Rebuilding Trust in Technology with Mozilla CTO Raffi Krikorian

What happens when one of Silicon Valley's most accomplished engineers decides the system he helped build is broken—and walks away to fix it? 

Today my guest is Raffi Krikorian, CTO of Mozilla and one of the most civic-minded technologists I know. We explore why the fight for open-source AI isn't just a technical debate; it is really a fight for who controls our relationship with knowledge itself. 

Raffi's career path is uniquely fascinating. He spent his early years scaling massive engineering teams at Twitter and launching Uber’s first self-driving fleet. But then he did something rare. He pivoted to public service, becoming the first-ever CTO of the Democratic National Committee to rebuild their cybersecurity from the ground up. He then went on to drive social-impact technology at Emerson Collective, applying his engineering mind to systemic issues like immigration and climate change. 

At Mozilla, he is now on the frontlines of the AI revolution. We talk about what it means to be "technically optimistic" right now—which also happens to be the name of his excellent podcast. For Raffi, optimism isn't about blind faith in algorithms. It’s about demanding that our tools are trustworthy, transparent, and built to serve humanity, rather than exploiting it. 

In our conversation, we explore:  
→ The Twitter crash that taught him his job was not to be the architect, but to create the conditions for others to do their best work  
→ Why he left Uber's self-driving program after discovering their models misclassified people based on skin color  
→ How a week of Google Sheets transformed an asylum-seeker nonprofit more than any AI chatbot could  
→ His conviction that we need seven billion AGIs—one for each of us—not seven controlled by massive corporations  
→ Why patience, not speed, is the leadership skill that actually builds movements 

"We have outsourced dreaming to a few people who are building companies and we all need to dream again." — Raffi Krikorian, CTO, Mozilla 

If you have ever wondered whether the technology on your phone is truly working for you—or for someone else—this conversation will completely change how you think about what comes next. 

🔑 Key Themes: Open-Source AI, Responsible Technology, Purpose-Driven Leadership, Digital Trust, Civic Tech, Cybersecurity, Technical Optimism. 

🎧 Related Episodes: 

• Kevin Scott — Empowering people with AI: 
• Navrina Singh — Building Trust in AI: Leadership, Governance, and Human Values 
• Rana el Kaliouby — Human-Centric AI (purpose meets profit in technology)  

🎧 Listen on Apple Podcasts, Spotify, YouTube, jpcourtois.com 

New here? Subscribe to Positive Leadership & You for one edition a month, written for leaders who want to build companies and communities people thrive in. https://www.linkedin.com/newsletters/positive-leadership-you-6970390170017669121/

Want to go deeper? Listen to the Positive Leadership Podcast on your favourite platform. 130+ conversations with the leaders, founders and thinkers shaping a more human future of work.

🎧 Apple Podcast: https://podcasts.apple.com/fr/podcast/positive-leadership/id1574911588
🎤 Spotify: https://open.spotify.com/show/1EdnJcYgh9nXPTFJ5euOBf 

JEAN-PHILIPPE COURTOIS: So Raphael, we'll start with a short suggestion by yourself and officially welcome you on the show. Okay. You're used to that. I you've done many of those podcasts, but a bit longer as some others. So don't be worried.

RAFFI KRIKORIAN: It's health care.

JEAN-PHILIPPE COURTOIS: Hello everyone and a warm welcome to the Positive Leadership Podcast, a podcast that helps you grow as an individual, as a leader, and eventually as a global citizen. I'm Jean-Philippe Courtois. Today's conversation is with a technology leader who has consistently operated at an intersection of innovation scale, commercial and public service, where the stakes are not just technical, but deeply human. You will hear from someone who helped scale TwitterX in its early defining years, we later played a pivotal role at Uber, including launching the company's first self-driving car fleet. We stepped into the heart of US democracy, yes, as the first city of democratic national committee to help safeguard elections when cyber security threats were surging. And then we drove social impact technology and a marathon collective. In all these efforts, at the frontier of trustworthy open source AI, as the city of Mozilla, championing the conviction that internet and AI should belong to all of us. He's one of the most innovative, civic-minded, and positively driven technologies I got to know. So with real gratitude and excitement, please join me in welcoming my guest today, Rafi Krikorian.

RAFFI KRIKORIAN: Jean-Pierre, such a pleasure to be here. Thanks for having me.

JEAN-PHILIPPE COURTOIS: A very warm welcome, Rafi, to the Positive Leadership Podcast. Thank you so much again. So starting at the very beginning, Rafi, as you may know or not, I always love to start with your early days, your bringing, where you grew up, your family, your supporting environment, the people who matter to you, the kind of values and inspiration you got in your early life.

RAFFI KRIKORIAN: Yeah, I mean, maybe I'll even go before me. I mean, a lot of this I think about comes from, you know, my grandfathers in a bunch of ways of like where I am. So like my I'm Armenian Filipino. Like my father is Lebanese Syrian, Armenian. My mother is Filipino and my father's father was an orphan. Like he was an orphan from the Armenian genocide. And so he spent his, he spent and grew his own life as a military man, as a carpenter, as someone who just built things, but like took the values of raising a family to heart and raising my father, was his own entrepreneur, came to the States as a jeweler because he wanted to live the American dream and then had the ethics of like teaching his family to like work really hard, to study really hard, to do better than to do better than the previous generation. And on my mother's side, you know, my grandfather

JEAN-PHILIPPE COURTOIS: Mm.

RAFFI KRIKORIAN: was fortunate enough he created one of the largest radio networks in the Philippines. And so what he did, he started in a small town called Kagean Dioro. So Kagean Dioro is like on a beach and then there's a bunch of mountains and then there's farmland on the top of those mountains. So farmers would come down from the mountains every day to sell their things at the farmers market and then they would go a multi-hour trek to go home again. And so my grandfather started his radio network but

JEAN-PHILIPPE COURTOIS: Hmm.

JEAN-PHILIPPE COURTOIS: Yep.

RAFFI KRIKORIAN: One of the things he did pretty quickly was he let the farmers get on the radio before they went home. So they would broadcast being like, I'm so and so I'm heading home now. So their wives, their families would know that they're on their way home. It'll be a few hours, but they're on their way home. They're safe and they're making their way home. They can prepare dinner, whatever they want. And for me, between the two of those, like from my paternal grandfather, I think a lot about like work ethic, bringing oneself up, the importance of family.

JEAN-PHILIPPE COURTOIS: You

JEAN-PHILIPPE COURTOIS: Hmm.

JEAN-PHILIPPE COURTOIS: Yeah.

RAFFI KRIKORIAN: And for my maternal grandfather, I really think about like technology can be used for good. Like that, like as a clear example, like you can build a radio station for money or you can make a radio station that helps your community. So then for myself, like I grew up in suburban New York City. So I grew up maybe 20 miles out of New York City, but really interested in just tinkering. Like my mother would just buy me random things and I would be taking apart one of my

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: Mm-hmm.

JEAN-PHILIPPE COURTOIS: Hehehe.

RAFFI KRIKORIAN: One of my former memories is when we all used to have landline phones in the house, I was disassembling one of the landline phones and then a phone call came in which electrocuted me. I'm like, I need to unplug the phone before I... And so from there, I was fortunate that I went to a really good high school, I entered into all these science competitions, I was semi-bored, so I just kept on building as many things as I can. And then I landed at MIT, which was a great experience because it was my home.

JEAN-PHILIPPE COURTOIS: Yep. You

JEAN-PHILIPPE COURTOIS: Yeah.

RAFFI KRIKORIAN: In fact, I'm still connected to MIT and MIT Media Lab today. I serve on the visiting committee there. I'm on the alumni board because it was just such a formidable time to find your own people, to to like, start building things.

JEAN-PHILIPPE COURTOIS: Huh. Yeah. Yeah, maybe I have a of follow-up questions, maybe based on that, Rafi. It's fascinating to hear, in a way, the legacy you had from your grandfathers and family as builders, but also as people very committed to spread information, I guess, with the radio broadcasting and more. So it's kind of interesting enough, as we'll discuss later, AI and the way language is being broadcasted now through agents and machine.

RAFFI KRIKORIAN: Yeah.

JEAN-PHILIPPE COURTOIS: What does it mean for humans? So any particular sensitivity on that side, in the way, again, you grew up and talked about languages maybe from the Philippines or...

RAFFI KRIKORIAN: No, mean like my mother speaks four languages, my father speaks five, I speak two and a half. I mean like I think the human connection is like one of the most important things in my mind. like I've always like, know, one of the graduate programs I did for my first master's degree was in social media. this was before social networks, this was before all those things. like, it was in the world where like,

JEAN-PHILIPPE COURTOIS: Wow. Wow.

RAFFI KRIKORIAN: Nokia would have an anthropologist on their team just understand how cell phones were being used. Or we saw these interesting side effects of, know, at the MIT Media Lab were some of the first cyborgs, like people who literally wore like x86 motherboards duct taped to their back and like a whole screen in front of their face. But so like I've always thought about like that for me that connection just like how do people really talk to each other? What's the implications?

JEAN-PHILIPPE COURTOIS: Yeah.

RAFFI KRIKORIAN: of how we talk to each other. Like I feel like that's one of the through lines that come through in like the way I think about the world.

