AI transcript
0:00:16 Hello, and welcome to the NVIDIA AI podcast. I’m your host, Noah Kravitz. For generations
0:00:20 now, people have associated the West Coast of the United States with the technology industry.
0:00:26 From Silicon Valley birthing semiconductors and so much computing technology, to Seattle
0:00:30 and San Francisco during the dot-com boom and beyond, many people around the globe hear
0:00:37 technology and think of Northern California, the Western United States, these places. Well,
0:00:41 these days, it’s the city of San Jose, California, that’s grabbing headlines for its innovative
0:00:47 approach to using AI in Silicon Valley. Our guests today, and we have two of them, we’re
0:00:53 going to dive into what it means to be an AI city today in 2025, both at a high level and
0:00:58 in terms of some on-the-ground engagement and projects and initiatives going on right
0:01:05 now in San Jose. So let’s get to it. Matt Mahan is the 66th mayor of San Jose. He’s been serving
0:01:10 since 2023. He’s a former teacher. Matt transitioned out of the classroom and into Silicon Valley
0:01:15 entrepreneurship and then city leadership, and his current administration has positioned San
0:01:20 Jose as a national leader in civic adoption of artificial intelligence. They’ve launched
0:01:26 initiatives including an AI upskilling program for over 1,000 city employees and what I believe is
0:01:33 the nation’s first civic AI grant program. And joining Matt is Jumbi Edelburn. Jumbi oversees
0:01:38 global business development for smart cities and spaces at NVIDIA, where he drives the adoption of
0:01:44 AI solutions for spatial intelligence and urban applications. With a PhD and extensive background
0:01:49 spanning chip design and leadership roles, Jumbi brings deep technical expertise and passion
0:01:56 to using AI to transform city operations and urban life. All of which is a longish way of saying we’ve got
0:02:03 the two best people I can think of here today to talk about what AI means to cities, to the people who
0:02:09 live in the cities, and to where our cities and our communities are headed in the future. So Mayor Matt
0:02:14 Mahan and Jumbi, thank you so much for joining the NVIDIA AI podcast, and welcome.
0:02:18 Thank you, Mel. It’s great to be here with you and Jumbi. Looking forward to our conversation.
0:02:20 Thank you. Glad to be here.
0:02:26 Absolutely. Mayor Mahan, I’ll start with you. Can you talk a little bit about what inspired you,
0:02:31 kind of your journey from the classroom to the tech world? And I’ll throw in real quick that
0:02:37 our stories are kind of parallel to there. I started off as a teacher as well, so a lot of respect for that.
0:02:41 But your journey that, you know, led you to now leading San Jose and how all of that kind of
0:02:47 shapes your vision for the city at large, but also, you know, thinking about and starting to use AI
0:02:53 in daily operations. Yeah, you know, no, it’s definitely a non-linear career path and not one
0:02:58 that I mapped out with any intentionality. I’ve just always been really curious about the world.
0:03:02 I’ve always loved problem solving. I think that’s what drew me to entrepreneurship.
0:03:09 But more than anything, I’ve wanted to think about how we make the world a better place. I grew up in a
0:03:15 little farming town and we had a lot of challenges in my neighborhood and my community. And it was coming
0:03:20 to San Jose when I got a work-study scholarship to a high school here that really changed my life.
0:03:25 I was exposed to the innovation economy, was privileged to get a great education, fell in love
0:03:31 with Silicon Valley. And I’ve just always wanted to solve problems in an entrepreneurial way, but apply
0:03:38 them to the biggest challenges we face of education and health and safety and climate and the housing
0:03:44 costs, the big issues that affect all of us. So I guess you could say that’s what led me here. I’ve
0:03:50 tried to bring a bit of that startup environment, entrepreneurial spirit into City Hall and run the
0:03:55 mayor’s office a little differently than it’s traditionally been done. And we’re having a lot
0:03:59 of fun just getting everybody focused on the outcomes we’re trying to drive and being a little less
0:04:04 worried about risk-taking and making mistakes. So we’re trying to bring a more experimental culture.
0:04:10 Fantastic. Jumby, from your perspective and kind of by way of introduction to the work that you do,
0:04:17 what are some of the biggest opportunities that you see for AI adoption in cities in particular?
