What we talked about
In this episode, Apurv Naman, Product Manager at NVIDIA, shares a heartfelt, visionary journey that starts with childhood cartoons:The Jetsons, Star Trek, Star Wars:and evolves into a career shaping physical AI: humanoids and self-driving cars. What began as wonder for machines you can “touch and feel” became a calling to work where hardware, software, and AI meet real-world impact.
Show notes
Apurv Naman’s path from mechanical engineering to product management at NVIDIA began with a question he kept asking that nobody would answer: why are we building this? As an engineer working on autonomous vehicles and steering systems, he grew frustrated that feature requests arrived from leadership with no context, and that asking “who is asking for this?” was treated as a distraction rather than a necessary input. He got his MBA at Berkeley specifically to gain the standing to ask that question for a living.
What we covered
- Apurv traces his love of physical technology to watching The Jetsons as a child, the show’s flying cars and household robots planted an ambition that led directly to mechanical engineering and then to work on autonomous vehicles, where he filed patents before making the shift to product management.
- His move to product management was driven by wanting control over what got built and why, not just how. He describes engineering as answering “how,” while product management is answering “what” with a “why”, and he found the latter far more energizing.
- Working on physical AI carries a safety burden that software products do not. He points to the collapse of Cruise as a concrete example: a mistake in a physical AI system can end a company, injure someone, or kill someone, which is why safety and traceability have to be top KPIs, not afterthoughts.
- He observed Waymo vehicles evolving from overly cautious behavior, where human drivers learned they could blow past stop signs because the car would always yield, to more assertive merging that mirrors how a human driver chips into traffic. That shift reflects an ongoing product development loop between user behavior and system response.
- On humanoids: he argues that because our entire world is built for human bodies with two hands, two feet, and ten fingers, a humanoid that can operate at human level would theoretically automate anything. But he also notes it is a far harder and less well-defined problem than self-driving cars, which at least operate in a constrained environment.
- His advice for product managers in tech: the one constant is change, and the product managers who thrive are those who are genuinely comfortable with the environment shifting under them, user preferences, engineering workflows, and what is even possible will all change, and AI has accelerated that pace significantly.
About Apurv
Apurv Naman is a Product Manager at NVIDIA, where he works on physical AI at the intersection of hardware, software, and machine intelligence. He holds an MBA from UC Berkeley and previously worked as a mechanical engineer on autonomous vehicles, where he contributed patents related to steering and vehicle control systems.
- LinkedIn: http://www.linkedin.com/in/apurv-naman
- Website: https://nvidia.com
Episode 71 of the PreVetted Podcast.
Full transcript
Federico Ramallo (00:01) Welcome back to the PreVetted Podcast, where we spotlight extraordinary people and remarkable talent reshaping our world. Today we are joined by Apurv Naman. He’s a product manager at NVIDIA, one of the most innovative companies in the world. He’s based in California. Apurv has been contributing to advancements at the intersection of hardware, software, and AI. We’re going to explore his journey into product management.
his experiences into one of tech’s most competitive environments, and his perspective on building products that drive impact at scale. Apurv, welcome to the show.
Apurv Naman (00:42) Yeah, hi, Federico. Thank you for having me. I’m excited to talk to you about my journey and things I’ve been doing and yeah, looking forward to this discussion.
Federico Ramallo (00:54) Great, I’m honored to have you on the show today. ⁓ So let’s start talking about your journey. What first drew you to engineering and what did it mean to you at the start of your career?
Apurv Naman (01:10) Yeah, no, definitely that’s a good question to start with. So believe it or not, like what drew me to engineering was like cartoons. So I used to, I used to watch a lot of cartoons like growing up and then I would, I would love this show called Jetsons. I don’t know if you are or our audience remembers this, but it was about like futuristic stuff of like flying cars or like humanoids doing stuff for you. And I always thought, man, that would be like so cool if we
could have these things one day and I can live in the world which these characters are living in and and just my desire to you know build those things or like my love with the robot like that was my first love with robotics which moved on to ⁓ cars when I saw like the engines for the first time I sat in the car and
I’ve figured out mechanical engineering would be something that will open all those doors for me. like automotive, robotics, anything that moves, anything that works out there, which I can feel and touch and is, you know, innovative in its own sense.
Federico Ramallo (02:19) Amazing. ⁓
I love the Jetsons. I used to watch it when I was a kid. ⁓ I’m still wondering where are our flying cars? ⁓ And instead of that, we ⁓ are settling with having physical buttons on the vehicles, right?
Apurv Naman (02:23) Yeah.
Yeah.
Yeah,
no, they are coming. are coming. I wouldn’t say they are far away. know, it’s like technology is easier to solve, but the bigger question is around other stuff like regulations, insurance and all those things. But yeah, definitely.
Federico Ramallo (02:54) Right.
I don’t know if you’ve seen this show called Futurama.
Apurv Naman (03:01) No, I haven’t seen that yet.
Federico Ramallo (03:04) So the creator of The Simpsons did the show called Futurama and they have a spaceship and they ⁓ talk about that the spaceship instead of transversing through space, the space is fixed and the universe is moving around it, right? ⁓ And that’s how this spaceship can go faster than speed of light, right? They kind of imply that, right? ⁓
Apurv Naman (03:08) Okay.
Mm-hmm.
Mm-hmm.
Yeah.