JEAN-PHILIPPE COURTOIS: Now going back actually to the choice you made to get into engineering MIT, why is that? I understand you were very curious as a kid, very interesting enough actually, same conversation with Kevin Scott, maybe you know Kevin, the CEO of Microsoft, friend of mine as well. And he was telling me on the same podcast that as a kid, he was also with a screwdriver, disassembling every piece of hardware in a house. Anyways, because of that, he came from rural.

RAFFI KRIKORIAN: Yeah.

JEAN-PHILIPPE COURTOIS: background as well and so rural areas, you know, he was very curious to understand how things get done and then get into engineering at one point. What about you? What decided you to get to?

RAFFI KRIKORIAN: Yep.

RAFFI KRIKORIAN: Yeah, mean, there was actually, was completely coincidental. mean, having parents who are immigrants who didn't really know the American college system. So like a lot of it is like I was just bumbling my way around and trying to figure it out. But my, the vice principal of my high school, I can't believe I remember this, like gave me a book called the MIT Media Lab. Like when I was a junior in high school. And that book was amazing. I mean, it described the place that lived 10 years in the future. It like always was 10 years in the future and how like their job effectively was to like figure out how that future should come to today. And so like at that point, they were talking about like, you know, 3D workstations and all these graphics capabilities. Like I even remember like when I first got to MIT, like the fact that there was a silicon graphics workstation that I could go play with was like the amazing thing.

JEAN-PHILIPPE COURTOIS: Mm-hmm.

JEAN-PHILIPPE COURTOIS: course.

RAFFI KRIKORIAN: But they're already starting to describe at that point, like what the future of the web might look like when most of us were still up on dial-up modems. so like, like understanding there was a place that was not just about building, but imagining was like the thing that drew me of just like, I want to go there. Like, like I want to go to a place where people are imagining the future. like, you know, I also remember there was this Australian TV show I used to watch when I was a young child called Beyond 2000.

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: Yes, the future.

JEAN-PHILIPPE COURTOIS: Okay.

RAFFI KRIKORIAN: was on some cable TV show, cable TV network, but it was always like, are the futuristic things going on now? And like when I was in Montreal as a young child, this was in the early 90s, my father took me, there was an exhibit in Montreal called Images de Futur. Like it was like, it was like a whole exhibit of like, what would the world look like in the year 2000? So like I've always been fascinated by like a, could it be? And so like my job, feel these days is like to help shape it.

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: shaping the future.

RAFFI KRIKORIAN: Don't make it the inevitable one, but like what could it be is how I think about it.

JEAN-PHILIPPE COURTOIS: It's about rebuilding a desirable future in many ways, right? Yes. We'll discuss that more, of course, in the rest of discussion. So, you know, we'll talk about different roles, we different missions. One is also very exciting, is when you became vice president of engineering on a major role scaling Twitter's platform in its hypergross years.

RAFFI KRIKORIAN: Correct, correct. One that we all could thrive in, like in some way.

JEAN-PHILIPPE COURTOIS: And I think you asked at the time to make these global service stable and available all the time, of course, the 99, 999 % that we all believe in that, of course, as hyper cloud providers, an example. Can you share a story of one of those farewell days when I think Twitter crashed and leadership lesson from building Twitter's engineering culture and infrastructure at scale? What it took? What are the painful lessons you got?

RAFFI KRIKORIAN: Yeah.

RAFFI KRIKORIAN: yeah.

JEAN-PHILIPPE COURTOIS: And I know how much it helps you today in every role you took.

RAFFI KRIKORIAN: Yeah, I mean like the most visceral thing for me frankly was the World Cup when it was in South Africa. I was maybe a year into Twitter at that point. So we were in our second or third office. It was still like the early days. And I just would remember like one...

JEAN-PHILIPPE COURTOIS: for you.

RAFFI KRIKORIAN: All the engineering team would wake up at odd hours in the morning because the South Africa time zone compared to San Francisco time zone was so far. But then two, you would sit there praying no one scored a goal. Because like the instant someone scored a goal, the entire country would start tweeting and Twitter would just fall over. And then we're all like racing to get this thing back up and running. So like I remember like the Japan versus someone game. And I was just like, they scored again. And then you saw the tweet speak a spike.

JEAN-PHILIPPE COURTOIS: Of

JEAN-PHILIPPE COURTOIS: Pshh!

JEAN-PHILIPPE COURTOIS: Wow.

JEAN-PHILIPPE COURTOIS: Yeah.

RAFFI KRIKORIAN: And then Failwell started showing up all over the site. And so like for me, that was like the most visceral experience of just like, I mean, in a good way, this service matters, right? Like people are celebrating on world scale what's going on, but also people, know, on the non-peak days, people are celebrating on microscale. if you look in the tweet streams, you see people celebrating their quinceaneras. They see people celebrating all these things.

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: yeah. Yes.

RAFFI KRIKORIAN: And so like it's like a celebration machine at that point, like now it's become a vitriol machine. back then it was like a celebration machine. And so my job after that World Cup was to figure out how can this never happen again? Like how do we make sure this never happens again? And so like the long arc was, you know, me and maybe one or two other people, we locked ourselves in a room and we drew up a whole new infrastructure for what Twitter looks like.

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: Hmm.

RAFFI KRIKORIAN: We assembled a team, we pulled the rock stars from all over Twitter together and gave them the plan of like, this is what Twitter should look like. And we started building it. We started building it. We built it all in some combination of Java and Scala. The original Twitter was in Ruby on Rails. We built it all based on the streaming fire hose of tweets, like how can we build a caching layer that would protect all of Twitter so that we can solve all these things. I would talk to the executive team about it. I never talked to the board about it. I would talk to the executive team about it. Like what progress is going. Like a lot of stuff was banked on me and my team solving this problem. We got to within a week, like quite literally a week of shipping this thing. And one of my senior most engineers, a guy named Steve Jensen came to me and was just like, I don't think this is a good idea. Like, I just don't think this is a good idea.

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: Hmph.

RAFFI KRIKORIAN: And was just like, what do you mean? Like we've been spending months on it. We had plans. Like we built it. Like, what do you mean this is not a good idea? And then he's just like, well, can you go talk to Aria? Can you go talk to John Colocchi? Can you talk to all these folks? But like, I don't think it's a good idea. And then the story started poking into it and like the team had built it. We're ready to ship, but they're all like, this is not a good idea. Like we, we don't think we built the right thing. And so when we debugged it, you know,

JEAN-PHILIPPE COURTOIS: Yep.

RAFFI KRIKORIAN: We found lots of problems. Like we found lots of problems and actually thankfully we never shipped that first version. We called it Woodstar. Everything is named after a bird at Twitter. Woodstar is a kind of woodpecker. So with the first version of Woodstar, are glad we they were like, no, we really shouldn't ship this thing. Like it has these it has a different set of failure modes, but it's not immune to failure. All these things are going to go wrong. And so then the question really came down to us. It's like.

JEAN-PHILIPPE COURTOIS: Yep. Mm-hmm.

JEAN-PHILIPPE COURTOIS: Yep.

RAFFI KRIKORIAN: Why did it take until a week before shipping for you guys to say anything? And it just really came down to, it's like, well, Rafi, you had a plan. We were following your plan. And I was just like, fuck. Like that, that's, that was wrong. And like, honestly, like, this is the thing I take from that, from every day of my career.

JEAN-PHILIPPE COURTOIS: Yes

JEAN-PHILIPPE COURTOIS: You're the boss! Yeah. Yeah.

RAFFI KRIKORIAN: And then we reversed it. was just like I had to do it massive me a culpa in front of the company in front of the executive team. I was just like we've been working for nine months on this thing and we're not going to ship it. And so like I did this massive me a culpa and then but my team then rose to the challenge. They're like OK if we don't have to ship it we can keep this we can keep this we can trash that we rearchitect this and do all these things and then we ship something that The World Cup happened and no one noticed. Like four years later, no one noticed. And like for me, this is a matter of just like, my job is not actually to be the architect. My job is to like provide the framework, provide the vision, provide the protection, the funding, the people, and to make the environment so all these people can do the best job of their lives. And these people did it. And they're now at some of the...

JEAN-PHILIPPE COURTOIS: Wow. Yeah. Hmm.

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: their best job.

RAFFI KRIKORIAN: best companies in the world leading engineering there. And so like, am so proud that that one engineer was able to be like, Raffi, we can't do this because that set a cascade of events that meant that we did the right thing. And I think the better thing.

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: You

JEAN-PHILIPPE COURTOIS: Yeah, I think it's a wonderful, nugget of wisdom, Rafi, and I'm sure you've been building on that nugget of wisdom in your career, Either way, you started.

RAFFI KRIKORIAN: You know, I often go into the rooms and say, like anyone who works for me is tired of me saying nine out of 10 of my ideas are bad, but I'm just like, but that's an invitation for you to tell me. No, that's a bad idea. I'll give you a better one.

JEAN-PHILIPPE COURTOIS: Please, let me know. Yes, yes, yes, yes, super. So after helping scale again a social media platform, you made also a very deliberate leap into a world where the consequences are physical and life critical. Moving from Twitter to Uber's Advanced Technologies Group, I think was not just a change of company. It was a shift of industry. pace of risk, profile, accountability, and the learning curve, I think which was very new to you, From automotive system and sensors to real-time perception, safety and engineering must have been just immense, right? So what drives your decision to make that leap from social media technology into this automotive world of autonomy, which I think you didn't know anything about, if I'm not mistaken.