0:04:23 Yeah, you know, first of all, I should say I’m super honored to be on the same show with Mayor Mahan
0:04:29 being a resident of San Jose. I can actually see some of the changes that the mayor is implementing
0:04:37 and they are pretty amazing. So, you know, there are just a large plethora of applications in cities for
0:04:45 citizen services, traffic management, public safety, energy optimization, infrastructure maintenance, etc.
0:04:52 But I’ll highlight two major areas where cities are seeing significant impact from AI.
0:04:57 And I think this, these will resonate with the mayor because I’m actually picking up some of the stuff
0:05:04 that they’re doing in San Jose. So, you know, the first big area is making city services more efficient
0:05:13 and accessible. So, AI-driven automation for providing city services, for example, permitting. And here again,
0:05:20 I believe San Jose is using it for streamlining ADU permitting, for example. And then there are other
0:05:28 services such as business licensing, filing forms, emergency relief support, looking for affordable
0:05:35 housing options. You know, AI can make the delivery of all these services, you know, much more streamlined,
0:05:42 comprehensive, comprehensive, personalized, and accessible. So, you know, just to highlight the
0:05:49 accessibility part, you know, the ability of AI to automatically translate the written and spoken
0:05:56 language means that city services can now be provided in a host of languages, which is, you know, of course,
0:06:02 important for a city like San Jose, whose residents speak over 50 languages, I believe. You know,
0:06:09 automated translation is also important, for example, for 9-11 calls. You know, imagine someone calling in
0:06:13 a language on the person on the other side, not even really understanding what that language is,
0:06:20 so they can find a translator. So, AI can automatically translate, understand the emergency,
0:06:23 and route the calls to provide the appropriate response.
0:06:29 You know, as I, not to interrupt, but as I was preparing for the episode, for the recording,
0:06:35 something reminded me of, I don’t know if either of you saw this movie or remember from,
0:06:41 this is 25 years ago, but during the first dot-com boom, there was a movie called Startup.com,
0:06:47 and it was these entrepreneurs trying to build, you know, state-of-the-art, then was web, you know,
0:06:54 web 1.0, right? But trying to build a web-based system where folks could engage with civic services,
0:06:58 and government. And I remember, if I remember correctly, one of the big dramatic points in the
0:07:03 movie was when one of the founders was able to successfully pay a parking ticket online,
0:07:09 you know? And so, I say all that just to say that listening to you speak about, you know,
0:07:15 translation is such a great example to bring that home of just that real-time. Yeah, so it’s amazing
0:07:15 how far we’ve come.
0:07:21 Yeah, it really is. And I’m, John B, I’ll return the fact, I’m honored to be here with you. You have so
0:07:24 much domain expertise here. It’s exciting to have this conversation. And as I noted,
0:07:30 as a San Jose resident, you are one of my one million bosses. So, I’ll be listening carefully
0:07:37 to what you have to say. We’re seeing in just incredible breadth of application for these new
0:07:42 tools. And John B spoke to many of them. Language translation is a great one. In our public meetings
0:07:50 now, without having to go hire dozens of interpreters, we have real-time language translation in dozens of
0:07:56 languages at every public meeting. And it’s two ways. So, I can be speaking in my native language,
0:08:01 English, and anyone who speaks dozens of other languages can see in real-time translation in their
0:08:06 native language. They can speak in their native language, and I can see translated on a screen
0:08:11 in real-time in English what they’re saying. And so, we can seamlessly use our native languages to
0:08:18 communicate in real-time in all public meetings now, again, without having to massively expand our budget
0:08:25 for this service. But to the point of permitting and those basic city services, we’re starting to use it
0:08:30 to automate the detection of potholes, to speed up bus routes. We’re just seeing so many applications
0:08:36 where we can take an existing city service and make it significantly better through automated
0:08:41 detection and service deployment and kind of building these more responsive systems.
0:08:45 Mayor, you said potholes, and I got a little triggered. I’m not one of your constituents,
0:08:50 so it’s not anything to you. But, you know, as we just started to scratch the surface of so many
0:08:56 amazing things that can be and are being done, what are some of the challenges when it comes to adopting
0:09:02 AI in a city? And, you know, this can be either, Jumby, to you sort of broadly from your work, but I don’t know
0:09:08 if Mayor Mahan, you’d be willing to take a stab at, you know, just some of the things that you’ve tried
0:09:12 to work on or would like to work on at the San Jose city level, but are just kind of challenging right now.
