo, I think she mentioned AI-pilled. When I first heard it, I said, Huh, that’s a interesting phrase. I never heard that word before. But I would think so because both Aoni and I we work I would call it kind of a almost a pretty pretty close to AI. You know, we’re not Anthropic, we’re OpenAI, you know, kind of at the hardcore model side, but we are consumer or enterprise-facing companies that are pretty AI driven, or at least you know, that definitely I would I would say the companies that have a lot of revolutionary changes because of AI. So I would consider us, you know, not trying to be humble or anything. I think we are pretty deep in the weeds. And also I think it’s a culture of Silicon Valley that. You know, people talked about here’s another podcast about this. So I try to it’s it’s very hard to keep up anyways, but I try to kind of listen to a few kind of podcasts, or interviews or read newsletters. It definitely feels like, you know, everybody’s drowning in all these changing landscape, but we’re we’re just crawling together. But I would consider us pretty like we’re pr pretty relatively caught up. You know, y in most of the cases, but yeah.

Grace Shao (14:19)

L let me follow up on this on the kind

Aoni Wang (14:20)

Yeah.

Grace Shao (14:20)

Of the frankly more negative side of it. Do you think then AI is becoming a disadvantage anyway? Like having A AI is a disadvantaging the role of the PM or the designer anyway? Do you do you ever think that, you know, AI, like you guys mentioned, requ even utilizing AI requires you to have a lot of strong judgment and taste and, you know, industry know-how, expertise? All these are buzzwords right now, right? That people are talking about. But at the same time, it took you guys more than a decade to get here. How do people obtain that, those, all those skills or tastes or expertise if they are relying on AI in the very beginning? This is a bit of a big open-ended question, but how do you guys view, I guess, the flip side, the negative side of AI’s impact?

Yiyang Hibner (15:06)

Aoni, do want to take that first? Since speaking of taste,

Aoni Wang (15:09)

No, yeah. I think Grace, what you’re trying to probe is kind of like how do people maybe starting out in the field today, like is there kind of like a fear of AI replacing their jobs? Like what’s the value of their job if a lot of it can be achieved via someone just telling AI what to do? Yeah.

Grace Shao (15:27)

I think it’s two questions. Sorry, I actually kind of I think I’m I’m muddled them up. One hand is, you know, we talked a lot about how AI has helped with your efficiency and helped a lot with your work. But are what are

Aoni Wang (15:37)

Mm-hmm. Mm-hmm.

Grace Shao (15:39)

Areas where you’re seeing AI is actually becoming negatively impacting, like negatively impacting your work workspace, workflow, or how your work was previously done? That’s the first half. The second half of the question is more similar to what Aoni, you’re you’re kind of alluding to, which is like, Okay, so if you’re a junior, you’re fresh out of school, how do you actually obtain the taste and expertise that took you a decade to learn and fine tune and obtain, right? You can take i either direction, but it feel free.

Aoni Wang (16:06)

Okay. Yeah. Yeah. I can yeah, I can I can speak I can answer the first question a bit about the negative impact of AI or whether people feel it is negative. To me, when execution becomes cheap, like everyone can come with a prototype. Like someone can create a prototype in minutes. If you have a like a hundred people on your team, you can have like a hundred prototype in minutes. That is definitely going to create some sort of chaos. And I think in some way, people are eager to build. In some way output can be mistaken as progress. But to me, I say that the tricky part is that thinking can be slow, can look slow and can be hard, and building is cheap and fast. So I think there’s a real risk that people can quickly fall in love or default to a specific idea before thinking about whether that makes sense to build.

Grace Shao (17:02)

Because the opportunity cost is significantly lower essentially. Like you don’t have to prototype for months and months. You don’t have to go through all this these hurdles to get to that place anymore, right?

Aoni Wang (17:12)

Yeah, yeah. And then in the product world, tangible things are always going to feel more convincing. So it is there’s a risk that a tangible prototype or a functional prototype actually underneath it is not a really good idea. So this is where also I look to you, Yiyang, to see if you feel the same way. And i if product

Yiyang Hibner (17:27)

Yeah.