RAFFI KRIKORIAN: No, nothing. No, I mean, I think it was two things really. One is like, it is a constant drive to try to make the world better. And like, look, I've been naive and I'll admit that. Like I worked really hard because I bought into the vision of making the world flat when it came to information spreading. so like Twitter in the beginning was amazing. Twitter today is not. And so like, and I worked really hard to make autonomous vehicles because you know,

JEAN-PHILIPPE COURTOIS: Uh-huh.

RAFFI KRIKORIAN: A large number of people die on the roads every year. large amount of geographic space in the city is taken up by parking. Imagine if we could just transform those things. But naive to all the side effects that might happen, but I admit that. One, it was a drive to make the world better. But two, it's a thing I tell anyone who asks for advice. I think what you should do is maximize learning for yourself.

JEAN-PHILIPPE COURTOIS: Totally.

JEAN-PHILIPPE COURTOIS: Yes. Yep. Yep.

RAFFI KRIKORIAN: And so I'm just like, well, I don't know anything about robotics, but like I want to learn. Like if you're giving me an opportunity that I can lead and learn at the same time, like who would turn that down? But you know, the one nice thing that came out of it is the problem I had at Twitter, which is like, I felt I literally could do everyone's job. Maybe not as well as them, but I literally could do it. I could not do any of these people's jobs.

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: Aha. Yeah. Any of those jobs.

RAFFI KRIKORIAN: And so like it was it forced me to have a very different leadership mentality of just like how do you lead people who are quite literally better than you at everything in this field kind

JEAN-PHILIPPE COURTOIS: Yeah. So can you walk us through maybe a little bit about that steep learning curve? You know, as you are ignorant of almost everything, I'm exaggerating, obviously, you know, and you've got smart people, and yet you got to learn something as well to direct your people, to lead them. What did you do? What was your learning system? What was your whatever it was?

RAFFI KRIKORIAN: No, no, true.

RAFFI KRIKORIAN: Yeah.

JEAN-PHILIPPE COURTOIS: to learn, absorb, question, and challenge as well the team, of course.

RAFFI KRIKORIAN: No, I mean, like in the beginning, it was a lot of humility and a lot of reading. Like, you know, I just read as much as I could, but then I would also be very like I would be I would try to be disarming about it. I try to approach every single room with like a for myself, be comfortable being the dumbest person in the room. And then. And then the second, though, like show up and just tell people I'm the dumbest person here. Like, but we are going to work this through. Like, I bring a different set of perspectives. And then eventually through those interactions, like I spent a lot of time with this amazing researcher named Drew Bagnell, who was a professor at Carnegie Mellon. He now runs a lot of autonomy at Aurora. They do self-driving trucks. He taught me so much just by me, by being patient with me, asking him lots of questions. Like whenever something came up and like, Drew, you have to explain that to me.

JEAN-PHILIPPE COURTOIS: Hmm.

JEAN-PHILIPPE COURTOIS: Thank

RAFFI KRIKORIAN: But then I also figured out what I could add value in those world like you know I think that like a lot of at that time I think robotics is in a very different place a decade later but at that time roboticists weren't product people like there were researchers like there was an R &D spectrum and like robotics was way more in the R than the D building a website is a D building a robot is more like an R.

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: Researchers. Yeah. Yeah. Yeah.

RAFFI KRIKORIAN: And so like I could bring a lot of the product mentality, a lot of like, wait, how is a user going to actually use this? Like what's the interface going to go look like? Like what does this milestone actually translate to for business objectives? Like, you know, one of our first versions of the car didn't actually truly have a motion control system. Like what it really did is like every single road in Pittsburgh kind of had an invisible line down it. And then the car would just like roll down that line as if it was on a track. And I was just like, okay, but we can't ship that. Like that can't be something like, because like there's too many, yeah, there's too many failure modes. It also doesn't make sense. Like it also doesn't tie to the way the marketing is going to work. Like how do we sell that? And so like, it's like those types of things that was able to actually help the team really think through and then like really make true milestones. Like

JEAN-PHILIPPE COURTOIS: Safe. Yeah.

JEAN-PHILIPPE COURTOIS: Yes.

RAFFI KRIKORIAN: I also remember the team built a car, like we built a car end to end and they were like, well, we're done. They didn't quite say that. And I was like, okay, let's drive it around the block. And it couldn't drive around the block. And so like, mean, like it could, but it couldn't. And so like, I'm like, okay, we're going to do a block. Then we're going to do a mile. Then we're going to do a neighborhood. Like, like help it, like figuring out how we actually stage all these things. Like I could find, I found my value.

JEAN-PHILIPPE COURTOIS: Wow. Yeah.

JEAN-PHILIPPE COURTOIS: One step at a time.

RAFFI KRIKORIAN: through asking questions, through learning, to try to piece together what was BS but was not BS kind of thing.

JEAN-PHILIPPE COURTOIS: Yeah. Oh, it's incredible. I picked one line you said, Rafi, which to me resonates a lot because it's very rare and I'm not criticizing at all all my engineers friends, but in many tech companies, you know, opening a room, entering a room and saying, I'm the dumbest person in a room as opposed to smartest in a room is very rare. So congrats on that humility because that's pretty rare.

RAFFI KRIKORIAN: Yeah. But like, I appreciate that. like, but I legitimately, I mean that first experience from Twitter is just like, my job is really to like get people to say all the words. Like don't hold anything, just say all the words and then we can figure it out at that point.

JEAN-PHILIPPE COURTOIS: Yes.

JEAN-PHILIPPE COURTOIS: Yeah. Yeah. Love it. In many ways, you are what in positive leadership we call a coach-like leader. Someone who is really there to elevate people, to make them grow by asking some structuring fundamental questions so they can go get the responses by themselves, by the way.

RAFFI KRIKORIAN: Hmm.

RAFFI KRIKORIAN: I mean, my whole management style, if you were to ask me, is some form of like an acknowledgement that we're not going to work together forever. And so like my job then is to figure out how to do a joint maximization of just like, are going to get the best out of you to make the company work, but we are also going to give you the best you can so you can work. So like, that's, feel like my, my, that's what my management style has become over the years is finding that joint maximization.

JEAN-PHILIPPE COURTOIS: Yes.

JEAN-PHILIPPE COURTOIS: Yes, I can feel. So I want to add another layer maybe on this onion. starting, I don't know if you started to peel or to open the onion. You know, what's striking to me, Rafi, is that leadership challenge is not only about learning fast, it's also learning responsibly. And I had a conversation recently with somebody, Naveena Singh. She's the CEO of Credo AI, a former Microsoft talent. And we talked about moving from abstract AI principles

RAFFI KRIKORIAN: Mm-hmm. Mm-hmm.

JEAN-PHILIPPE COURTOIS: to operational governance, right? On so-called responsible AI. Who owns the risk? What gets measured? What happens when the data tells you something uncomfortable? So tell us a story, because I think you've got a few stories about the self-driving program where you encounter the very uncomfortable safety or, know, or signal, ethics signal that really worried you a lot.

RAFFI KRIKORIAN: Yeah, I mean, I can tell a whole bunch of stories around that line. mean, one of them is just basically like a. You know, there's a difference between looking at something numerically and looking at something with your gut instinct and like you have to figure out what's the right way to balance both. So in the early days, you know, we would have a safety driver in the car. And the safety driver could like take over at any moment. You touch the wheel, the system disengages. There's a big red, literally a big red button. You hit the big red button, the system disengages. And we would measure, and this still measured this way in the industry, like the number of interventions you would do per mile. Like how many times in one mile of average driving does the safety driver have to take over? And really the goal shouldn't be interventions per mile. The goal should be miles between interventions. So like we were still in the early stage of interventions per mile. And like I remember like our one of the senior leaders at Uber, I won't necessarily name names, came to Pittsburgh and wanted to ride in the car. And like we drove down the road and

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: Mm-hmm.

RAFFI KRIKORIAN: And luckily that day we had basically no interventions. Like the safety driver didn't touch it for this road, for this mile or so that we were driving from the office. And the leader at the time was just like, you guys are being, you guys are being chicken shits. Like you guys are being too wimpy. Take the safety driver out. Like we never, you never touched the button. Like you clearly don't need him. And we're sitting here being like, the math doesn't make sense, man. Like the literally the math doesn't make sense. Like.

JEAN-PHILIPPE COURTOIS: We don't need this.

RAFFI KRIKORIAN: We were just on a lucky mile, but if we look at aggregate miles, but no, I mean, like he overrode us all. So like we sat in the back, him and me, with the safety driver sat in the US, the driver's side is on the left. He sat, the safety driver sat on the right. We had no one in the safety driver's seat and we drove for a mile and like we're sitting here nervously with our hand over the red button just in case. like this, like that is an example in my mind of just like,

JEAN-PHILIPPE COURTOIS: Wow.

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: Yeah.

RAFFI KRIKORIAN: bravado and numbers are coming into clash with each other. Just because it worked doesn't mean it's safe by any means. And so, I don't know if I've even told my wife this story because she would have murdered me when I got home. Like, what do mean you rode in the car with no one in the front? I wouldn't even let my kids in the car at that point. And so, that...

JEAN-PHILIPPE COURTOIS: Hmm.