0:09:17 You know, I think innovation can be hard in government for a couple of reasons. One is we have a fairly
0:09:23 risk-averse culture because unlike a startup, which is the world I’m coming from, we’re really not allowed
0:09:30 to fail in a big way. First of all, we often have lives at stake, so we have to prevent that kind of risk.
0:09:36 But also, we’re the only game in town. That tax revenue is coming every year and we’re relied on to provide
0:09:41 essential services. If a startup tries something and it doesn’t work and they fail and they run out of funding,
0:09:48 there are often a hundred other startups in the space that are pursuing similar ideas. So we have a risk aversion.
0:09:54 We also have much more need in the community for the essential services we provide than we have resources for.
0:10:01 So it can be hard to carve out time, staff time and budget to try new things when we feel like we’re
0:10:07 underwater just scrambling to do the things that are already identified as core needs and core services.
0:10:13 But I’m lucky to have a great chief innovation officer, Stephen Cain’s on my team. We have very
0:10:18 forward-looking leadership at the city within the city manager’s office. I think we’ve all recognized we need to
0:10:25 carve out that time to pilot new tools because the only way we ultimately make government more productive
0:10:31 and do more with limited resources is through innovation. It’s technology adoption, but it’s also,
0:10:37 and I’m sure we’ll talk more about this, so much of it is about workforce development, giving people
0:10:43 permission to experiment, training them in new tools, giving them space to learn, to grow, to try things,
0:10:49 things to fail. That’s where I think it takes the most effort to break through and actually have that
0:10:51 kind of innovation and better outcomes.
0:10:56 Jovi, do you see similar challenges, similar patterns in the work you’re doing?
0:11:01 Yeah, absolutely. You know, one of the things I’ll highlight is that initially when we started really
0:11:08 digging into the whole sort of gov tech area about five years ago, there was a lot of resistance from
0:11:15 all kinds of constituents around AI. You know, they wanted to have the proper guardrails. They wanted
0:11:22 to make sure it wasn’t biased and all that good stuff. And, you know, Mayor Mahan and the chief
0:11:27 innovation officer, Stephen Keynes, are the ones who actually led the charge on coming up with a
0:11:34 framework. It’s called the Gov AI coalition, which actually was an initial attempt to lay out, you know,
0:11:39 know, some basic guidelines. And that, in my opinion, really gave a boost to a lot of cities
0:11:43 looking at it and saying, look, guys, you know, we have all these guidelines now. We could just, you know,
0:11:49 go for it. Let’s not just hold back. So rather than, you know, waiting for a massive amount of regulation,
0:11:55 you know, even these kinds of basic guidelines that Mayor Mahan drove initially, you know,
0:11:57 I think made a big difference in the adoption.
0:12:02 Well, thanks for mentioning that, Jambi. I’ll take the opportunity to note that our
0:12:11 second annual Gov AI Summit is in San Jose, November 5th and 6th. This coalition, if you will,
0:12:17 has taken off like a startup. We didn’t really know what we were starting, but I’m very proud of our team
0:12:24 at the city for convening other public agencies. And as Jambi was saying, we needed to create a permission
0:12:30 structure and a set of guardrails and guidelines to give people confidence to take risks.
0:12:37 And so this entity has grown into a platform for over 800 public agencies all over the country at
0:12:43 every level of government. It has AI policy manuals where people can take something off the shelf and
0:12:49 adapt it for their own agency, but they have those guidelines in place. It has tools like an incident
0:12:54 response plan. If something goes wrong, how do you dissect it and do a retrospective and learn from
0:13:01 it? It has a vendor registry where we can collectively initiate procurement by essentially understanding
0:13:06 what are the use cases? Who are the vendors? What are the, you might consider it like the nutrition
0:13:12 facts on your box of cereal. What do we know about these vendors? And then we have a library of actual
0:13:17 deployments of use cases where we say, well, here are all of the pilots we’ve run and here’s how far we
0:13:21 got. Here’s what we did. Here’s, here was the data. Here’s what we learned. And so it’s just
0:13:28 become this incredible platform for collaboration and it’s allowing public agencies to leapfrog off of
0:13:32 one another. We can’t test everything all at once in San Jose, even though we are a bigger city in
0:13:36 Silicon Valley. So we, we look to other cities and say, oh, somebody else is doing something really
0:13:41 interesting around procurement or permitting. Let’s go try that. We can jump off from wherever they,
0:13:44 wherever they are today versus starting back, you know, at zero.