Aoni Wang (17:29)

Managers also play a significant role now in kind of defining the direction and making sure we’re rowing toward the same thing, at least on the design side when A lot of things can all come at once. The effort is really in ensuring that these different pieces or different features, different ideas that people build can connect together into a cohesive experience and that takes time. So

Yiyang Hibner (17:52)

Mm.

Aoni Wang (17:52)

I think fundamentally the risk is treating speed of output as the same as speed of progress.

Grace Shao (17:59)

That’s a very interesting way of putting it. Okay. Yeah, what’s your take on it?

Yiyang Hibner (17:59)

That’s true. Yeah, that’s a great way of putting it. Yeah, the first one about the negative part. I could see a few things. Number one is that like Aoni said, once you have a prototype, it looks like, so easy, it’s done. But you had to think, okay, you know, to build a house versus seeing a 3D model rendering, that’s a big difference, right? It and then the second part that where where I want to highlight more. Is that sometimes AI can be too agreeable? Let’s say somebody has a really bad idea. You partner with Claude, what do you think? Blah blah blah. Cloud say, You sound great. You’re genius,

Grace Shao (18:32)

And they’re like, You’re a genius.

Yiyang Hibner (18:34)

This is great. I love this, you know. Especially some AI tools. If you don’t give kind of specific instructions, for example, for my own flow, right? I would actually in my Claude memory, I want Claude to challenge everything I say. Right, be a true thinking partner. Like the default should be don’t agree with me; tell me where it could go wrong. I think that is kind of a almost like a slider in your brain you had to turn on because AI could make you go to a deep dark hole where you are so lost that you don’t even know, right? And then the third thing that could be negative is that you don’t know what you don’t know, right? Let’s say if I want to pull some data on some things, right? And I just ask AI, you know. If I don’t cross-check the work, right, if I don’t ask my data science partner to verify, it’s like, okay, this is a great number, let me use it. Because you know, the AI quoted this source, this source from this data, from this data, that looks good. And then if I just present it to leadership with that number, and somebody in the room is gonna say, how do you get that number? Right? How did you verify with people? If I didn’t, then that would look really bad because you just take whatever AI said for granted. When you don’t have the expertise, right? So the better way to combat this is that sure you can get your number but verify with an expert, right? So don’t make a mistake of a rookie person that just takes the hallucination as a source of truth, right? I think that’s a third part I want to highlight. Regarding your second part of the question on kind of a junior people, it is challenging for them, no doubt, right? Because Now every company is gonna say, hey, do I really need to hire a more junior person or I could just buy more tokens? Right? The AI could work day and night versus juniors they might need to really take time to ramp up. But on the other hand, AI can make almost everybody be a builder. You could be a solo builder. If you have an idea, maybe you can you know find product market fit, you can run with it. So again, the entry to Doing something great has been lowered, right? You don’t need a technical co-founder, or sometimes you do, but most likely you could explore a lot more than before, right? So for the juniors at the same time, that they have to also acquire taste, right? Maybe through kind of study of the great products, we’re talking to other peers or finding mentors, you know, from the industry mentors who have seen, you know, a better times versus just take whatever AI. That’s default.

Grace Shao (21:01)

That’s really interesting. Do you I just wanna kinda double click on this again, just s as in how do you think AI will then change your role in the next three to five years? How do you think your roles as a pot product manager and a designer actually will change like fundamentally? Like the requ the skill sets required, how will that be different?

Aoni Wang (21:22)

Yeah, I can yeah, I can continue with kind of like this idea of using prototype as like a very quick way to get something out, to illustrate an idea or to build something really functional. I think as the more AI enables to create. The more important actually design becomes. It’s just that where design adds value and what designers will focus on will shift. So like everyone can yeah, everyone will be able to build. But just like Jan said, I think like these expertise in different fields, it’s still going to be valued. For example, the on the data side, on the engineering side, like these expertise are still huge hugely valued. And for design expertise, I see the value in defining the visual language. So defining like the style, this design system, defining what does a good experience mean and how to scale that. And also connecting everything into a coherent experience that is designed for human. I think this is where if I were to think about, I a lot of things that I do today, AI can’t do really well in three to five years. There’s this is how I see where design continues to add value.