JEAN-PHILIPPE COURTOIS: Yes. Yes.

JEAN-PHILIPPE COURTOIS: Cross.

RAFFI KRIKORIAN: It's like that kind of sense of just like, you've got to know when you look at the numbers and you got to know when you look at your gut. And this was clearly an example of like, you got to look at the numbers, not your gut.

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: Wow, that's a scary story. I think you had another one as well when it comes to diversity of people on the streets, right? Do you want to share that? Because I think ethically these days, mean, more than ever, of course, with the eye, there's a lot to say about the diversity of the models and what is it really? Yeah.

RAFFI KRIKORIAN: Yeah, no, that's a good one. That's yeah, that's also a good story. Like when we

RAFFI KRIKORIAN: Yeah. So when we first when we first started training the car, we I mean, we lived in Pittsburgh, Pennsylvania, Pittsburgh, Pennsylvania. For those who don't know, it's basically on the northeast of the US, not getting technical Midwest, but like it's cold. gets wintertime snow happen. So like you can't run a self-driving car program year round in a place like that. So what we did is in those type of times we sent what we call the data collection vehicle. So basically a self-driving car, but we wouldn't make it self-drive. We just drive down the road and record as much information as we can. We'd use it in training simulation, model generation, et cetera. And we drove it in the place in Phoenix, Arizona, which if you don't know the US, it's like very bright, very sunny. The roads are super wide, really easy to drive. And we just collected all the data that we could and trained models based on the data we drove there and brought it back to Pittsburgh. And the car mostly worked. you know, roads are mostly roads, et cetera, et cetera.

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: Yeah. Clip.

RAFFI KRIKORIAN: But one critical problem is that there turns out to be very few African Americans that walk the roads in Phoenix, Arizona. And so our training models for training, the way people look, pedestrians look, wasn't indicative of what it looked like in Pittsburgh, Pennsylvania, where there are more dark-skinned people walking around on the streets.

JEAN-PHILIPPE COURTOIS: Phoenix, yeah.

JEAN-PHILIPPE COURTOIS: Yep.

RAFFI KRIKORIAN: And so it's not that the car hit any of them, but we would just be misclassifying them all the time. And for me, that's also a really good example of just like, you also have to understand your operating environment. You also have to understand diversity. have to operate, understand all these things that like truly make up real world experience. Like I think about this all the time when it comes to AI right now of just like, do we have

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: Yes. Yes.

RAFFI KRIKORIAN: Do we have the right representative data sets in this thing that we can actually make decisions for entirety of humanity? You know, I might actually be coming to the opinion that like we have too much diversity, it's impossible to cram into one thing. So like we have to think about these questions of just like what does that truly mean, especially in these worlds.

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: Yeah, think it's a huge debate, actually, very important, but we'll discuss again later. I'm adding more discussion to come, particularly with the Global South. I I think to me, which I've been traveling my life in many of those countries, that's a big question in terms of the way those models are working or not and are relevant or not. They will get there. So moving on to your career, Rafin, 2017, after the tumultuous

RAFFI KRIKORIAN: Yep.

JEAN-PHILIPPE COURTOIS: 2016 election in the US for listeners. the way, our listeners are global. There are many in US, but they were also in Africa, in Asia, in Europe, everywhere. You made a dramatic pivot. You left Silicon Valley to become the CTO for the Democratic National Committee, the first in its history. Your move reminds me, actually, a story I had of another guest on a podcast. don't know if you know her, Reshma Sojani. She's the founder of Girls Who Code. a pretty very nice NGO in the US. And also she's a leading voice on courage, rebuilding careers that she shared with us on the podcast. And she described with us the moment where she left a comfortable corporate pass to run for public office. At the first time, she did something truly brave, very bold, and it unlocked her purpose. She was not elected, but so what? And so, of course, you are not yet, I think, going for election.

RAFFI KRIKORIAN: Yeah.

JEAN-PHILIPPE COURTOIS: Maybe you, well, who knows exactly. But I love to understand in your case, Rafi, what inspired you, what inspired that leap that you did? And what are the moments that made you say, have to help secure our democracy?

RAFFI KRIKORIAN: One day.

RAFFI KRIKORIAN: I mean.

RAFFI KRIKORIAN: I mean, the 2016 elections for me was a major disappointment in America. Like I'll wear my politics on my sleeve for a second. And so I was trying to figure out what are the things I could do ever since that election night in 2016. And I didn't quite figure it out. Like I just didn't know, I didn't come from the space. Like I just didn't know anyone. I was fortunate enough that I was invited a few times. to join the Obama White House on different roles, but I never took any of them up on it because I was like government, whatever, like kind of thing. I have bigger things I'm working on. But then the real gut punch for me was I was sitting in a hotel room in California. I was living in Pittsburgh because that's where their self-driving team was. So I was in a hotel room in California because I visiting the Uber office on inauguration day in 2017, watching inauguration on the screen.

JEAN-PHILIPPE COURTOIS: Hmm. Yeah.

JEAN-PHILIPPE COURTOIS: Yep.

RAFFI KRIKORIAN: And at that moment, I quit. Like in my mind, I was just like, I have to leave this job and I have to figure it out. Even if leaving, even if I didn't know what it was, the act of leaving would make me figure that out. So I left. It took me a few months to unwind myself and get myself out the door. But I had mentally at that point decided I needed to leave. And for me, it was just trying to figure out like, was there a place where a technical person could help? And, know, thankfully or not, thankfully, there's all these stories about the social media machine, all the data stuff that the Trump campaign did as part of the 2016 election. So there's clearly an opening in my mind. It just like, I might be an alien to these folks, but I like I think they should find me valuable. And I even remember in my one of my first conversations I had was I figured out how to talk to the chairman of the DNC Democrat National Committee was a guy named

JEAN-PHILIPPE COURTOIS: Hmm.

JEAN-PHILIPPE COURTOIS: You

RAFFI KRIKORIAN: secretary, who is a secretary Tom Perez. So I spoke to Tom and I remember like him having his first conversation with me being like, well, you're going to solve my cyber problem, right? Because you remember the Russians had hacked the DNC and I was like, yes, but not only that, like solving your cyber problem was a path to not losing. It wasn't the path to winning. And so like, so trying to understand that world and trying to understand how I could even talk to these

JEAN-PHILIPPE COURTOIS: Yes, yeah, yeah. Hahaha! Yes, to win, yes.

RAFFI KRIKORIAN: people was like the biggest thing is just like I am going to be an alien. Like I'm I'm used to being the the big fish in the room. But like I went into a world where like the tech team is like knocking on a door being like, can you let us in? Like like maybe can we just sit in the corner? Like like going from that big dog position to this position was very interesting. But it was it was literally that moment. It was watching the President Trump get inaugurated. was like I have to figure out something. Yeah.

JEAN-PHILIPPE COURTOIS: To help you? Yes. to do something. So what's amazing is actually the way you talk about it. You kind of left the job without having a new job in that new space. So it was a big leap of faith. And then you enter. I'm sure she was very, very nice. Yes, she is probably. And I love to understand the way you approach leading through that cyber security crisis, fusing in a way your Silicon Valley culture, right?

RAFFI KRIKORIAN: Yeah.

RAFFI KRIKORIAN: I have a very kind wife.

JEAN-PHILIPPE COURTOIS: with politics, which is a very different world. I never got into politics, but I met with many politicians, leaders in my life as well. So can you describe maybe one initiative that you spearheaded at a DNC that you are the most proud of and where you could actually blend that culture, those cultures together to make sense of it?

RAFFI KRIKORIAN: Well, mean, like, but I can talk about projects all day long, but I think you hit exactly on the thing, which is the strategy I came in with is how do we blend these cultures together? And so like, you know, I hired this amazing woman as my deputy, Lindsay Chou Cortez. She was a veteran of the campaign space. She's done all, she worked inside labor unions as a technologist. did campaigns as a technologist. She did data analysis. So like a I brought her on board. I spent so much time trying to convince her. And then I remember the day she took the job. And I was like, oh, thank God. So I spent all this time trying to convince her to be my deputy. And so we tried to portray a leadership team, which was that blend of we have people who know the campaigns, who know politics, and we have the aliens. So the people who are like, we know all this other stuff, but we're open and willing to learn. And so I hired. a 40 person team, which seemed small, but it was the largest single team at the DNC was the technology team when I was there. That was this blend of Silicon Valley technologists, like people from Twitter, people from Facebook, people from all from Google, from all around the Valley. And then we brought all these veterans from the campaign space, people who've done tech and campaigns like the Barack Obama campaign, the Secretary Clinton campaign had technologists. We brought some of them in and we brought campaign folks who were like the users of technology but not the developers of technology. So this 40 person team was like a very mixed family. And so like our job was like we had to figure out how talk amongst ourselves. Then if we can talk amongst ourselves then we could talk to everyone else. So that was really the strategy and like.

JEAN-PHILIPPE COURTOIS: Hmm.

RAFFI KRIKORIAN: Because then we were able to take on pretty contentious things like, for example, the way the voter file works in America, you know, the voter files, the list of all registered voters, and you need to know who the registered voters are because that's who you want to spend time talking to. Like, it's not not worth it talking to people who can't vote or who aren't registered to vote is every single one of those voter files is housed and is the purview of a state. So the state of New York, the state of California, the state of Arkansas.