Federico Ramallo (03:33) Anyway, I find it interesting that they mentioned that because I always, when I saw the Jetsons, I was thinking, how do these cars fly? How does that happen?
Apurv Naman (03:45) Yeah. Yeah.
No, definitely. It’s like interesting, right? So as I said, technology always fascinated me ⁓ from watching cartoons to Star Trek or Star Wars or so on. I grew up with those shows and those things and just…
whatever we imagine, eventually technology is what makes it possible. Think about video calling, right? We are doing it now. When I was a kid, I don’t think this existed or this was even accessible to people like us. But I would read in comics and stuff and they would show superheroes doing a video call. And now today we are on a video call. So technology is like a paint through which you
Federico Ramallo (04:31) Right.
Apurv Naman (04:36) create the image that you can imagine basically. So that’s the way I put it. Like if you can imagine something, like not right now, maybe, but someone is going to build it in the future.
Federico Ramallo (04:48) Yeah, but I think it’s the dreamers that envision what doesn’t exist today. ⁓ And, you know, ⁓ it’s what allowed us to dream to the future and try to push where we are to get there. ⁓ And that’s one of the reasons I love science fiction, On books, movies, because it’s kind of that idea. Right. And.
Apurv Naman (04:55) Mm-hmm.
Yeah.
Mm-hmm.
Okay.
Federico Ramallo (05:17) And it’s wild when you see those movies, old movies, or what was etched at the time. Now it’s obsolete. ⁓ But from their perspective, it was the future.
Apurv Naman (05:28) Thank
Yeah.
Federico Ramallo (05:37) I was watching a video recently where they were comparing how the implementation of video calls, you mentioned video calls, on Star Trek, how accurate or not and what they missed, you know. ⁓ So yeah, I find all of that very interesting.
Apurv Naman (05:46) Yeah.
Yeah, yeah, yeah, I’m
pretty sure like in that show they didn’t have a PM to look into like all the features of video calls. So, but now we have them. ⁓
Federico Ramallo (06:02) you
So moving on, ⁓ can you describe a moment in your engineering days that planted the seed to becoming a product manager?
Apurv Naman (06:17) Yeah, I would say it’s like not one moment that I could like put my finger on saying, yeah, like, you know, this thing happened on this particular day at like 430 PM. And I was like, I want to be a PM moving forward, but it was more like a gradual thing. ⁓ So I started off as engineer, really loved my work. ⁓ Did my masters, started working in like autonomous vehicles, built like some really cool stuff, had some patents and I was like enjoying
my work as well as my life was good at that time and then I just grew ⁓ with the technology and everything and I wanted more control over things that I was building and
Like I did not honestly, did not know like PM because the companies I’ve been working in wasn’t like very agile or they did not had like product manager as a role. It was more like a water flow, like traditional engineering orgs. But I wanted to be like more leader leadership based roles, more influential and grow ⁓ in what I’m delivering as a product and have a say in what features I want to build or why we are building.
So the question of why like there were sometimes requests
coming in from leadership that we need this particular feature in this vehicle or in this ⁓ algorithm. And I would always question that like, why do we even need it or like who is asking for it or maybe there’s a better solution to do it. ⁓ Which is, you know, which is like a fair question, but at times I wouldn’t get the answer to those kind of questions in my engineering days. And I was just like, yeah, that is a very good question. I’ll get back to you. But right now we focus on doing this thing right now.
So I think so things like that and drove me to doing an MBA degree, ⁓ which was like, want to gain more leadership skills and grow up in ranks and have more control over things that I’m building. And that’s where I got introduced to.
this concept of like product management or how product management field works. I was in Berkeley in Bay Area and product management is like all the rage over here. So ⁓ I met with a lot of alumni who were doing product management in organizations like Meta, Google, even Nvidia. And that like caught my interest. ⁓ that’s what I wanted to do, right? Like figuring out why because
That’s what product managers do is like find the question like why do you want to build this thing or like which thing do you want to build?
⁓ based on those wise. So that’s what led me to product management as like a field where I was interested in and that’s when I decided I want to do product management and then I also interned with Nvidia for my summers while I was doing my MBA and I had a very good experience ⁓ at the company, ⁓ amazing team, very innovative culture and I was like this is something I want to do long term and as see it as my
career so I ended up working for them eventually.
Federico Ramallo (09:31) Right. Right. Yeah, because on the technical side, we kind of try to answer the question, how can we make it happen? And but we don’t ask the question, what should we build? Right. We We focus on the how, not the what or a little bit of the why, if you like. Right. And and then when you move to product management.
Apurv Naman (09:40) Yeah.
Yeah.
Yes.
Federico Ramallo (09:59) you answer the question of what with a why, right? ⁓ Yeah. How has the shift from ⁓ engineering is ⁓ things oriented versus product management is more people oriented, right? How was that transition for you? Was that a challenging transition?
Apurv Naman (10:02) Yeah, yes, exactly.
No.
I wouldn’t say it was challenging for me. As I said, like it’s…
It kind of depends on the person as well. So I always wanted to be more out there interacting with people like talking to customers or talking to stakeholders. And I always ⁓ as an engineer, even that was something I felt was even missing in my day to day is like more interactive, more ⁓ people facing ⁓ responsibilities, more
live like you know talking to customers and talking to different stakeholders getting their viewpoints having those discussions or having ⁓ like debates over things ⁓ rather than building stuff which was like again as i said it was pretty cool like great experience and like i learned a lot from doing that ⁓ but
I think so I just have it in me of like going out there and talking to people. Even during my engineering days I was very, I wouldn’t say like extroverted but like more driven towards like people oriented type of roles. So it wasn’t like very difficult for me. I kind of just like fit in as a glove I would say. I’ll let my manager answer that question honestly.