Yiyang Hibner (22:33)

Yeah. I see I think that the boundaries of roles could be blurry, but eventually it will still people should still kind of drill down to their specialties. What I meant is that you could be there’s a new word member of product staff or product builder, all these names, but the nature of these are still product managers figure out how to build and help lead a team with a vision to execute it, right? Whatever title that is. Okay, so how the role, right? How the role would evolve. Okay, yeah. So my perspective is I think initially the boundaries would be blurry, but later on the lines might be clear again. One example is that I listened to a podcast, there was like a VP of Facebook, Meta. He walked back his previous podcast mentioning that PM should ship production code. Right, Meta was all about, people should ship production code, whether you’re a PM or designer. At that time, I was thinking, huh, is that the best use of your time? Right. Like what if PM breaks something? You know, who’s gonna fix it? You know, it’s gonna take more time, it’s gonna be more chaotic. Of course that VP figured out the same thing a few months later. So he actually said he realized that’s not the best way of using PM’s time. Right? Sure, anybody can ship production code, but for a person who has been vibe coding versus a person who’s been coding for 10, 20 years, the quality of code is dramatically different. Right. So initially it looks like the lines are blurry, you can do this, you can do that. But what I meant is that eventually whatever you are good at, you should be the best at it. But at the same time, whatever you were not good at, you should be at least mediocre. Right. So I think that’s where you kind of draw the line, right? PM could ship code at some companies, right? If you don’t have that many staff members, you’re tight on resources, you know, you roll up your sleeve, do a bunch of other things. But at larger companies, they’re hiring you for doing something in your domain really, really well. And then of course you have other bars you don’t want to fall behind either. But your specialty and your niche will become more important than the rest of the abilities. That’s how I think.

Grace Shao (24:43)

I agree with that. I actually think it’s like becoming a generalist, actually that bar has been lowered. It’s really easy to obtain basic knowledge and scratch the surface a lot of things now. But to stand out, you have to be even, if anything, better than you were before to really stand out because there’s gonna be so many people that are kind of mediocre at doing what you used to be able to do because of the tools. Okay, I wanna shift the focus. I really appreciate you guys giving me a lot of insight into like how you were thinking about your industry and your roles. But one thing, Aoni, I’m gonna direct this to you is what you were saying was that previously you really lived on Figma, right? And Figma was so useful because of the collaborative nature and people in the work the same workplace. How has that really changed, especially now everyone’s starting to build different environments? Like you mentioned, everyone’s kind of within their own, like Claude conversations, how does a team so maintain shared resource and how do they now collaborate and work together?

Aoni Wang (25:40)

Yeah, so I can speak from my experiences. So to me, today general AI is really strong at gathering and synthesizing work context, connecting all the relevant information about a project. And like Yiyang mentioned, like if I have like question about data, today like AI can likely just give me an answer without me bothering my data coworker. And the design capability of these general AI models is r also really strong, which is why most of my work today they start from conversations with AI, just clarifying a lot of the context, gathering all the information I need to make design decisions. And even they can sketch out ideas. So I would say nowadays I probably don’t sketch out ideas in Figma or on my own anymore. I got those ideas or wireframes just from chatting with AI. And to me, after sketching out the ideas, I usually would turn to Figma for the high-fidelity work, working on the pixel-level details myself before handing that off to coding agent to make interactive prototype. So I’d say Figma is still great for its shareability. The collaborative nature and also a great tool for commenting and providing a bird’s eye view of the design or documentation for example having each screen of a flow laid out as a source of truth versus someone having to find the prototype and click through the prototype to understand the flow. And then I think you brought up a really good question around teams when teams start to have like when each person uses AI on their own, setting up their context on their own. How do people continue to maintain a source of truth for work or shared artifacts? I think Figma can still provide some of that value, but I still feel like this is definitely an interesting problem to solve in the AI era, like how you connect everyone’s work and how you keep a source of truth for the team versus one-to-one for everyone working with their own AI agent.

Grace Shao (27:45)

That’s very interesting. Yiyang, how do you think, you know, AI is affecting the old software stack or like, you know, have you shifted some of the workflow from previous softwares to now all like these AI products, or are you kinda using them in tandem the same way the Aoni is?