JEAN-PHILIPPE COURTOIS: Such as, yeah.

JEAN-PHILIPPE COURTOIS: Yes.

RAFFI KRIKORIAN: have their own voter files. But the problem with that is that they're all in various states of assembly, not assembly, various states of other metadata, not metadata. Like it's very helpful to know for a particular voter, like it turned out, this is years ago, it's not as true anymore, but it's very helpful to know what magazines does someone subscribe to? Because like if they subscribe to,

JEAN-PHILIPPE COURTOIS: Yeah.

RAFFI KRIKORIAN: a hunting magazine, it probably tells you something very different than if they subscribe to a quilting magazine. And, like all the different states would not have metadata, would not have all these things. So we worked really hard to build a pipeline that would take voter files from different states, deduplicate them, mesh them together, annotate them with all this other metadata that you could then use for targeting voter outreach, et cetera.

JEAN-PHILIPPE COURTOIS: Hmm. Yep. Yes.

RAFFI KRIKORIAN: and provide that as a service back to the state parties. But, know, state parties have their own processes, have their own business plans. have all these different things. So like doing that in a way that wouldn't trip all the trip lines for a state party was actually quite hard. Like, and also it turns out that state parties use this as a little bit of power and leverage because if you were a campaign, you need that data. So you got to play ball with the state party. So

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: Gross.

RAFFI KRIKORIAN: Trying to navigate the politics of all that to just even build the right technology we needed was a huge challenge and that blended team strategy was the only way we got that done. I had a woman on my team who was a veteran of the Texas State Party. If I didn't have her, there was no way the Texas State Party would have played ball. They played ball because they had a woman on the inside, how they perceived it.

JEAN-PHILIPPE COURTOIS: Yeah.

RAFFI KRIKORIAN: And so like figuring out all those nuances, it's like, it's an engineer challenge unto itself.

JEAN-PHILIPPE COURTOIS: Yes, it is. And I guess because of that, you build some legacy and some capacity, which is still existing at DNC. if you were fast forwarding to today, next elections in the US, we could have municipal elections in my home country in France and then others presidential elections in a year. And we've got a national database of all the citizens. It's actually nationally managed.

RAFFI KRIKORIAN: Correct.

JEAN-PHILIPPE COURTOIS: But listening always to the politics in the US and understanding actually the underlying system you just described, Raf, is like, wow. OK, now I understand why there could be some controversy about the ballots at the borders, right? And so if you fast forward all that today in 2026, given as well the scale of fake data, of AI.

RAFFI KRIKORIAN: Yeah. Yeah, white gets crazy.

JEAN-PHILIPPE COURTOIS: manipulation coming from abroad and from within the country. mean, in all of our countries, not just the US, same here in France. It's, mean, technology should play a critical role in preserving democracy, isn't it?

RAFFI KRIKORIAN: I 100 % agree, but technology also is the problem, in a lot of ways, to your point earlier. When I was doing this, we had an amazing security person on staff, a gentleman named Bob Lord. He used to be the head of security at Yahoo when they were hacked at Twitter, et cetera. And what we stood up was a disinformation monitoring program. We wanted to get a disinformation weather report every day. This is before Twitter got even as bad as it is today.

JEAN-PHILIPPE COURTOIS: I agree with you.

RAFFI KRIKORIAN: So we would get reports from the field of just like, hey, there's this piece of content floating around that we need to figure out what to do with. So we would do rapid response of like having to brief the campaign, but also then we'd have armies of volunteers who would all hit the report button to try to trigger a signal on the platforms. like trying to figure out not just how to use the technology, but to be good users and stewards of it was also such an important part of that job.

JEAN-PHILIPPE COURTOIS: Yeah. So let's keep moving on because after helping at DNC, you move to another very interesting organization I found as I was doing some kind of research, Emerson Collective in 2019. And you'll talk about Emerson, but my understanding is Emerson is a very different context, a social change organization, working, education, climate, health, integration, and even more. And again, you move into an unknown territory with excitement. So what prompted that move again to get into again, okay, I have to learn everything from scratch all over again. And how did you reinvent your leadership focus again with a very different audience or audiences that you have to deal with?

RAFFI KRIKORIAN: Yeah.

RAFFI KRIKORIAN: Yeah, I think learning is that phrase again for me. Emerson Collective is this wonderful place. It was founded by a woman named Lorene Powell Jobs. she's, like you said, education, immigration, health, and the environment are the top four focuses. But really, there's a lot of the social good, nonprofit sector, mission-driven sector is lot of the work they do. And so for me, it was like a...

JEAN-PHILIPPE COURTOIS: For sure.

JEAN-PHILIPPE COURTOIS: Yeah.

RAFFI KRIKORIAN: You know, I've always been interested in it, but I was playing in a very particular corner when I was at the DNC. And when Laurene and her chief of staff approached me to potentially come and work there when I was leaving the DNC, it was a chance to sort of like take what I did in just the election space and potentially do it again at scale across an entire sector, entire across nonprofit sector. honestly, like while I started, I remember You know, Emerson Collective does these things called a demo day once a year where they put their top 10, top 10 is the wrong word. They curate maybe five to 10 entrepreneurs, social entrepreneurs to go on stage and to show off the work they've done. So, you know, I was a little on the fence of joining, but I showed up at their demo day and it was jaw dropping. Like it was like a whole other group of people. I've spent all this time in the capitalism world, I've spent all this time in the election world with politicians, but then to see all these social entrepreneurs who are doing just as amazing work, just with a different funding model effectively, was jaw-dropping. And then when I realized that the number of technologists, at least in that time, the space, was like you could count on one hand, I was like, well, it's clear I should go figure this out too.

JEAN-PHILIPPE COURTOIS: Yeah. Yeah. Yeah.

JEAN-PHILIPPE COURTOIS: Absolutely.

JEAN-PHILIPPE COURTOIS: That's right.

RAFFI KRIKORIAN: A, it would be fun and exciting and B, technology could make a real big difference. So the bet I was placed with the head of philanthropy at Emerson was, do we think the world is better when you have technologists who are learning something about social good or do you teach social good people something about technology?

JEAN-PHILIPPE COURTOIS: Mm.

JEAN-PHILIPPE COURTOIS: Yeah. Hmm.

RAFFI KRIKORIAN: And so like that's been the bet that I've been trying to play for all this time. And it was a lot of fun.

JEAN-PHILIPPE COURTOIS: Yeah. So I know you launched many initiatives and I heard about, of course, on AI for nonprofits, which really resonates with me, Rafi, because as I share with you, I'm running an NGO in my home country. I'm spending a lot of time with philanthropists. And just like you said, tech is really under invested in philanthropy, which is dramatic anyway. And so, for example, I think you build Ralph, an internal AI chat bot.

RAFFI KRIKORIAN: Yeah.

JEAN-PHILIPPE COURTOIS: that helps Emerson staff query their data. And you could develop a chatbot to assist immigration lawyers, something pretty, I would say, that many people discuss these days in helping asylum seekers. So how did you decide again which problems to tackle with AI? And maybe can you expand on one of those tools and what it did to the society? And obviously, the kind of the pride you got out of that.

RAFFI KRIKORIAN: Yeah, actually, let me back up for a second. wasn't even, AI was this thing that showed up later in the game. The things that you could do with technology, the bar was actually quite low and you could still make a huge difference. So there was this one organization that works with asylum seekers in the United States, help them find jobs, et cetera. And when we, they're an amazing group of folks, so, so heartwarming. But when we talked to them,

JEAN-PHILIPPE COURTOIS: Right.

RAFFI KRIKORIAN: And we asked them, how many people are in your service? These numbers are not exactly right. I don't remember. But let's say they're like, we have 20,000 people who signed up to work with us. I was like, OK, but what's the total addressable market? How many asylum seekers are there in the US? And the number is way bigger than 20,000. And so I'm like, why? Why are we capped? How can we help? And so my team spent time just understanding their system, et cetera. Paper. They were just using paper for everything.

JEAN-PHILIPPE COURTOIS: Yeah. Of course.

JEAN-PHILIPPE COURTOIS: Paper.

RAFFI KRIKORIAN: And so we did a week long sprint. We built a bunch of stuff on Google Sheets and Google Forms, like not even anything complicated by technology standards. And then we checked back in a few weeks later and they were onboarding 20,000 a week just because we just completely changed their process. So like those are the opportunities and like honestly, those are feel good moments for technologists. Like you could do something small that has very high leverage. It's such a, it's like a great thing that you can go do.

JEAN-PHILIPPE COURTOIS: Yeah.

RAFFI KRIKORIAN: But then you're right, AI did come around at some point. And so like we felt the most responsible thing to do was to start experimenting, was to understand what was possible. And so one of the things that we figured out was there are lots of people at the time when they were crossing the Southern border and a aid groups are trying to help them. There are too many regulations. There are too many opportunities to help. There are too many programs.

JEAN-PHILIPPE COURTOIS: Mm.

JEAN-PHILIPPE COURTOIS: Yeah.

RAFFI KRIKORIAN: could we ingest that into a chatbot that allowed a legal aid worker to rapidly query given a certain person who's at their door. So we built that and that actually had pretty good effects. It actually streamlined a lot of the efforts on what was going on. But for me, that still reminded me that there are still all these very low hanging fruits that you can just do to make things better. Like AI sounds great.