But yeah, I would
Federico Ramallo (11:47) Right.
Apurv Naman (11:48) say I kind of liked it. I as a job, like it never felt as a drag or like, you know, something that drains my energy after like doing a week of product management. It’s something that I look forward to going into meetings where I know people are going to say, no, this is not going to work. Or like, we are going to have to wait. It’s about, know, why my solution is like not the best one or they have something better or even like getting people aligned on.
having a decision saying that ⁓ all of us needs to move in this particular direction, like we don’t need to do 10 different things. This is the right thing at this moment of time and as a team, we should focus on that. So, yeah.
Federico Ramallo (12:30) Right,
and you mentioned something interesting that we usually think of product managers of deciding what to build, but it is a nuance, a little bit more subtle, what you’re saying is you’re also deciding what not to build so you can focus. And…
Apurv Naman (12:43) Mm-hmm.
Mm-hmm.
Federico Ramallo (12:55) we never see what hasn’t been built. you know, it kind of goes under the radar, but it’s important because if you don’t have focus on the product you’re building, ⁓ you’re all over the place, you’re wasting resources, right? And you’re not moving the product forward, right? So…
Apurv Naman (13:11) Yeah.
Yeah, exactly. Yeah.
No, that that makes sense. I think so as product management, like one of the things that like it’s very overused, but it’s basically the truth. Like prioritizing is your like main job as a product manager. How do you prioritize? Which thing do you prioritize for and who do you prioritize for? Right. So
that is one of the most important skill or like most important work as you will do as a product manager and yeah you will go wrong definitely or you can go wrong but learning from those mistakes and then you know trying to figure out like what’s the best case like my concept in general is users
So understanding the users and then deeply getting to understand their pain points ⁓ and then prioritizing on the user needs other than like other goals we have. So if you build something for a user to make it like user friendly so that they use more, all other KPIs will follow, right? Like no matter what you’re like, you want to look at monthly average users, you want to look at like revenue dollars, you want to look at like graphic card users or whatever, like, know, it’s like as long as
you build a product or a tool or a service that a user loves and they will keep on using it ⁓ because it’s solving one of their pain points you’re gonna get the whatever metrics you are chasing after
Federico Ramallo (14:44) Right.
How did your technical background shape how you approach product management?
Apurv Naman (14:53) No, I think the technical background was really helpful when we talk about product management. I don’t think so it’s required for sure. It might be required in some roles in some companies, but having a technical background as a product manager anywhere helps me better understand my engineering team.
And, you know, I like, there’s one thing about understanding users and their user needs of like what they want, how they want. And then the other aspect is like understanding what can be done and how much time would it need if I want to pull something off like this. Right. And.
Engineers are also like coming back with their own ideas, which are sometimes even better, right? Like they are amazing or but I need to understand that idea like why it is better or where they are coming from and Sometimes some ideas are not good and then you have to tell them in a way without like saying that this is not a great idea, but you have to
exactly understand where the engineer is coming from because how they are thinking about things, how they are looking at the problem and how they came up with the solution to identify if it’s a good solution or a bad solution or how to handle that situation. So I think my technical background helps a lot in my day-to-day.
Federico Ramallo (16:18) Amazing. What happens to me is that when I try to keep the balance, right, but sometimes I get too ⁓ intrigued or attracted to ⁓ the technical aspect. So ⁓ I kind of go into this mode of, this technology is cool, this feature is cool, right? And I kind of forget.
Apurv Naman (16:37) Mm-hmm.
Yeah.
Federico Ramallo (16:48) thinking about the user, right? And then sometimes I realize that and I go the other way around, right? I think more on the user, not much on how to implement it, right? ⁓ So, and when I find that I’m going to these extremes, then I try to go back on finding the center, right?
Apurv Naman (16:52) Yeah.
in
Yeah, no, it’s a balance, right? As you said, have to understand. So like engineers are your main stakeholders, if you think about it. Like users are important, but there wouldn’t be any product managers without any engineers, right? Like you need to have them and then they have like…
They are the ones who are building stuff to help get you there where you want it or solving user needs. I look back at my time as an engineer and whenever I talk to engineering managers and leaders, I think one thing that I’m very particular about is telling them why instead of just telling them we need to build this or things that I felt as an engineer, like why I’m being kept in blind. ⁓
More
than that, I give them the whole user story is like, ⁓ see, like, I want you to build this particular feature in this product because, you know, if you think about a user, he wants to do X, Y, and Z and to do X, and Z, he needs these kinds of A, B, and C abilities. And if you want to enable these abilities six months or a year from now, we need to start thinking about them now and we need to build our infrastructure in this particular way. So laying it out like this.
because you know I have both my technical background as well I am connected to the users is how you get alignment between different engineering team, marketing teams, pricing, UX teams, what not.
Federico Ramallo (18:45) Right.
⁓ So talking a little bit about physical ⁓ AI, is something I’m very interested to discuss with you. ⁓ Working on physical AI and humanoids feels like science fiction made real. What excites you the most about it?
Apurv Naman (19:01) Mm-hmm.