Yiyang Hibner (28:03)

Yeah, I think fundamentally, right now for at least for work, I use company-approved tools, right? You know, I want to make sure everything is, you know, secure, especially working in payments. It’s a lot of sensitive data. Versus in other companies in the past where from I heard from other people, some companies might not be that strict, right? So you could explore let’s say AI note-taking tool running in the background, nobody’s asking questions. Or some other companies, they encourage you to try out different flows, you know, you especially some AI tools you have, like you know, trial period or something. So I think each company has different policies. But for me, regarding the software, because now I follow only kind of the company-specific softwares, it actually makes me think, if I use cloud, what are other tools that could replace, let’s say in the past I might have three or four different tools to manage one thing. Now I use cloud or Claude Code. Or Gemini you know the tools, what are the best ways for me to centralize my usage? So I don’t actually get distracted by three or four different workflows. I think that makes a difference. Which means that having limitation actually open up a lot more opportunities, if that makes sense, right? Because you have to think, what are

Grace Shao (29:14)

Yeah, it’s interesting.

Yiyang Hibner (29:16)

Yeah, what are the use cases that I really want to solve versus you know for for me in the past or other kind of peers Like, you want to try all these like different new tools to see okay which one, you know, would work. But now it’s like, hey, I have limited sets of tools that are approved by policy. So what are the best ways to approach the current problem I have? You know, if that’s even something worth solving at this moment. And that’s a mindset shift for me. And another thing I would say is that you know, for for my per but in my personal time, right, I do want to see, hey, how does this new AI agent work or how does this new popular tool work? So I basically I keep the boundaries pretty clear for the personal learning versus I for work, actually put things into production and stuff.

Grace Shao (30:01)

How does the effect like I’m assuming a product manager have a lot of like project tracking, product tracking tools, like the monday.Coms and Trellos of the world and whatnot, right? You have more more of these d do you actually still use these tools or are you actually vibe coding your own project management tool essentially?

Yiyang Hibner (30:17)

I do still use these tools. But in you know, in some previous lives, right, if I didn’t like the tool as much, maybe I will try to vibe code or see how it works versus now that in the tools are very specific. It might not have my use cases, but then I will think about okay, Claude, can you pull me a spreadsheet of all the things that are assigned to me, give me kind of status updates versus, you know, finding out the perfect project management tool for myself.

Grace Shao (30:43)

Right. Okay, I think I wanna also bounce it back to Aoni. Like in terms of creative work, I think what’s been really fascinating. So I’ve actually been interviewing a few companies in the creative image generation, video generation side. Of course, I mean it’s not directly all the same, but just there’s such an array of products and it feels like it’s less obvious there’s a clear winner right now. As a creative, what are you really looking at when you are trying out these tools and products? Like Is it like sophistication of the end product? Is it the ease of the use case and communication? I guess how do you really convey your visual thinking to these products and how do you judge what’s a good creative product?

Aoni Wang (31:26)

Mm-hmm, that’s a great question. I think with a lot of new tools like popping up every day, it is easy to get FOMO. So to me, I think that yeah, there there are a lot of different tools for creative purpose. And I think it’s interesting and aspiring to try them out as they may spark creativity and different ideas. But to me At the end of the day, it is more about what am I trying to accomplish? And with that goal, most of the time, general-purpose AI can be quite helpful already. And especially when it comes to if we think about projects or or work, once the AI is connected with a lot of your existing system, actually general-purpose AI knows what you’re trying to work on the best. So when you go to another creative tool just with zero context, sometimes it’s harder to steer it around the end result that you that you want or you have in mind. So to me, when I think about using different creative tools like or AI tools for creative work, I’d say being able to get to what I want quickly, like spending less effort in a lot of explaining and tuning and that is probably what I’d look for the most.

Grace Shao (32:41)

So these tools also need to understand your lingo, your jargons, like how you convey things. But end of the day it’s really also the context that you already took like built around it.

Aoni Wang (32:50)

Yeah, yeah, I think for some it’s the context. For some, I think I’ve seen like creative interactions where like you give like a reference and then you like can use really simple words to describe and work off of a reference. I think some tools are built to optimize that way. So depending on what you’re trying to achieve and how you wanna communicate with the AI, like with work, it’s probably the general AI where it’s already connected with a lot of the work context, the user research, so that it knows the product that. That I’m trying to build, but for other types of creative work, like if you want to build off of a reference, sometimes general AI maybe can do that too, but I know other tools that could that are really specialized in that type of creative work where it’s working off of a reference and they’re probably more tailored to that type of workflow.