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: Yes. Yes.

RAFFI KRIKORIAN: But it wasn't the solution for everything. Like literally just doing a digital transformation to not paper could still be a way bigger impact than to build the sexy chatbot thing. It's kind of the lesson I took from that.

JEAN-PHILIPPE COURTOIS: Yeah, absolutely. Yes. No, it's really a... To me, it's a very exciting place, by the way, and for many engineers listening to the podcast as well, I mean, I would really encourage you to go and spend time with some of those NGOs and profits in your city, in your state, in your countries. There's so much you can do with little help, honestly, with a big leverage.

RAFFI KRIKORIAN: Yeah, I mean, I'll give you another example. I am fortunate to be on the board of an organization called Medic. And what we do is we build open source software to help community health workers in Africa, in Asia, and other parts of the world. And community health workers are volunteers from a community that walk around and just check up to see how everyone is doing. And if there's something wrong, they can record it, maybe refer you to a doctor, et cetera. again, it's a digital transformation thing like

JEAN-PHILIPPE COURTOIS: is

RAFFI KRIKORIAN: The sexy thing to do would be like, let's give them an iPhone app and we can do like a tricorder and scan them and like figure out all this stuff. But really what they needed was they need effectively a CRM that works on a mobile phone in an offline way so they can just record things. Because what they used to do is carry a book on their back and just write all these notes down. And so like just like just helping them meet them where they are.

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: Right. Easy. Take notes. Yeah.

RAFFI KRIKORIAN: with technology is like the biggest opportunity.

JEAN-PHILIPPE COURTOIS: Yes. No, it's great. You just opened the borders right now, or are you going outside of the US? Because I wanted to go there, actually, in my next question. So let's zoom out for a moment, because what strikes me is that what is called AI for Good is no longer a niche idea, I think. It's becoming a global movement with institutional backbone. So I know these days the UN is very criticized, but I would say the UN ecosystem through the ITU's AI for Good initiative has been convening garments, researchers, companies, NGOs, to try to push out to all the sustainable development goals, by the way, because we see that those sustainable development goals for this, as you may know, and to move again from inspiring pilots to solution that can scale. you know, I don't know if you check that work actually from the ITUs and the UN, Rafi, but what do you think that they get right or not? Or what do think we should do? globally as humanity to take the next step again for AI for good and what it means for you actually.

RAFFI KRIKORIAN: Yeah, I mean, for me, the phrase AI for good is actually bigger than a bunch of the SDG, which are very important. Don't get me wrong. And I think that like, I think a lot of these questions can be answered through the lens of like, are as technology helping us or as technology not working on our behalf. And so like I think about the SDGs through the lens of like, what are really the metrics that we really want to optimize for? Like,

JEAN-PHILIPPE COURTOIS: Yep, yep, yep.

RAFFI KRIKORIAN: Do we want to optimize for number of lives saved through AI? Do we want to optimize for like, how do we want to phrase all the different things so that we can then storytell to the world, to the researchers, to the implementers, like the things that we actually want to push forward. And I think that like the way that STGs are framed, which are great, don't really tie the dots close enough to like, for the implementers to go do something. And this is a similar problem.

JEAN-PHILIPPE COURTOIS: Yes.

RAFFI KRIKORIAN: I used to face when in Silicon Valley, like, you when I was hiring for the DNC, it was very easy to explain to a person why they should work on this. Like, you see it every day on your news and I can give you a direct sense of like, if you join my team, we can work on this problem right now. But when I moved to Emerson Collective, I had a harder time doing that. And like, you know, I would go talk to people at the big five companies all the time to be like,

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: ways you can do. Yes, yes.

RAFFI KRIKORIAN: you need to work on education, you need to work on immigration. And their answer to me was like, along the lines of like, Rafi, we're not jerks. Like, of course we would want to help, but you're not telling us how. Like, you're not giving us a clear pathway that connects the dots. And I feel the same way around the SDGs of just like, I think we can mobilize the entire world of engineers if we can give them clear pathways.

JEAN-PHILIPPE COURTOIS: Yes.

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: Hmm.

JEAN-PHILIPPE COURTOIS: but actionable plans. Yes. Yes.

RAFFI KRIKORIAN: on actually how to go do it that's actionable and then they can feel like they made a difference in that work. Now there's a bigger question for me and this AI for good question, which is a little bit of like a how do we use these tools? Like are we owners? Are we renters? Like who owns the data? Do I control the system? Does someone else control the system? So there are these bigger questions I think which also. we need to be grappling with because especially if you race into a future where machines are making decisions for us all the time, how do we have trust in those decisions? Like if those machines are now attempting to solve the SDGs, which would be fantastic, like I would be all over it. I just want to know that they're not solving it for 10 people at the expense of a hundred or like how are they making those trade-offs? Like I think those are, those are actually like

JEAN-PHILIPPE COURTOIS: Yes.

JEAN-PHILIPPE COURTOIS: Right.

RAFFI KRIKORIAN: the philosophical and important questions of like how is this like machine to human decision going to go work?

JEAN-PHILIPPE COURTOIS: Yeah. I think I've got a new mission for you, Rafi. You should become the first ever city of the UN. Maybe we'll get in a better place for the world. I'm serious, actually. I'm serious. So let's get into your current role.

RAFFI KRIKORIAN: hahahaha That would be hilarious.

JEAN-PHILIPPE COURTOIS: inspiring as well. You you made this move that feels to me again, very deeply value driven. You joined Mozilla as CTO. For listeners who are not all tech, by the way, Mozilla is a unique origin story. It traces back to 1988 when Netscape open sourced its browser code to defend an open web. And today's structure is just as very distinctive, right? I there's a Mozilla foundation that safeguards the public in transmission. Well, the Mozilla Corporation, if you understand well, builds products like Firefox to fund and advance it. In that sense, it's a rare mission. It's a rare first tech mission, you could argue. Somewhat comparable, you could argue, to OpenAI because there's this kind of structure. And I we can debate that. Of non-profit on top of commercial. So what specifically first, Rafi, pulled you to our Mozilla? its mission open source DNA. And you know, why was that again, the right platform for you now to versus joining a Frontier lab or open AI, right? Any one of those companies today, why did you make that choice?

RAFFI KRIKORIAN: Yeah.

RAFFI KRIKORIAN: Yeah, no, that's a great question. mean, so I view Mozilla in a similar way to, like I mentioned Signal earlier, that like what we do is we build technology on behalf of people. Like, so in your story of Netscape open-searching the browser, which then became Firefox, which I hope some of your users use, like the real goal of Firefox became

JEAN-PHILIPPE COURTOIS: Yep, Yes.

RAFFI KRIKORIAN: to take on Microsoft when Microsoft wanted to control the web. So if you remember in the 90s with Internet Explorer and ActiveX and all these things, Microsoft, I mean, I used to build some ActiveX components, so I understand too. I mean, like, but I also was also a Firefox user at that time too, on my desktop Linux box and everything. Like,

JEAN-PHILIPPE COURTOIS: I was there on the other side to be... Yes, yes.

RAFFI KRIKORIAN: Microsoft is trying to exert power over the web and make the web an extension of the Microsoft ecosystem. And you you could buy a Microsoft web serving stack. You can do all the stuff on your server closet and on Microsoft environment, et cetera. And what Firefox did was to provide a credible alternative, which forced Microsoft to the table to talk about standards, to talk about interoperability, to make sure the web remained open because we had a that pushed the conversation. And so like I think about that moment a lot because like I feel like we are in the similar spot right now. So like you know the term user agent is usually used to describe your web browser like your web browser is your user agent on the open web. It protects your privacy. It blocks ads. It does all this stuff on your behalf and.

JEAN-PHILIPPE COURTOIS: Hmm.

JEAN-PHILIPPE COURTOIS: Yep.

RAFFI KRIKORIAN: The user agent we're starting to all turn to is no longer the browser, but these chat agents are these chat prompts or other things. But those user agents aren't on your side. They're on the side of someone else who has different incentives than you do. They're on the side of a company like chat GPTs talking about running or open ads talking about running ads inside it. like, which again makes you wonder about like the results that they're providing, like are they being incentivized in different ways or are they for me? And so like that I feel is actually one of the biggest questions that we all have to tackle with. Like our portal to information could potentially be owned by a company. And is that the right thing or should it be owned by me? So like I use this phrase owners, not renters. Like I think we should own our our our knowledge portal, not rent it from someone else. And so like that's why I think Mozilla for me was the right place. Like, yeah, I mean,

JEAN-PHILIPPE COURTOIS: Yep.

RAFFI KRIKORIAN: I'd like to think I could have gotten a job at the Frontier Labs. I don't know if I'm smart enough, but I'd like to think I could. But this is the place where I really want to build technology for people. Like I don't want seven AGI's in the world by the time all the big companies duke it out. I want a billion AGI's. I want seven billion AGI's. I want all of us to have our own AGI that works for me. And like that's I think what the fight, that's I think the fight of our.

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: Yes.

RAFFI KRIKORIAN: for the soul of our civilization that kind of looks like right now.

JEAN-PHILIPPE COURTOIS: Wow, that's a big statement. And indeed, I think it's a big important statement you made. I'd you to maybe to do two minutes kind of elevator speech as less technical as possible on the open source, AI open source stack, compared to the other commercial AI stack that everybody's using. And by doing that, Raphael, I'd you to basically explain why you think that

RAFFI KRIKORIAN: Yeah.