Well, I think the same thing that I started with like, ⁓ it’s technology that you can like feel and touch and you can interact with. Right? Like apps are like, I’m nothing against software or apps because they are driving the world anyways. ⁓ But they make your life easier for sure. ⁓ Right? Like you have an Uber app on your phone, which helps you get a cab. But at the end of the day, it’s the cab, like the technology in Uber app or whatever.
you
cannot touch or feel it but like having an autonomous vehicle or a humanoid doing stuff for you it just like the physical fact of it is like oh you can sit inside a Waymo or you can you know you can grab the door and open it or like you can be sitting in a Tesla behind a wheel doing its own driving that aspect of like physical AI or just robotics is something that excites me a lot and
It just makes me feel how intrigued, how entangled we are when we have technologies that become part of our life.
Roaming or like this to be if you take an example of like the robot vacuum cleaner iRobot right? I just I just love it. It’s such a simple technology But it does like so much for me and then I can see it. It looks cool It’s in my house And then I can like hear it see it do it work, right? Those things are something that you wouldn’t have with just like an app or like, you know a chatbot which again
They are amazing. They are what driving those hardware technologies in the end. But just the aspect of seeing them, able to feel them, hold them in your hand makes makes me excited.
Federico Ramallo (20:54) It feels almost like magic, right? When you see, ⁓ you know, like these ⁓ humanoid robots that can jump, you know, walk up and down stairs and obstacles, right? I see those and I mean, I know that there’s a lot of software behind it, but my reaction is, this is magic, you know, it feels like magic. It’s amazing. ⁓
Apurv Naman (20:57) Yeah.
Yeah, exactly. Yeah.
Exactly,
exactly. And then the potential is like, if you think about the potential is insane. Like, you know, the industry, haven’t even scratched the surface of what is possible with physical AI. Like humanoids, in theory, if you have a humanoid that can behave as a human, you can automate subjectively or theoretically anything, right? Like anything in the world could be automated. Like cooking in your house or like even driving,
do you need fsd like or self-driving cars you can just have a humanoid sit behind a wheel and drive for you but like maybe like that’s not where we are headed like obviously self-driving cars without a humanoid are more effective they don’t have latencies they are faster they are purpose-built right which which makes sense but i’m just saying like it’s
Our world is built for humans. Everything is built around us for humans with like 10 fingers, 2 feet, 2 hands and what if you can automate that? Right? Now you have automated everything.
Federico Ramallo (22:29) Yes, wouldn’t that be great? I mean, that would be amazing.
Apurv Naman (22:33) Yeah, yeah, it’s like, I think it’s one of the toughest problem to solve and it’s gonna be like that for a while. AI has enabled us to like, you know, unlock a lot of things that were not possible before, like self-driving cars in itself is, it was a complex problem. It’s a problem that is well-defined in some context. So…
it becomes easier in that way compared to humanoids which is not well defined at this point. ⁓ But again, it’s when you start solving a problem look into it everything looks like monumental but eventually you are able to solve it and it’s like the amazing feeling you get.
Federico Ramallo (23:18) ⁓ I’ve been watching, observing the Waymos autonomous vehicles and I haven’t got into one yet. There is ⁓ some sense of… ⁓
Apurv Naman (23:23) Mm-hmm. Yeah.
okay.
Yeah, we need to change that. We need to be on our way
more.
Federico Ramallo (23:38) Yeah, mean, for me, it’s unsettling the idea of not being able to grab the wheel, right? And I know it’s probably a backwards thinking.
Apurv Naman (23:48) Yeah, no,
no, I, that’s like you are there with like thousands and like even millions of people out there. Like, you know, that’s a genuine feeling and that’s something definitely there’s a product manager looking into that user experience and trying to make it as smooth as possible for people who have this issues. ⁓ And like that’s, that’s a genuine problem and that needs to be solved. like once you have, ⁓
experienced it, you won’t regret it is what I would say like as a customer as a user I don’t know about like something but as a user I would say like first time I sat in it like the way it drove it was like like why why do I even have a car that I have to drive why like I’ve been working in automotive for a while and I hate driving so
Federico Ramallo (24:33) the
Right, yeah, I can understand that.
And I can see the benefit of getting into a car that drives you around, right? ⁓ The stories of ⁓ people that was horse riding and the horse would be able to drive, quote unquote, drive back home, right? When the rider was incapacitated, right? ⁓
Apurv Naman (25:00) Yeah.
Mm-hmm.
Federico Ramallo (25:08) There many stories like that, right? ⁓ So I can see the benefit, ⁓ but that doesn’t mean that I feel unsettling, right? Yeah, and what I notice is that the Waymo vehicles has been very careful on the driving side, ⁓ very careful on the driving. And I noticed recently that they have been…
Apurv Naman (25:16) Yeah.
Federico Ramallo (25:33) more and more aggressive, ⁓ they accelerate faster, they take turns that, you know, I’ve been observing that they wouldn’t do that before, right? So I find that very interesting and I think that’s part of the development of the product that they’re doing, right?
Apurv Naman (25:49) Yes,
yes. Yeah, no, those things will evolve, right? Like if you would have been in initial stages of…
Being in an AV, what would usually happen once we as humans figure out that this is an AV that’s gonna follow all possible rules in the world that are laid out and it’s a right and it’s not gonna make a right turn unless you know if you are on the crosswalk or you know it’s even if you don’t have the cross signal on and if you step on the crosswalk that car is going to stop no matter what.
people start taking advantages of those things, right? Like even if it’s not your right of way, you are blowing past a stop sign, the Waymo will stop it. It’s not gonna honk at you or be mad at you, right? So once you get that understanding, people start adapting to those things, right? They would start like behaving in a certain way where they were like, it’s a Waymo. I don’t care about my stop sign. I’ll probably just blow past it, which is wrong on like so many levels, but.