Grace Shao (33:38)

I think we already really went into details about how a lot of the roles are blurring in terms of like the lines between them and how you’re empowered to do a lot more than you were able to before. But on that, you know, if execution, like Aoni said, becomes dramatically cheaper and creating new things becomes increasingly faster, what really becomes the most important thing a human can add value to?

Yiyang Hibner (34:03)

I think the interpersonal communication skills, interagent communication skills, these are pretty important. Because let’s say everyone can build prototypes. But which product should we fund? Right? You know, we still have that question. Any company would have limited resources or limited just like how human have limited bandwidth, right? You cannot work 24 hours a day. Otherwise people will burn out, have health problems. Similar for a company, you want to have a healthy company. So you want to put your eggs in different, potentially different baskets, or you want to make sure you threw money in the places where the money will grow, right? So then if everybody has the same, you know, high quality prototypes, let’s say mostly high quality prototypes, good vision docs, then the question becomes which one do we think? Can take into execution, can take to the finish line, right? Who has the charisma or the relationships with your cross-functional partners to get it to carry the team to the finish line? I think that’s very important. Be because ultimately, you know, it’s it’s interwoven when you work with others, right? Unless you’re a solopreneur, right? I’ll do everything. Me and my agents and I do everything. But if you as long as you work in a company, you still need to influence across, beyond, above, below, you know. So I think that the relationship management, how do you how do you manage upwards peers, that become really important. So people have trust in you, right? I think I think that’s a differentiator.

Grace Shao (35:30)

Aoni, what do you think?

Aoni Wang (35:31)

Yeah. Yeah, I resonate with Yiyang. I think when it comes to designers thinking about this world where things are just dramatically cheaper and faster to make, I would say it comes down to prioritization, judgment, which are similar to what Yiyang mentioned, and communication, systems thinking. And to me, building on top of the previous point of a person has to stand behind AI’s work. It is the same, like when you’re in the room with people, you’re presenting It is you who has to listen, respond, disagree, explain your thinking and build trust with the people around you, not not AI. Like AI can coach you, can help you prepare, can help you rehearse, but at the end of the day, you are the person who is building that relationship with other people to work on something together.

Grace Shao (36:19)

Are we already seeing agents delegating agents to do agent work? Like each of you have an agent, your PM has an agent, your designer has an agent, your developer has an agent, your big boss strategy guy has an agent, and then are we already seeing any of that happening where actually there’s very little human communications in the loop?

Yiyang Hibner (36:44)

Agents from humans. I think they’re mostly interacting with humans right now, yeah.

Grace Shao (36:49)

So yeah, that narrative I think is a bit sci-fi and people talk about the potential threat AI could bring to workflow and the workplace. But end of the day, from what the takeaways I’ve gotten from this conversation is really still that AI is not really replacing the roles. It’s just that making you kind of delegate your time differently and your expertise differently. But you still need the humans to make these decision making, right? Look, guys, I think we’ve actually gone through most of the topics in detail. Like a lot of it was just prompt because you guys were very eloquent. You answered like five things in one answer. But I just want to ask, like, is there anything you think we should be talking about? Because you guys are in the Silicon Valley Valley bubble. And actually, you guys are not the researchers. So, you know, you’re not the ones spreading the doom and you’re not the ones preaching the good either, but you guys are plugged in. I just kind of wonder if you guys have something. You want to share with the general public, the people who are very interested in AI and how it’s developing, anything you think that’s missing in the in the conversation.