JEAN-PHILIPPE COURTOIS: that open source stack is not a threat to the sovereignty of countries and others necessarily. And it's so important in this new AI era for people.

RAFFI KRIKORIAN: Yeah. So there are lots of components that make up an AI stack, like everything from the compute that you'd hear about all the time, Nvidia, Nvidia this, Nvidia that, to the models, whether it be open AIs models or whether it be Anthropics models, to the data sets that are powering them, to the applications that are built on top of them. So if you think about say like a Google stack for a second, Google has its own dedicated hardware that's something they call the TPU, the Tensor Processing Unit. They have their own models, things like Gemini. They have their own data sets because they've scraped the web. Like they are the masters of knowing what's going on in the web. And then they build applications on top of it. So that verticalized stack is all under the control of Google. And look, I actually really like Google. I actually think Google is a good player. But like,

JEAN-PHILIPPE COURTOIS: Yeah!

RAFFI KRIKORIAN: If Google decides to make a one decision, it actually impacts all of us and we have no control of it because they control every part of that stack. And you know, those decisions could be anything from like the YouTube ranking algorithm or like small design decisions on how you and I interface with AI. It could be that when you talk to Gemini, Gemini is only returning YouTube links and as opposed to other links, like those small things are shaping the way that our world works.

JEAN-PHILIPPE COURTOIS: Yep.

RAFFI KRIKORIAN: in their benefit somewhat. I want to live in a world where you could build a stack of components that anyone could build, anyone can contribute to, and anyone can inspect to understand what's actually going on, how are decisions being made. And look, I'm not, I'm literally not saying, I'm not literally saying that my mother-in-law, who's a very lovely British woman, should be able to go in and read the code, but the fact that someone

JEAN-PHILIPPE COURTOIS: Mm-hmm. Yeah.

RAFFI KRIKORIAN: can read the code I think is incredibly important. So like I want to live in a world where like the models are you know what data went into a model and you know how they trained it and what they trained it to do. So like in the case of my attempt, my Arizona example of driverless cars, I want to know that they've like properly have pictures of people like me in the data set. So they'll misclassify me as part of the situation. I want to know.

JEAN-PHILIPPE COURTOIS: Someone is able to do it.

JEAN-PHILIPPE COURTOIS: Yep.

JEAN-PHILIPPE COURTOIS: Hmm. Yeah. Yeah. Yeah.

RAFFI KRIKORIAN: how those decisions are being made. So I want to be able to understand when it's when in model takes a particular action. I want to be able to trace it back and understand why was that action taken. So I can also know that no one's benefiting like it's actually the right thing as opposed to someone's making money on the side. Like it's like when you go on a when you go on a tour guide at operation and they bring you to a restaurant are they getting a kickback from the restaurant? Like how would you know? And so like I want I want to understand that on the model side. And then all the way up into the user interface, like I want to know that this user interface is actually trying to answer a question for me as opposed to keeping me engaged for as long as it can because like these systems currently make money because you stay there. I want a system that is incentivized to give me the right answer, not keep me engaged. And you know, that's the same problem that social media has and stuff like that. So I feel like the

JEAN-PHILIPPE COURTOIS: Engaged, busy. Yeah.

JEAN-PHILIPPE COURTOIS: Attention. Yeah. Yeah. cross.

RAFFI KRIKORIAN: the most logical attack to those problems are to open up every single layer of the stack. And so like there is a vibrant ecosystem globally of people building parts of the stack together. I view my job of like, how do we stitch it all together? How do we build applications or how do we enable every developer to build applications on this open version of the stack and make it a credible alternative?

JEAN-PHILIPPE COURTOIS: Yep.

JEAN-PHILIPPE COURTOIS: huh. Yeah.

JEAN-PHILIPPE COURTOIS: Yes.

JEAN-PHILIPPE COURTOIS: Mmm.

RAFFI KRIKORIAN: to building on a closed version of the stack. And now I'm not even saying everyone needs to build open, but the fact that it exists is going to be enough to move the industry.

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: So I think it's a super important topic. Now, if we were to expand that discussion, Rafi, with the geopolitics of the world in which we are living, many Europe, you in the US, some of the folks in China, others in Africa, there's a variety of models in US. As an example, Lama 3.1 from Meta in China. There's been a lot of news as well from Quen, DeepSic, of course.

RAFFI KRIKORIAN: Yeah.

JEAN-PHILIPPE COURTOIS: with a fast open source ecosystem developing. In Europe, we've got Mistral, my home country as well. If you were, maybe it's another job for you, the CTO of the US administration, yes, and you could whisper in the presidents here to Mr. Trump, what would you tell him and how should the US balance open source, open models with national security and export controls? I know that's a big question.

RAFFI KRIKORIAN: you

RAFFI KRIKORIAN: Yeah, no, it is a very big question. I mean, like, I do think there is a need for national security concerns. So like, I'm not going to fault anyone for building closed systems in the sense of like, we're going to need to do national defense and and likes like that. You know, I might have ethical issues with harming others, but for defensive purposes, and we can argue about the blurring between offense and defense. Like I like I, I understand that need. But I also would say that like we have as a country advanced to the place that we are now due to open science, due to open funding of science. Like DARPA in the 1960s or the equivalent of DARPA in the 1960s, the equivalent of all the stuff that gave birth to the space race was because the US government was funding foundational research, was funding foundational engineering. And I want to get back to that world. I want to get to the world where like

JEAN-PHILIPPE COURTOIS: Yes.

RAFFI KRIKORIAN: It is not just a company who's willing to dump a billion dollars into something because actually the biggest sources of cash are places like the US government. So imagine the kinds of advances we could do if every single computer science researcher had a huge budget from the US government to do foundational work. We can do amazing things on top of that. And, you know, to talk about national defense, like

JEAN-PHILIPPE COURTOIS: Yes.

RAFFI KRIKORIAN: They could also be sitting on top of that and then building their own things from there. So like I like to think of it more of just like we need to get back to a world where like we're doing large funding of this foundational work and then we can do because with large funding comes risk-taking comes all these other explorations right now because all the funding of computer science is basically being done by the big tech companies. They're all doing it specifically so they can do make money off of people like you and me. I have no problem making money.

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: Yes.

JEAN-PHILIPPE COURTOIS: Bye.

RAFFI KRIKORIAN: but that has a downstream effect of focusing what research looks like. So right now, the vast majority of AI work that's not robotics is basically looking at transformer models and figuring out how to productize them. I have this feeling that transformer models are a local maxima. There are other solutions out there, but we're going to do everything we can to eke out a dollar of transformer models. But if we did large funding of basic research from the federal government,

JEAN-PHILIPPE COURTOIS: Yes.

RAFFI KRIKORIAN: Some research could be out to exploring the other spaces of what might be possible. There might be a world where you can figure out how to get the equivalent performance of transformer models in the tenth of the hardware. We might figure out completely different ways to represent the human brain. We're not doing any of that right. We are doing it but not at the scale needed to move the And so like that's the kinds of stuff I would be whispering in Trump's ear. just like

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: The scale.

RAFFI KRIKORIAN: You can't just outsource this all to companies because their incentives don't allow them to go explore and we need to explore.

JEAN-PHILIPPE COURTOIS: Yes.

JEAN-PHILIPPE COURTOIS: I mean, again, it's super insightful and such a big discussion to have that just, of course, in your home country, Rafi, I would say in mine, in Europe, other democracies as well. It's big implications globally. Now for the last section, because time is moving very fast, so much we could discuss, I'd to get into the last segment, Rafi, about leadership. So from the world to you as a leader.

RAFFI KRIKORIAN: yeah.

RAFFI KRIKORIAN: Okay.

JEAN-PHILIPPE COURTOIS: You've been into some amazing missions. Now listeners understand the scope of missions you took on in your life and more to come, as I said, because you've got a few more exciting jobs ahead of you that I suggested. I'd love to understand the way you've been learning, I'm sure, to inspire people to join in a mission, particularly when you work in nonprofit, philanthropy, politics, not just in the largest well-funded companies in the world where you get money. to high engineers for dozens of millions of dollars, Or even billions sometimes crazy. So how do you do that? How do you pick those talents, inspire them to bring their best to the kind of mission discussing this podcast?

RAFFI KRIKORIAN: Yeah. So like one of the things I discovered in my career at Twitter was the power of telling a good story. And so like the way we motivated, so my engineering team, worked on what we called multi data center failover a lot of just like what happens if like a squirrel chews through the power lines of a data center and goes offline? How do you recover from that?

JEAN-PHILIPPE COURTOIS: Hmm.

RAFFI KRIKORIAN: But explaining it like that is super boring. Multi-Center Data, Multi-Data Center Failover. So we did a presentation at one of the All Hands is at Twitter, where we literally did an animation of Godzilla trampling through one of our data centers. And all of a sudden, everyone in the audience got what we were talking about. They all understood it in that moment. And so for me, it's always been, from that point on,

JEAN-PHILIPPE COURTOIS: You Yes.