Federico Ramallo (26:34) you
Apurv Naman (26:51) We do that as human beings. So technology has to adopt in some way to change that perspective or change that driving behavior or you keep getting stuck. So I was in a way when it just like stood there because no one was giving it way to enter, right? Like it was trying to cut in and everyone was like super aggressive. ⁓ It wasn’t able to cut in like how you usually cut in is like you slightly start moving and then the person behind you.
gives you space how much like you know mad they are at whatever but that’s how it works now I sit in a Waymo it’s like aggressive if it has to cut in it cut ins like a human driver like you know just start chipping in and and merges with with the traffic which is pretty cool so yeah these things evolve as we evolve with these things so it’s it’s it’s a dynamic problem ever changing
Federico Ramallo (27:49) Right, right, and ⁓ it’s building this relationship between the users and the product, right, that it’s something that is evolving, right? So what impact do you believe self-driving car would have in society and daily life?
Apurv Naman (27:57) Mm.
⁓ it’s like, again, as I said, we are just like scratching the surface. Like impacts are huge.
Right now, we are able to drive like people imagine people who are like old and they’re they’re driving license or driving privileges have been taken away because you know, they cannot see in night or like it’s like difficult for them to react. We are enabling them in some way or the other to drive to mobility. think so mobility is kind of one of the human rights. It’s not there, right? Like it’s but it should be. You should be able to move and get to from place A to place B to go to your favorite spots to eat in your favorite
restaurant like you know go to the movie that you wanted to watch in a theater without thinking of like I need to I cannot drive I you know I don’t have access to a car or the parking is going to be intense or whatever reasons you cannot drive you should have and I agree public transit solves a lot of these problems and public transit is one of the biggest solution and one of the biggest competitors for self-driving cars but
Like in my ideal world, ⁓ combination of public transport and AVs for last mile door to door services is what is going to like save, like help us or like improve our lifestyle or.
help us get there where like you know anyone who’s even blind can go and sit in a car and go to where they want to go or like children if you did they can sit in a car like you can book a cab for them without any safety concerns because there is no driver in it it’s gonna be safe it’s school drop-off school pickups they can be automated right like we are not there yet we are like ways beyond that but eventually if we we get there those things can increase our quality of life to the next
level, right?
Federico Ramallo (30:02) Yes, yes, I agree. mean the…
I was talking with a founder that was proposing to use ⁓ his building a platform so the car could do your daily chores, right? A self-driving car. So ⁓ negotiating with the… ⁓
Apurv Naman (30:15) Mm… yeah.
Federico Ramallo (30:21) vendors on the other side. So could go and pick up your medicine and then a human would put it in inside the car, right? You know, and things like that. Or your groceries, right? Just the car shows up at the, you know, at grocery store and then a human put it in your trunk and then the car drives back to your house, right? Or as you said, having, you know, being able to…
Apurv Naman (30:22) Yeah.
Yeah.
Yeah.
video.
Federico Ramallo (30:47) put a kid in it and just drive it to school. There’s so many things like that that’s interesting. ⁓ And then you mentioned something interesting about public transport. ⁓ My concern with public transport is sometimes ⁓ you lose privacy. ⁓ Having your own space ⁓ sometimes become important. ⁓
Apurv Naman (31:12) Mm-hmm.
Federico Ramallo (31:14) So I mean, if you’re going in your car ⁓ to one place or to another, you’re in your safe space, right, within the car. So I see the benefit of that and having this idea of shared self-driving cars, which is a mix of public transport but with a ⁓ private space, ⁓ that would be life-changing for sure.
Apurv Naman (31:20) Yeah.
Yeah, no, no, because like that’s like there’s there problems to it too, right? Like if you have a lot of like self-driving cars, the congestion is just going to go up like cities won’t be livable if they are littered with like these vmOs or soaks or whatever is on there on the road like and you cannot do it like for everyone. It’s just how humans are like we are like so many of us out there and we have a distinct need so no kind of like one
Federico Ramallo (32:05) You
Apurv Naman (32:09) solution that fits all is going to work like there are people who argue like why do we even need self-driving cars public transport is the best thing and like if you go to Singapore very well connected amazing public transport right and I went with my parents and then they are like old and they were having difficulty even walking to the public transport right from from our house ⁓ what happens when it’s raining right like it becomes even more difficult for them to do that
But even you cannot leave a child alone in public transport. However, Singapore is an amazing super safe city. Still, I wouldn’t trust anyone leaving my child with that. But I can do it with a self-driving car which has cameras in it which I can also watch or whatever technology we figure out to do that.
So it’s like there’s no one solution that fits all. You have to go hybrid. You have to do something like which serves all our purposes and makes the city better and makes our commute better and makes our life better, right?
Federico Ramallo (33:15) Right, right. And I think that the other issue that public transport does not solve, that self-driving cars would, is routing. You kind of mentioned it about, you your parents going to the public ⁓ station would be challenging. ⁓ But then once you have the routes on public transit, they’re fixed. So if you go to some other place, you have to…
Apurv Naman (33:25) Mm-hmm.