Yiyang Hibner (37:50)

I can go first. I think my relationship with AI definitely has shifted. Probably will continue to shift. I think definitely I had a fear, right? You know, Claude Code, how does that work? But what I want to give the audience as a little takeaway is that, you know, just do it, right? Just just do it. Once once you dive deep, it’s just an empty file, you know, the text markdown file, right? You know, you keep memories, you know, you interact with agents. It’s maybe I’m just not humble enough. It’s not that complicated. Let’s not overcomplicate ourselves. That I mean the models could change, right? Maybe I don’t know how it all works behind the scenes. But for my day to day, I think AI has brought a lot of joy, you know, so far, right? But at the same time, I agree, I think it is overwhelming, right? There are so many tools. And to be honest, right, I don’t try all of the Every single new tool, right? I did not try OpenClaw. I just want the dust to settle down. And then, you know, hey, let the security flaws surface themselves. And then, like, once people expose all the issues, now, okay, maybe now I’ll see if it’s still worth following. There there is a saying is that if you are slow enough, eventually you will catch up to other things, right? And I think that’s true in the AI world because if I didn’t learn this thing, guess what? Two months later, I’ll learn the latest. You know, don’t feel discouraged when the speed is insane. It just, you know, if you have a specific use case, think about how to address it. The fundamentals are still there, right? You know, ask AI or kind of partner with AI how to solve it. But you don’t have to feel super discouraged or pessimistic about just like, AI taking over jobs. Let us enjoy the things that AI could help us more be more efficient without overcomplicating things. That’s my that’s my takeaway.

Aoni Wang (39:36)

Yeah, I can add some more from the design side. Similar to what Yiyang mentioned, I believe AI has somewhat leveled the playing field among designers as well, that designers, no matter what background or skills you have, you can build something decent on your own, which is why I think it is increasingly important to have some sort of point of view on taste, to have your work really demonstrate what you believe a good experience is and also be able to articulate it. And to double-click on the point of system thinking, I feel that a lot of design projects will become not only the execution of a local solution, but also thinking about the product as a system. So having that type of system thinking will also be important for designers. And I also think this is a great time to be a designer because anything you can imagine, you’re able to build it versus before. Probably you can imagine something, but you don’t know how to build it, or you don’t have the right tools to build it, or it’s gonna take so long to build it. And with a lot of AI-native products, there are also opportunities to introduce new interaction patterns that haven’t been standardized. So that is why it’s just I think there’s definitely some fear around AI replacing designers, but the opportunity to imagine and create new interaction patterns is also exciting.

Grace Shao (40:57)

No, that’s very I love that. So what are some interesting use cases you guys are using AI in your personal life even? Like outside of work?

Aoni Wang (41:05)

I can go. So I think one, AI is really good at turning brain dump into some structured way of communication. So to me, a lot of times like I it may take me some time to respond to people just because I struggle to find the way, the right way to say certain things. So with AI, I’m actually being becoming a more responsive person because whenever I get something that I need to respond to in my personal life, I can just brain dump to my AI and say, this person just asked me about this. And here’s what I think. Can you help help me structure it into a message or an email? And so then I can quickly just respond to people versus sitting on my own.

Yiyang Hibner (41:46)

Yeah, for me it’s more about kind of a health and fitness part. So I have two little kids, you know, I gave birth earlier this year and I feel like I’m in the health journey. As a mom of two, it’s like, okay, how do I, you know, build more muscle and also, you know, get rid of some more fat. So whenever I have some kind of body scan or some kind of, like, you know, in the health journey, there’s like health tracking where I talk to like coaches, a fitness coach and stuff. I would kind of put our kind of conversation transcripts into the AI, you know, have like a lot of things and also track my weight. And then I also kind of wear Fit Mid Air. I haven’t really configured that part, but I basically have a lot of recommendations of what I should do or like repetitions or even you know, like AI I asked AI to generate kind of what are, like, postpartum-friendly workouts. In city features. So I don’t have to watch YouTube videos or I can just look at the paper with like, you know, five different you know, weightlifting positions I could do. I think that makes it easier for me because I could see a world that I can build my own app. The whole user base is just me. It’s tailored to my use cases, you know, my weight or fitness goals, you know, it knows my habit, you know, what time I if I have time, when I can work out shifted, you know, 20 minutes. Dumbbell lifting, you know, into my routine and stuff. I do see that I think in the future as well. I’ve talked to some other PM friends that when they build their personal project, if it’s fitness related, you know, they have a habit tracker or they have a to do list. So because they want to explore, you know, how Claude Code works and they also want to build something useful for themselves. So I could see a an app of one or an app of two, right? You know, you and your partner could share the app or something. I could see that be more powerful in the future.