RAFFI KRIKORIAN: How do we tell that emotionally resonant story that really hits someone of just like, I get the problem, I think I can do something about it. so when I talk about, well, DNC was fairly easy because you can see horrible things on the news and I could draw a direct line of like, we can do something about this. When it came to nonprofits in the social sector, it's helping them tell their story in a way that an engineer could grok it and be like,

JEAN-PHILIPPE COURTOIS: Yep.

JEAN-PHILIPPE COURTOIS: Yes.

JEAN-PHILIPPE COURTOIS: to speak.

RAFFI KRIKORIAN: I could build something that can do something like that. And now with Mozilla, I'm just really trying to focus on the like, like, imagine a world of like, if all this goes wrong, like we could end up in the matrix, like we could end up in all these other places, but you actually could do something about it, like, and here are the five ways. So like, I actually spend a lot of time these days just watching masterful storytellers do their things, like.

JEAN-PHILIPPE COURTOIS: Yes. Yes.

JEAN-PHILIPPE COURTOIS: Yes.

RAFFI KRIKORIAN: watching Maria Reza tell a story of democracy gone wrong, watching Nick Thompson talking about how to be a runner and be a leader. Those are like, it's not actually the content, but the way they tell that story is the thing I aspire to. I aspire to tell a story the way they tell a story.

JEAN-PHILIPPE COURTOIS: Mmm.

JEAN-PHILIPPE COURTOIS: Yes. Yes.

JEAN-PHILIPPE COURTOIS: The emotion, No, I think it's a golden nugget. I'm so much with you on this one, Rafi. I think people need to inspire and they need to tell stories because humans are all about stories sharing from millions of years, by the way. No, I'd to dig into your personal observation because you work through many of those steps. It reminds me of a discussion I had with someone you may know, Whitney Johnson. Whitney Johnson writes Amina Adder as a guest and...

RAFFI KRIKORIAN: yeah.

JEAN-PHILIPPE COURTOIS: And she's been talking about using the S-curve of the innovator's dilemma, but applied to you as a person, right? Going through your own transformation. And so what she says is that real growth often happens when you choose to leave mastery behind you, step back again into beginner's mode and climb a new curve, you know, before circumstances force the change. You've done that so many times, Rafi, as you share those stories.

RAFFI KRIKORIAN: Yeah.

JEAN-PHILIPPE COURTOIS: Again, moving from corporate engineering, commercial, public service, and so on and so forth. So the big question, of course, is how have you managed those transitions personally as a leader? And what have you learned about yourself now that makes you a stronger leader? Because of all those moments of doubts, of failure, probably, in reinventing yourself and be the dumbest person in room, as you said, as opposed to the smartest.

RAFFI KRIKORIAN: I mean, I think it comes down to patience. Like I think patience is the thing I've learned. And I think that like, it's not just through my work. You know, I'm a father of two young boys. You need a lot of patience to be the father of two young boys. But I think that like, you know, earlier in my career, I was impatient to make action. And then the problem is that like, if you want to go fast, you go alone. If you want to go long, you go together. And so like,

JEAN-PHILIPPE COURTOIS: Patience.

JEAN-PHILIPPE COURTOIS: Yes. Yes.

JEAN-PHILIPPE COURTOIS: Yes.

JEAN-PHILIPPE COURTOIS: Yes.

RAFFI KRIKORIAN: You just got to have enough patience to be like, if I can convince someone else, then that's my job is to convince someone else and then they convince someone else and they convince someone else because then you have an army that can show up. It's not just me as a person. And so like, you know, it is very challenging when I feel like I can just do it right now. But then you have to realize I can just do that one thing. I can't do the entire battle.

JEAN-PHILIPPE COURTOIS: Yes.

JEAN-PHILIPPE COURTOIS: Yes.

RAFFI KRIKORIAN: And so like having patience, I can find the people convinced of people, let's build that army. Then we can take on that entire battle that I think that's for me, the thing that's come out.

JEAN-PHILIPPE COURTOIS: Love it. People may not know on the show, but you have your podcast as well, called Technically Optimistic, which is intriguing, of course, in an era where many are very pessimistic about tech as well. So why are you optimistic about technology's future, Rafi, and how do you maintain that optimism personally?

RAFFI KRIKORIAN: Correct.

RAFFI KRIKORIAN: Yeah, I'm optimistic. It's a great question. It's called technically optimistic. Check it out on iTunes and Spotify. I'm optimistic because I believe in people like I believe that if you can clearly explain, if you can educate, if you can then give them their mission that people will actually change the world. And I feel like we're in this moment right now.

JEAN-PHILIPPE COURTOIS: Yes. Yeah.

JEAN-PHILIPPE COURTOIS: Yes.

RAFFI KRIKORIAN: which is both a combination of like learned helplessness when it comes to technology. I was just like, well, this is what Sam Alpin said, or this is what Apple is gonna give me. And I'm just like, that's not, what? Like that's not the world we live in. Like there are better things that we can do. Like don't be, don't learn helplessness. Like it's not inevitable. You can choose it. And then on the other hand, I feel like we're in this crisis of imagination when it comes to technology.

JEAN-PHILIPPE COURTOIS: Yes.

RAFFI KRIKORIAN: When I talked about the Media Lab early before, people used to dream about what you could do with technology. And we seem to have lost that. Like we have outsourced dreaming to a few people who are building companies and we all need to dream again. And I acknowledge the world sucks, but like we all need to dream in some ways. And so like for me, technically optimistic was exploring that line of like, how do you get people to dream about what's possible?

JEAN-PHILIPPE COURTOIS: Hmm.

RAFFI KRIKORIAN: And how do you give people enough information so they can be like, I'm on the wrong end of the deal here and like actually then do something about it. So, you we told stories about data and surveillance for women's reproductive rights. We talked about what the future of the entertainment industry could look like. We talked about what future social media could look like. So it was a fun set of conversations.

JEAN-PHILIPPE COURTOIS: Yep.

JEAN-PHILIPPE COURTOIS: Yes.

JEAN-PHILIPPE COURTOIS: Hmm.

JEAN-PHILIPPE COURTOIS: Yeah. Yeah, so my last couple of questions, because I know you're very busy with your big mission with Mozilla. Something I ask many of my guests, how do you keep yourself energized and grounded in this chaos of information that could make you really sad and depressed all the time? How do you do that every day?

RAFFI KRIKORIAN: My sons are some of the most creative people you'll ever meet. Like my youngest, I have no idea where he got it from because I'm an engineer, is an incredible artist. He makes small comic books. It comes from school every day and there's a new comic book in his hand. The other day he learned, he saw me using cursor and then he asked to learn to use it. And then now he just takes my laptop and vibe codes complete video games that he shares with his friends and stuff like that.

JEAN-PHILIPPE COURTOIS: Yep.

RAFFI KRIKORIAN: My oldest child is constantly tinkering with things. And so like, for me, it's the joy and surprise of what like their minds can bring. And I'm just like, well, if we can do that for everyone, like that would be awesome. So like that, that, that's, that's what keeps me grounded and excited. Yeah, exactly. Exactly.

JEAN-PHILIPPE COURTOIS: Yes.

JEAN-PHILIPPE COURTOIS: That's the way you recharge your batteries every morning, I guess. It's wonderful. Now, what is your core purpose as a leader? What is your personal mission, your North Star, Rafi?

RAFFI KRIKORIAN: I mean, it is to basically storytell and motivate because it's that joint maximization I told earlier of just like, when you and I, we will not work together forever. So let's get the best work of our lives done in this moment and let's set you up for that next thing. So that joint maximization of motivating, exciting, mentoring, I think is like, that is the thing I find energizing about being a leader.

JEAN-PHILIPPE COURTOIS: Yes, my very last question now. What do you hope the world might look like in, let's say, five years as a result? And what do you hope, again, in a few words, what is your hope for the impact that you and your like-minded peers can make by 2030?

RAFFI KRIKORIAN: I mean, in five years, I really hope we can get to a place where like AI is working for me and not for someone else. Like I want to live in that world of personal agents, but I want that personal agent not to be someone I've rented. I want it to be someone that lives with me. Like already my phone is probably the device that spends the most amount of time on my body. And I want to know that that phone is operating on my behalf and not someone else's. But like when it comes down to like what I want to, you know,

JEAN-PHILIPPE COURTOIS: Hmm.

RAFFI KRIKORIAN: My wife is a computer science professor. She's way smarter than I am. And so when my son was five, he was asked in this school, what do mama and baba do? And he said, well, mama builds robots, mostly accurate, not exactly accurate, but mostly accurate. What does baba do? And he said, which still makes me tear up to this day, it's just like, he tries to make the world better. And so like, that's what I hope.

JEAN-PHILIPPE COURTOIS: Yeah.

JEAN-PHILIPPE COURTOIS: Yes. Well, that's a wonderful way to close the conversation. Rafi, it's been such an inspiring, insightful discussion, giving me hopes. I'm an optimistic too, by the way. So that's why I'm interviewing people like you, because I think your story shows really deeply that when tech and humanity meet together through positive leadership, amazing things can actually happen. So I hope our listeners have enjoyed this episode. So if that's the case, please leave a rating five stars. Don't be shy. And tell your friends as well. And of course, take care of yourself until we have the next episode together. Thanks again, Rafi. It's been a wonderful experience for me.

RAFFI KRIKORIAN: Such a pleasure, thank you.

JEAN-PHILIPPE COURTOIS: Thanks. Tom?