0:13:48 Open source government in a, in a way, exactly. Yeah.
0:13:54 How are the city workers taking to using the new tools? And, you know, I think you mentioned earlier,
0:13:58 well, you mentioned the upskilling program in particular, and, um, I think you, you’ve trained
0:14:01 over a thousand city staff at this point, which is fantastic.
0:14:04 That’s the goal. We’re in process toward a thousand.
0:14:06 That’s the goal. Okay. Got it. Got it. Okay.
0:14:11 You also mentioned, I think, or maybe it just always pops into my mind, but you know, the notion of,
0:14:17 you know, AI comes up and you still have people who respond in different ways and often, you know,
0:14:21 maybe one of two, one of three ways, but one of them is, I’m not so sure about that. Yeah.
0:14:25 What’s it been like rolling this out to city employees and kind of the range of reactions
0:14:29 that you get and, and just kind of how’s the whole process going with the people?
0:14:36 Yeah. I think Jumbi would probably agree with me that the innovation isn’t just about new tools and
0:14:43 technologies. We often get very focused on the new app or the new, the new device, the new gadget,
0:14:51 and really the, the, the greatest unlock in terms of innovation really comes from humans wanting to
0:14:56 to use it, wanting to experiment, being creative in the application of new tools and finding new
0:15:03 applications with, of, of these capabilities. AI is so, so robust. The, the number of applications is,
0:15:11 is basically limitless. And so what we’ve tried to do is essentially create a, a system through which
0:15:17 our workforce can become more comfortable with and adept in applying AI and then turn them loose to
0:15:24 go apply it in their daily work. We have an MOU with Nvidia that includes a pillar around workforce
0:15:29 development. We also have a close relationship with San Jose State University, and it’s right here next
0:15:37 door to city hall. And we’ve created this curriculum for AI upskilling. And rather than do this in a top
0:15:42 down way where we came to our city workers and said, AI is coming. It’s a big deal. You should
0:15:47 be scared for your jobs and go take the course or you, you know, you’ve got to, you got to quickly
0:15:54 learn this thing. We did the opposite. We said, we see AI as not a labor replacement so much as labor
0:15:59 enhancement. And we want to empower you. We want to give you the training and tools you need to keep
0:16:04 growing, to be more productive and to have frankly, a more interesting job. Let’s take some of the drudgery
0:16:11 out of your day-to-day work by automating the things that are kind of rote and routine. And we saw a
0:16:18 great response. We’ve already graduated two cohorts of employees who voluntarily signed up for a 10 week
0:16:24 AI upskilling program. We’re in our third cohort. As you mentioned, the goal is to have a thousand
0:16:30 employees by the end of the year who sign up voluntarily, go through a 10 week upskilling
0:16:35 course. And then they go back into their departments. We have over 20 different departments. We like to
0:16:41 say we do everything from A to Z. We literally run an airport and we have a, we have a small zoo out at
0:16:47 Happy Hollow. We, we want them to go out to their departments and start building those custom GPTs and
0:16:52 finding different applications. And we’re seeing that happen. And it’s incredible. We’ve already saved
0:16:59 over 10,000 hours of staff time this year, just with those first cohorts beginning to
0:17:03 create these new workflows using, using these new tools.
0:17:09 Fantastic. I will, Jumbi, before I get to you, I just want to do a quick shout out promotional drop.