Grace Shao (43:30)

That’s very interesting. So then basically all the old like tracking weight loss, like fitness apps might just lose their lunch there. But I can see what you mean. It’s much more tailor-made because fitness itself is so personal.

Yiyang Hibner (43:43)

Mm-hmm.

Grace Shao (43:44)

Last question, really. Actually, no, second last question. First is what are some tools you guys would actually recommend to others? Like you think people should check out? Not vibe coded out, but just tools available to them to the mass. What are some things you guys have encounter that are really good. Yeah.

Yiyang Hibner (43:59)

I’ll let Aoni go first.

Aoni Wang (44:01)

Actually I’m I’m very like I’m very anti-FOMO like I’ve been I’ve been just sticking with Claude and ChatGPT for a long time. I feel like they kind of meet all my needs so far.

Grace Shao (44:15)

No, but I like that about I’ve always even though we haven’t seen each other in decades, you’ve always been like that. You’re just like very calm. You’re like, I know what I want,

Aoni Wang (44:26)

Actually actually Granola. Actually, I may

Yiyang Hibner (44:29)

Mm-hmm.

Grace Shao (44:29)

Okay.

Aoni Wang (44:29)

Maybe I’m late to the game. Yeah, I just I just had my aha moment with Granola. And for context, I love note taking. I love note taking like handwritten notes. Like it helps keep me focused in meetings and conversations, helps me sort out my thinking. But what I found out recently is I no longer do pen and paper in meetings anymore and I can be more engaged in the conversation versus trying to catch up to the latest. So I think in that sense Granola is a really great product.

Grace Shao (44:58)

Okay. You know what? Because I was a journalist by training, I’m also one of those people where I have a stack of notebooks on my desk and I actually still take notes when I’m in meetings. I to your point, I think it actually helps me solidify the key thoughts because now there’s all these transcripts, I actually never read them because it’s overwhelming. You’re like, I’m gonna sit here for another hour to read our meeting notes, you know? So interesting. I’ll check out Granola. I’ve heard of it, but I haven’t tried it yet. Yiyang, what about you?

Yiyang Hibner (45:24)

Yeah, I would say personal usage, Wispr Flow is really good. I’ve been using Wispr Flow

Grace Shao (45:30)

I like them.

Yiyang Hibner (45:31)

I’ve been a user, yeah, for three years, before they were big. But I do think they really stood out. You know how Zoom stood out when there was Webex, when there was Google Hangouts. I think Wispr Flow is the same way there are other dictation tools. You know, Apple has dictation tools, but having that just be being able to speak out my random thoughts, you know, when when I’m thinking or even for personal project, right? Like if I have a LinkedIn post, I just want to spit out all my thoughts, blah blah blah blah. And then I will just have Wispr Flow turned on so it kind of organized through my raw notes and I’ll partner with Claude to think, okay, what’s the best way to write it out? Yeah.

Grace Shao (46:13)

Okay. Last question I have for you guys, which is a question I ask every single person that joins the show. What is one differentiated view you hold? Something you think is non-consensus? It doesn’t even have to be about I, although usually people end up talking about AI. So I’ll leave it to you two.

Yiyang Hibner (46:28)

So you’re saying what is something, a unique view I hold?

Grace Shao (46:32)

No, just like, you know, non-consensus, differentiated, something you think people most people think a certain way. You don’t think that way. The most hilarious one I heard was a friend telling me that he thought socks should be worn inside out because of the line wouldn’t actually be on your toes. And I was like, That is fair. It’s something I’ve never thought of. And most people will not agree with you, but I mean Others usually would talk about how they view AI diffusion, China-U.S. AI race, you know, it could be anything from the spectrum.

Yiyang Hibner (47:05)

I need a few minutes. ‘Cause I feel like I’m a I don’t know, like th this is a good question, very thought provoking question. Something that’s non-consensus, okay.

Grace Shao (47:15)

I’m looking into your soul here.

Yiyang Hibner (47:17)

Yeah. I’m looking into the wall. Okay. Let me think.

Grace Shao (47:20)

Ha ha ha. I’ll need you to have one.