Mm-hmm.
Yeah.
Mm-hmm. Yeah.
Federico Ramallo (33:43) to do combinations, whereas with a self-driving car, you can set the destination and the car will be able to route towards that destination. Whether it’s in a closed loop, closed
Apurv Naman (33:53) Yeah.
Federico Ramallo (34:01) section, know, where you get into a vehicle that only deals with, you know, like a little train, you know, it only deals with other cars, vehicles like them, right? Or it goes into the public roads, right? ⁓ So, but either those scenarios. ⁓
Apurv Naman (34:20) Amen.
Federico Ramallo (34:27) The idea of avoiding routing, think that’s critical, right? Because then you have the flexibility to go door to door, right?
Apurv Naman (34:30) Mm-hmm.
Yeah, like yeah, I agree. It’s it’s again as I said like it’s more like product management as well, right? These things are going to be iterative It’s not gonna be like one company comes up with a solution and then everyone is like, this is the best and that’s it like even like Like imagine Facebook that we were all on five or ten years ago. They keep changing
Technology is like, that’s the best part about tech, right? Or being in tech is like, nothing is constant. You have to evolve, you have to change, you have to iterate. You have to make things better, right? So like these solutions, as we are talking about, like even like maybe these are not the best. So we don’t know, right? Like we have to do maybe an A-B testing. So to figure that, yeah. So, but like you have to like be iterative and find solutions and keep doing that. So.
Federico Ramallo (35:12) Bye.
Right.
Apurv Naman (35:33) Which is pretty interesting and which is cool and then I’m like excited to see where this world of like physical AI with these self-driving cars, humanoids, what not like where it’s gonna take us and how we are gonna evolve as a society and yeah.
Federico Ramallo (35:52) So how does working on products that move, that interact with the physical world, how is that? That sounds fascinating.
Apurv Naman (36:04) It’s
challenging because so like with all the stock of iteration and improvement room for error in these products are very minimal right like if you think about it like you can you can do an AV testing on an app which might have an API call gone wrong and
Your user might not be able to log in. Okay. Like what’s the worst they could happen. They didn’t get in cab without, because they were not able to log in and they might have to use another app to do it. Right. Like you lose a customer or what? Imagine like a self-driving car making a mistake. can.
Federico Ramallo (36:40) You
Apurv Naman (36:44) break or make a company right like we have example of Cruise going down because they made a mistake and it didn’t end up well for them. You can kill somebody potentially with physical AI like robots are like they look feeble but eventually like they are machines right whatever you do and they have potential to like physically harm you. It can go bad in like so many ways so
Federico Ramallo (37:10) Right.
Apurv Naman (37:14) The main difference I see with physical AI is like safety and traceability. That has to be like one of the top KPIs that a product manager or a team or even engineer has to always keep in mind when they build those things or deliver those things. So that is the most challenging part of ⁓ working in a physical AI space compared to any other space we are in.
Federico Ramallo (37:42) Right, right. ⁓ Yeah, because I’ve seen robots are using industrial engines, know, motors, and they have a lot of strength. So it will be very easy to, you know, crash a human or break some arm or whatever, right?
Apurv Naman (37:48) Mm-hmm. Mm-hmm.
Yeah, like
even like if you think about your car, so there’s a there’s like a motor in your steering wheel, right? That I worked on that too in my initial roles and those things are like so they have so much power is like no human being can like oppose them. So if a steering wheel wants to go right.
however much strength you apply on that wheel to turn go straight or go left like you you won’t be able to do that it’s impossible humanly to do that unless you have a software that takes care of it like which will just release the motor as soon as someone like nudges it or touches the steering wheel so even your like manual like the car that you drive today if there’s something goes wrong in that motor and it jerks towards right
Like there is no way you are trying, you are being able to pull it back to the position you have. The amount of torque it has is like insane. So we have a lot of technologies that is like technology, is way beyond like human strength and you, and we cannot override them.
So you have to be like very safe about it. It’s more like we take a lot of, so that’s why we say a lot of lessons we have in aviation industry and automotive industry. I think so they are gonna be applied to robotics, humanoids and all those things when we look into those kinds of things.
Federico Ramallo (39:27) But I think that the situation now is that the level of innovation is ahead of regulation. ⁓ that gives you lot of, ⁓ know, the engineering, it gives the engineer a lot of freedom to experiment taking risks because the regulations hasn’t catch up to what are the things that you cannot do.
Apurv Naman (39:37) Mm-hmm.
Mm-hmm. Yeah.
Federico Ramallo (39:57) And as you mentioned, aviation, caught up and, you know, unless you’re doing a 103 experimental aircraft, which is a very limited category, for any other airplanes, you have to go through years of test and certification before they deemed.
the machine flyable, right? Where you can actually fly. And then after you fly, you have to go through other tests, right? So that hasn’t been invented for the self-driving cars, right?
Apurv Naman (40:32) No, for self-driving cars, it’s there, right? Like there’s a lot of regulation, like not everything, but we have a lot of regulations and a lot of like ⁓ legal ways of getting approvals ⁓ or it’s like, it’s anything that’s out there right now, like let’s say Waymo’s or Zoos or whatever, like I don’t know the details, but I’m pretty sure like they have gone through very strict ⁓ regulations and testing validation to get them there.