0:17:15 We had Dr. Teniente Mattson, president of San Jose State on the podcast. The episode came out
0:17:22 late September, wherever you get your podcasts. But yes, a very, she, a lot of similar themes talked
0:17:27 about. Definitely worth a listen there. Glad you mentioned her mayor. But Jumbi, I was going to ask,
0:17:32 you know, kind of along these lines, what do you see as some of the biggest opportunities
0:17:38 to help public sector employees drive civic innovation through the use of AI? You know,
0:17:42 and the mayor was talking about this. And this is kind of how I got into it myself in my own work and
0:17:48 learned really what works for my workflows and then how to share and learn from colleagues was just
0:17:52 kind of getting a sense of the guard, guardrails, right? What, what should I just not do in the work
0:17:58 environment? And then just playing with it and seeing how it could work in my daily flows. Jumbi,
0:18:03 what are you saying? So, you know, I’ll absolutely echo what Mayor Mahan just said that, you know, in the
0:18:10 public sector, I like to think of AI more as assistive intelligence. You know, there really is an artificial,
0:18:16 anything artificial about it. It’s real stuff that people can use in their real work to make it, you know,
0:18:23 more interesting to remove some of the sort of repetitive stuff that, you know, AI is now good
0:18:30 at doing. So I think the two broad areas of opportunity I see in this kind of innovation is
0:18:36 using AI to unlock, you know, potential of large amounts of data that cities have
0:18:44 to streamline services and provide sort of more holistic insights. And, you know, in the end, make
0:18:51 better informed decisions. You know, the ability to take large amounts of multimodal data, and that could
0:18:59 be, you know, data, audio files, videos, etc., consolidate it, analyze it, is where, you know,
0:19:05 AI can be incredibly useful. And, you know, creating these kinds of applications, as, you know,
0:19:12 Mayor Mahan was saying, is now, you know, a lot easier to do than it used to be even a couple of
0:19:17 years ago. In fact, you know, the amazing work that they’ve done in upskilling the city’s workforce
0:19:24 and giving them the training to create new tools, you know, with the amazing amount of data that cities
0:19:29 have, they can really truly become, you know, engines of civic innovation.
0:19:29 Yeah.
0:19:36 And, you know, one thing I’d add is, you know, once cities start unlocking the value in their data,
0:19:42 they may actually find opportunities to even monetize some of the data and the analytics
0:19:46 to boost revenues, which is, of course, you know, something all cities are looking to do.
0:19:51 Along those lines, Mayor Mahan, are there, you know, you mentioned potholes earlier,
0:19:56 and, you know, connecting the virtual dots, if you will, kind of from something like AI and talking
0:20:01 about data and, you know, online services are something that are made real, right? When a
0:20:07 citizen can log into the public works website or whatever the interface and have something happen.
0:20:12 But connecting it back to the physical world of core public services, services, you know,
0:20:16 filling in potholes, cleaning up waste, public transit issues, these kinds of things.
0:20:21 What impact have you seen in these core services at the street level,
0:20:24 directly resulting from AI-powered initiatives?
0:20:32 Yeah, I think that the availability of relatively low-cost cameras, other types of sensors that we
0:20:39 can put on vehicles or otherwise distribute out through the built environment, and then the incredible
0:20:45 data generation, real-time insight, data and insight that can be generated from that is really
0:20:50 powerful and we’re just scratching the surface. My favorite example in San Jose is bus route
0:20:58 optimization. We partnered with a local startup, Light AI, to put these sensors on buses and then
0:21:04 controllers on red lights to create signal prioritization. Signal prioritization for public
0:21:10 vehicles is not a new concept. People are probably familiar that many fire departments around the country
0:21:17 will give that signal prioritization to a fire engine when it’s going lights blaring down a road,
0:21:23 but it has been a very blunt instrument and one that has had downsides. So you’ll see a sea of green
0:21:29 lights for miles ahead as the fire engine goes down the road and it’s holding everybody else up,
0:21:36 creating congestion and creating all these negative externalities. This new AI-enabled system
0:21:44 is dynamic. It doesn’t guarantee a green light for every public bus, but it tracks the velocity of
0:21:51 the bus, tracks what’s happening more broadly with the signals and can dynamically lengthen or shorten the
0:21:58 duration of a red light so that cross traffic is still going, but you’re reducing the probability that a
0:22:03 bus ever hits a red light. And so what we’ve been able to do without massively disrupting traffic
0:22:09 flows in the city is reduce the likelihood that a public bus in San Jose hits a red light by over 50%,
0:22:17 which speeds up their overall bus route time by over 20%, which means on a route that’s a loop,
0:22:22 you can have better bus service by 20%. You not only have a faster commute, but you can increase the
0:22:27 frequency of buses stopping by a bus stop by 20% without having to buy any more buses, hire any
0:22:35 more drivers. We can literally just make your commute faster, have more bus passes and optimize the whole
0:22:41 system in a way that encourages public transit use, which has a further spillover effect of fewer cars
0:22:47 on the road and less congestion. And all because we’re dynamically monitoring velocities of buses and
0:22:52 slightly modifying the duration of signals. And it’s, it’s just a really cool, it’s like kind of a neat,
0:22:57 neatly packaged use case that has really meaningful real world effects for everybody.