Aoni Wang (47:22)

Hmm. Can I say like maybe like swags are underrated? I don’t know. I and also swags are hard to get right. Like I’m as a designer, I’m I’m really passionate about designing swags as well. Just like the free stuff we get and then designing things for the team. And I just have like I just have like 20 t-shirts, maybe 10 water bottles from different companies and events. So which is why I say like it’s it’s underrated, it’s hard to get right because You can really be thoughtful about it. And a thoughtful swag can rule.

Grace Shao (47:51)

What is a good swag? I like what’s a good product? Like I think Silicon Valley is now quite intense because I’ve heard about like pillows, bedding, like these days, like getting things are getting more and more intimate. Like w what are some good swags you’ve received? I just sorry, I just want to say something. I

Aoni Wang (48:07)

Okay. I can yeah.

Grace Shao (48:09)

Hate stickers. We’re not children. Like I have a toddler, she wants my stickers, but I don’t understand why companies give me stickers. Like what I’m gonna do with all these stickers. I’m not gonna put it on my products, like my actual computers and my wallet, you know? So I’m just gonna say it out there.

Aoni Wang (48:21)

Yeah, yeah, exactly. I think it I think yeah, t shirts, water bottles, stickers, these are just like like the baseline. And I designed a pair of crocs for my team last year and it is quoted the best swag ever and I would happily take that praise.

Yiyang Hibner (48:40)

That’s really nice. I feel like that’s that’s very useful, right? Because you get to show off your pride. But I do like stickers. I think y if you guys if you can’t see, but this laptop in the back, it was all stickers from previous partners, previous companies, different conferences, stickers on stickers, you it’s such a war going on. I’m trying to think. Okay. I think my opinion is that okay, the Baseline philosophy I have is that there’s no winner for all remote or all hybrid, like all working in-office. I think companies have different personalities, right? You know, I’ve seen so many things that

Grace Shao (49:17)

Mm.

Yiyang Hibner (49:17)

People just like fight, all remote is better, it’s the best, you know, you save hours of commute time, you don’t have office small talks, you know, we are the best. Versus the other company would be the other extreme, sorry, the other spectrum. Is that people who are loving in the office culture, you’re inclusive, you have a clear boundary between work and life. Versus, you know, if you’re remote, sometimes, you know, you just keep working until you’re tired or, you know, unless something interrupts you. But if you commute, you have clear transitions of I don’t pick up my kids, or I’m gonna train for marathon. So you have that division. Right. So those two spectrum people like to fight. So my opinion is guys, it’s not that difficult, it’s not that serious. Just work for the company that fits into your personality and lifestyle. You could join a hybrid, you again at will employment, right? If you want to change the culture of the company, you either fit in or you fit out, you know, find a place that you like. No need to argue, debate, you know, like attack each other. That’s my opinion. I work for a remote company and I do go to office. So I’ve seen you know both cultures, right? So I feel like it’s not that hard, you know, figure out a company culture that you like, right? Every company is different, right? Some companies embrace 996, right? But if that’s what works for them, best of luck, right? Maybe that would not work for me, right? So that’s my opinion. Just like no need to fight about these kind of topics.

Grace Shao (50:42)

I completely agree with you, Yiyang. I think just that can be applied for to almost everything in life. Like, find what works for you. You know, like

Yiyang Hibner (50:47)

True. Yeah.

Grace Shao (50:49)

As a new mother, like young mother of two, I’m sure you also get unwarranted, like constant motherly advice. I get that constantly, I’m like, you do what you want to do, I do what I want to do, like it’s fine, like everyone’s fine. Like, I totally agree. I think this can be applied to almost everything in life. Just find what works for you and stick with that. But anyway, guys, thank you so much for your time today. I really, really appreciate your insights. I think it was really interesting to hear from you guys directly because, like I said in the opening, you know, you’re in the Bay Area working with these tools, even at AI companies. And so much of so often all we really hear from are the company executives and the research. And of course, that’s really important to know where the technology is going towards like the frontier, but it’s also very interesting and meaningful to hear from people who are. using this technology in day-to-day life and seeing how their work is changing, evolving with the time. Thanks again.

Yiyang Hibner (51:44)

Yep. Thank you guys. I really enjoy our conversation.

AI Proem is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

Get full access to AI Proem at aiproem.substack.com/subscribe