What I’m saying, like humanoids is something that is untested for. We don’t know. We still figuring it out. Right. That’s you don’t have any humanoids in your house right now. So you have no idea what to do. Yeah.
Federico Ramallo (41:01) Right.
Right,
going back to the Jetsons, want the, I know it as Robotina because I was watching it in Spanish, but the, you know, but the Robomade, you know, that would be so much fun to have, you know.
Apurv Naman (41:23) Yeah. Okay, yeah, yeah.
Yeah.
Exactly,
like so that’s what’s like from aviation like automotive industry is like very regulated. ⁓ You have to do like thousands of certifications, SAE levels, ⁓ OBD compliances that you have to qualify in order to put a vehicle on the road. ⁓ It takes years of work to get that done and it’s regulated. But like again, as technology evolves, those things weren’t built for self-driving cars. They were built for having humans behind it and
Federico Ramallo (41:39) Right.
Apurv Naman (42:01) Our government is doing great job catching up, adding more regulations and that’s why we see these cars out there on the roads or else you won’t but…
We have to do those things moving forward with like humanoids, robotics, all those things. ⁓ Industries is a controlled environment, right? Like industries, it’s a private place. You can train your workers to behave in a particular way around the robot. You can build like safety nets, ⁓ laser gates that shut the robots off and someone enters and things like that. But if you want to put your technology out there, physical technology out there in a public space, it’s not regulated.
anything can happen. A kid can run on a road or like a dog can like fall from or like come from your side or you know things can go wrong a lot of things can go wrong and you have to be able to respond to that and behave in a safe way.
Federico Ramallo (42:57) Right. Do you see a human element, empathy, ethics, values, playing a bigger role when designing AI that lives in the physical world?
Apurv Naman (43:09) yeah, I think that kind of moves into this conversation, right? What we were discussing. If you feel something, see something, same thing with empathy. If you think about empathy, as you said, like you are skeptical about sitting in a car where you cannot hold a steering wheel, right? So the first part, empathy comes in like trying to understand that feeling because…
If you are in the industry or working with it, you just get used to of like having an idea of car drive itself because you know in and out how these things are built or like what safety measures are there or what protocols or they have been certified. But as a user, you don’t have any idea what went through in like building those things. So that is an empathy that you have to come up with and then you have features built around that so that you do that. There’s a lot of research in behavioral science, which I think so every product manager should look at.
for any product they building. You can do a lot with behavioral science. You can build nudges in your product that help a user guide or their experiences. So making the car a very friendly environment or trying to…
So initially there was this concept of steering wheel moving really rapidly. I’ll give you an example. It makes it easier because if you’re taking a right turn, have to like to a rapid steering wheel turn and that was making users very obnoxious because they were like sitting in a car which like their steering wheel was like moving a lot or like it’s what and then there was other concept where the steering wheel was not moving at all. So you can possibly do that. Steer by wire. The steering wheel is not connected.
to the rack or to the wheels so you can have the steering in theory like being still there and then your car going around and that experience for users was like this is unreal like i don’t want to sit in this car this doesn’t feel very good and then then companies reverted back saying no no we need the steering wheel to move because user needs to feel or see physically what’s happening in the car if you don’t move the steering wheel and car makes a right turn they registered it as an
process or like what’s happening so they added features like this now we have cars that doesn’t even have steering wheels
Right? Like they have, yeah, they don’t even have steering wheel. They don’t even have brakes and that’s where we are headed. But like it’s, slowly like trying to get user used to it. People who are using Waymo are, I would say they are more likely to move to a car which doesn’t have a steering wheel instead of someone. And then again, this is not backed against any data. So don’t hold my word against it. It’s my feeling.
Federico Ramallo (45:27) Right.
.
Apurv Naman (45:49) It would be difficult for someone from this going from driving their car to sit in a box that would drive them around compared to if you let them sit in a car which has a steering wheel which is moving and you can see the car moving so and then next step would be going in a car with nothing in there. So yeah, yeah, you have to get used to those things.
Federico Ramallo (46:07) Right.
Yeah, I mean, it’s a transition that people get used to, get used to things until there’s this ⁓ revolution or evolution that pushes towards a different mindset. then, because before that, there’s a lot of resistance, and then when people make the change of mindset, then everything happened faster, right?
Apurv Naman (46:30) Yeah. So that’s the empathy part where you have to have empathy with the user of how they’re going to, you know, feel about this thing and use it. then those things. And then on the ethics part, like I think so there could be like essays written about it. like how do you define ethics? Like there are frameworks that we use for ethics or ⁓ establish on, on, on
on products that are interacting with humans and they can possibly cause harm to humans and ⁓ then right now it’s more about safety about you know having universal overlay of like you shouldn’t hurt a human being like whatever happens like that’s the main ⁓ arching like overview of a product that is in physical world or like a robot or a
FS car or like whatever but eventually I think these things have to evolve and be more dependent on like what use case you’re trying to figure out have better ethical frameworks have better ethical thought process in building those products
Federico Ramallo (47:42) Right, right. I remember the story in New York when they shifted out of horses into internal combustion engines, right? And, ⁓ you know.
Apurv Naman (47:50) Thank
Yeah.
Federico Ramallo (48:01) before they had the issue that if a horse died, they would leave it dead on the street, right? So they had issues with that and then with the manure, right? You have a lot of horses, you have a lot of poop, so they had that issue, right? That’s why in New York, the houses has stairs. It was to rise above the manure, right? Yeah, the poop, yeah.