0:23:02 Now that’s beautiful. And the down, the downstream effects you mentioned, you know, just fantastic.
0:23:08 Uh, Jumbi, you know, yeah, I was just going to add, you know, this is an amazing application and just
0:23:16 imagine now sort of extending this application across the city to not just prioritize bus routes, which is,
0:23:20 you know, of course important for public transportation, but really to orchestrate,
0:23:26 you know, all the lights in the city using AI. Uh, one city that actually deployed a solution
0:23:32 orchestrating literally all the lights across the city to optimize for traffic patterns, et cetera,
0:23:40 saw 50% reduction in overall congestion, uh, which means, you know, big reduction in times that cars have to
0:23:45 wait at lights, which also has a positive environmental impact because, you know,
0:23:50 as they’re waiting at the lights, you know, they’re, uh, have a bunch of emissions.
0:23:57 So, you know, intelligence signaling, uh, can, you know, like the mayor said, be prioritized
0:24:03 public transport or to make, you know, passage of law enforcement and emergency management vehicles
0:24:09 quicker and safer, but applied more broadly, it can actually have huge impacts and congestion
0:24:10 within the city.
0:24:14 I love that. You’re going to have to share that use case with us, Jumbi.
0:24:22 Absolutely. Our guests today are Matt Mahan, the 66th mayor of San Jose. Matt has been serving since
0:24:29 2023 coming from a background in entrepreneurship and the tech industry. And we are speaking with him
0:24:36 about AI adoption in San Jose and beyond, uh, more broadly in cities all over that can benefit from AI
0:24:41 from what San Jose has been doing, sharing out some of the work they’ve been doing, um, as we talked
0:24:48 about. And we’re also speaking with Jumbi Edelburum. Jumbi is an NVIDian who oversees global business
0:24:54 development for smart cities and spaces at NVIDIA. And we’ve been talking about some of the ways that
0:25:01 San Jose, other cities are starting to use, have been using, uh, and getting measurable results out
0:25:07 of bringing AI into day-to-day civic life planning, everything that goes into running and governing a
0:25:11 city. And, you know, this is no surprise to you guys, given your work, but as we’ve been talking,
0:25:16 as I’ve been listening to the two of you talk, just been thinking about the parallels between
0:25:22 what you’re talking about doing with AI in civic operations and overseeing communities. Guests we
0:25:28 have from enterprise, uh, running large factories and, and all different situations, medical research,
0:25:33 what have you. So many parallels, so many things kind of boil down to these, you know,
0:25:39 getting started with the data and the guardrails and making sure that the people using the system sort of
0:25:44 understand, you know, well, what does this do? What does it can’t do? What can I try? And,
0:25:49 and, you know, what do we have to really get, you know, more governance and policies and kind of
0:25:55 production ready mindset into before we get into, and just fascinating thing about all the things that
0:26:01 go day-to-day into, you know, being the mayor of a city, let alone one on the size of San Jose.
0:26:07 Mayor Mahan, San Jose just unveiled the winners of your first ever civic AI grant program. So along the
0:26:11 lines of all these things that go day-to-day into running a city and innovating for the future,
0:26:16 what kind of problems, what kind of, uh, issues are these startups solving and working on?
0:26:22 And what is a program like this? What does the investment mean for, for the city of San Jose
0:26:25 and the notion of AI innovation in cities nationwide?
0:26:32 Yeah, we, we, we, this was a first of its kind program. As far as we know, we’re, we’re trying to
0:26:40 create an AI ecosystem in San Jose that’s broad, diverse and touches every sector. And I, I think
0:26:47 downtown is a particularly good place to see the diversity of that ecosystem. We’ve talked a lot about
0:26:52 what we’re doing inside city hall. We mentioned our partnership with our wonderful public university,
0:26:59 San Jose State University across the street, which has 40,000 students in downtown and graduates more
0:27:03 students with technical degrees every year than any other university in the state of California.
0:27:07 Mayor, forgive me, but for people who, listeners who may not know the scale,
0:27:09 you mentioned the size of San Jose State. Yeah.
0:27:15 San Jose, around a million population, third largest city in California by population. Is that right?
0:27:19 That’s right. Yeah. We’re the largest city in Northern California with a million people,
0:27:26 only LA and San Diego and Southern California are larger in the state. And I believe we’re now 12th
0:27:32 largest in the country. Okay. So it’s, it’s a big city and we have big challenges, but it also means we
0:27:37 have the resources to be more innovative. And it’s particularly helpful that we’re here in the heart
0:27:45 of Silicon Valley with just an incredible workforce, incredible universities and the, the AI startup
0:27:52 incentive, the grant program that we created was a program through which we wanted to get to know who are
0:27:59 all the local startups working in AI that might want to work on civic and social issues, might want to
0:28:06 have a presence in our downtown next to our large public university. As we try to create this ecosystem,
0:28:11 we recruited plug and play to open an AI center of excellence downtown. The university is doing
0:28:18 incredible work. We have the MOU with Nvidia that I mentioned. And the other piece is making sure we are
0:28:24 more conducive to startup formation. I come out of that world. I help build two startups.
0:28:31 Barriers to entry are high. Housing costs are high. We have a lot of regulatory barriers in California,
0:28:37 and we just want to clear the way. And so we, we move forward to incentive programs. One was a very
0:28:43 simple office leasing incentive that’s open to anybody downtown, which says, if you lease new office space
0:28:49 downtown, you get, you have no business tax for two years and free parking in city garages for two years.
0:28:57 So that’s just kind of a baseline. But the incentive program or the grant program for startups solicited
0:29:05 applications for AI startups that wanted to be in San Jose, ideally work on a civic or social problem.
0:29:14 And we would give grants, which we did. We had, we were blown away. We had 170 applications. We chose our
0:29:20 four winners that are working on a diverse range of issues from maternal health to food waste reduction
0:29:28 to a satellite company and offered them the cash incentives, but equally importantly,
0:29:35 essentially a network and a community that they’re now part of. And we connected them with VCs,
0:29:44 with professional services firms. We had CBRE, the real estate brokers and, and then a venture
0:29:53 company, J2 Ventures offer pro bono professional services for the applicants and, and for the winners.
0:30:01 And, uh, and then NVIDIA very generously offered the ability for the startups to, to, uh, actually apply
0:30:07 for these, the inception program, which Joomby can probably explain better than I can, but we think
0:30:13 that’s a great opportunity. But this ecosystem is how do you bring all these resources to bear in one
0:30:17 dense environment to spark more innovation and entrepreneurship? And Joomby, do you want to
0:30:20 share a little more about inception and how, how the, that works?
0:30:26 Yeah, absolutely. You know, NVIDIA has the program, the inception program for startups.
0:30:31 And essentially it’s exactly what you said. We want to build a community of startups around
0:30:38 particular industry segments. It could be, you know, urban technology or, you know, smart city
0:30:45 technologies. And then of course we have healthcare and retail and manufacturing, et cetera. So, you know,
0:30:50 once these startups in the program, they get a host of benefits, you know, free cloud credits to test
0:30:55 their applications, et cetera. And of course, get connected to all the experts within NVIDIA
0:31:01 that can actually guide them in terms of product development. And all this from the NVIDIA side
0:31:07 is completely pro bono. So it actually makes for an extremely attractive package of stuff,
0:31:08 may or mayhem that you guys are putting together.
0:31:16 I think it’s great. By the way, I’ll just say, I think what we see when these ecosystems start to scale
0:31:24 and generate the next crop of great companies is you need mature companies like NVIDIA to help
0:31:29 facilitate. It doesn’t, these, they don’t just pop up by accident. They tend to be near large,
0:31:37 companies like NVIDIA and large universities like San Jose State or Stanford. And so we’re really
0:31:44 starting to see a lot of green shoots, if you will, here in downtown San Jose and in North San Jose.
0:31:51 Between the two, we have nearly 100 AI startups with venture backing, between them over $5 billion
0:31:57 worth of venture backing. So we’re seeing that next crop of startups come online, which is pretty exciting.
0:32:02 And by the way, Noah, if I could just add one thing, the mayor didn’t mention this, I’ll mention it.
0:32:09 But every month they actually organize a San Jose innovators meetup. In fact, there’s one this
0:32:14 evening and I’ve been going to those and kind of the networking goes both ways. I’ve actually met
0:32:19 companies there that we didn’t even know about and that weren’t in our inception program.