Apurv Naman (48:07) Okay.
Hmm. Yeah. Yeah. Yeah.
⁓ I see.
Okay, cool. ⁓
Federico Ramallo (48:31) ⁓ When the car was introduced, it was a very quick shift towards vehicles, towards internal combustion engines, that a years later, horses were ⁓ outlawed within New York City. Because other than ⁓ ceremonial horses on ⁓ Central Park or the police has some horses, but…
Apurv Naman (48:46) Hmm. Hmm.
Federico Ramallo (48:59) Other than that, know, horses were banned from the city, right? So I believe that something similar is going to happen with, you know, self-driving vehicles, that there’s going to be a lot of resistance until you get to the tipping point where it’s safe enough, people trust enough, and then it’s like, we don’t want to go back, right?
Apurv Naman (49:17) Yeah.
Yeah, no, that’s gonna happen for sure. Same with manual cars, right? We had manual cars and now everything is switching to automatic. Die hard fans who love driving manual, they still love manual. I would think so. I’ll be pretty confident in saying 95 % of cars in America, North America right now would be automatic compared to anything. So it’s a gradual shift, technological shift that’s happening.
That brings me back to my old point of like change, right? Like being in this work of like technology or being in tech, like you have to be very comfortable with change or like being open to, know, change it’s gonna happen. It’s like everything would change. Like your user’s preferences will change. Your technology will change. Your engineering workflows will change. And…
Again, as a product manager, we’re working in tech. You have to understand this concept of like, there’s only one thing constant and that is change. And you have to be very, yeah. And you have to be very, you know, used to like working in an environment that is going to change rapidly. And then AI has just like increased the pace of this change to next level.
Federico Ramallo (50:24) Right.
Right. So do you know about the three laws of robotics by Isaac Asimov?
Apurv Naman (50:50) Yeah, I think so. I’ve read it somewhere. Yeah. Like, do no harm yourself or something along those lines. Yeah.
Federico Ramallo (50:55) So, yeah. Yeah,
do not harm, obey humans.
Apurv Naman (51:03) Mm-hmm.
Federico Ramallo (51:06) the robot must protect its own existence, right? ⁓ So I’m wondering, ⁓ now that we’re talking about humanoids ⁓ and self-driving cars, ⁓ do you feel that these laws of robotics are being implemented into the physical devices using AI?
Apurv Naman (51:09) Mm-hmm.
Mm-hmm.
I think so, like, I think so that’s like first ever user stories written by someone for this field, if you think about it, right? Like that’s something that you start with. So I wouldn’t say like they are hard coded, like, you know, in plain English, like don’t do any harm or like, you know, obviously not. But like a lot of underlying principles that go around like these safety frameworks that we have or algorithms that take in account of like human safety.
or safety of other things or safety of the machine itself into account are based on that. Like I would say what he gave us.
was very rudimentary top level you know story points but now what we are building is a very complex machine which have varied use cases and I wouldn’t say it’s like just one of those two top three principles that’s gonna guide us but a very complex set of maybe decision trees or ⁓ like you know distributed decision trees that you can imagine for
for like sake of imagining that will guide these machines or these robots ⁓ cars all those things.
Federico Ramallo (52:48) I think that those laws, those three robotic laws are more ⁓ philosophical than practical. And it’s great to write for, you know, as a…
Apurv Naman (52:56) Yeah, yeah, yeah, yeah.
Federico Ramallo (53:03) element to write the stories, right? Because you can talk about the laws, right? And things like that, right? And the conflict that generates. But ⁓ in practicality, and as you mentioned, the complexity of the systems, you know, not necessarily allow them to implement them, you know, like that, it’s more embedded within other coding and software, right?
Apurv Naman (53:05) Yeah.
Yeah.
Mm.
Yeah,
yeah, and which is true, but like again, like I don’t know, like I don’t know history, how he came up with those things or like how he thought about it. Maybe he was a genius, still to be seen. I think so he was, that’s why we know him. ⁓ But yeah, it’s like those are things that you have to keep in mind for like when you are working. And then that brings us back to your ethics and empathy question, right?
Federico Ramallo (53:35) you
Right.
Apurv Naman (53:54) You need
to have some guiding principles because ⁓ these machines can do some real harm if they want to.
Federico Ramallo (54:01) Right. ⁓
We’re running out of time, but first I want to appreciate you being here today. I think we learned a lot about ⁓ physical AI. ⁓ But before we wrap it up, would you like to share some final comments?
Apurv Naman (54:12) Mm-hmm.
Yeah, no, I would say this was a good experience talking to you, Federico, and this is…
an exciting place to be in like I’m blessed to be in like this working in this great company and on this like great stuff like self-driving cars or like humanoids robotics all those things and I would say like it’s ⁓ it’s like as I said before is like if you just my advice to if so I know your broadcast is more about like product management and helping people you know who want to be a product manager or break into the speed or want to be better I
say like one advice if I can get give to any product manager in this time and date is like you have to be very open to change and how we do things and what things we do right like every especially in tech it’s like so much change around you and around the products that you see and if you’re comfortable with that you will be a very very good product manager
Federico Ramallo (55:29) Amazing, amazing. So Aperv, thank you for joining us today and I’m looking forward to hear what you’re going to build next.
Apurv Naman (55:34) Mm-hmm.
Yeah, definitely. Thanks Federico for having me.
Federico Ramallo (55:46) Hehe