0:32:24 And we kind of introduced them to NVIDIA and how we can help them. But that also is a great program
0:32:29 where, you know, companies, again, build their networks within the city and they get exposed to,
0:32:35 you know, much larger bunch of companies, other collaborators, et cetera, in the region.
0:32:39 So that itself is an excellent initiative, mayor.
0:32:42 Well, thanks for mentioning it. I’ll see you there tonight.
0:32:42 Absolutely.
0:32:43 Absolutely.
0:32:49 So bringing people together on the AI podcast, that’s what we do. That community, that ecosystem
0:32:53 that you’ve both been talking about, though, is so important, you know, for the individual companies,
0:32:58 the startups getting involved. But then, as you said, it just it grows, right? You plant seeds and
0:33:03 you harvest them and you get, you know, you get more than the way the sum greater than the individual
0:33:09 parts, all those cliches. But it’s really true. It’s fantastic to hear about. Along those lines,
0:33:15 Jumbi, I wanted to ask you, technology, AI powered technologies, new technologies being used in
0:33:21 urban settings, government, civic, you know, urban communities. Are you coming across other
0:33:27 innovations, other things that you’re excited about that you’re seeing maybe in other cities or just
0:33:34 kind of in the startup world? Sure. You know, what I find particularly exciting right now is the
0:33:43 potential of the so-called digital twins and AI based simulation technology. You know, digital twins are
0:33:52 essentially digital representations of the physical world. But now, AI makes it possible to run all kinds of
0:33:59 of interesting simulations of what if scenarios that could be used to model, for example, the impacts of
0:34:07 urban development or extreme weather events or fires or accidents. You know, for example, if you see the
0:34:16 impact on traffic of a new development, you know, rather than giving citizens a 100 page traffic impact
0:34:23 analysis report, you can actually simulate the changes and show people visually, which is a lot more
0:34:30 compelling and understandable. And running these kinds of simulations can give, you know, public agencies,
0:34:39 get them much better prepared for development or emergency scenarios. And more importantly, be able to share
0:34:46 information with citizens in a compelling way so they too can be better prepared. So applications around those,
0:34:52 you know, across in cities across the world, I find is a particularly exciting area in terms of urban innovation.
0:34:59 To mention her again, I believe Dr. Tiente Mattson from San Jose State was talking about a digital twin project
0:35:05 they’re doing. And yeah, the possibilities for modeling and the implications of that are fascinating.
0:35:11 Yeah, we’re really excited about it. It’s an emerging technology for us. We’re not as familiar,
0:35:19 but starting to test the waters. And we’ve spoken with some of your colleagues at NVIDIA, as well as the
0:35:27 university. You know, the LA fires really caught our attention when it comes to better modeling things like
0:35:36 fuel load out there, the vegetation risk, which can be modeled, evacuation routes, and just really
0:35:42 getting a better handle on how do we both prevent and prepare for and then respond to disasters,
0:35:48 natural disasters when they happen. So we see, whether it’s that or climate smart or traffic
0:35:52 patterns, as you mentioned, Jumbi, we’re really excited about this emerging area.
0:35:57 Yeah. And, you know, going back, Mayor, to the comment you made about economic development,
0:36:04 of course, attracting businesses to the city, you know, they can even be used to visually show the
0:36:10 businesses, the areas within the city, you know, potential sites for locating their businesses.
0:36:17 And think about applications around tourism. You know, you can give tourists a much richer experience
0:36:25 through things like 3D wayfinding and giving them timely and relevant information on their surroundings,
0:36:31 you know, as they walk around the city. So yeah, just a plethora of applications. That’s a super exciting area.
0:36:36 I love that. Jumbi, if you ever want to leave NVIDIA, we’ve got a roll open for you over here at the city.
0:36:42 So you beat me to it, Mayor. I was going to offer him a spinoff podcast where we’re like,
0:36:47 we take, you know, leaders from different sectors and walk them through all the possibilities.
0:36:53 Yeah. Mayor, on more of a serious note, perhaps, what is next for San Jose?
0:36:59 When you’re thinking about AI and San Jose AI journey and already, you know, as we’ve said,
0:37:05 kind of taking a really bold forward approach, what are some of the projects, maybe technologies
0:37:10 or partnerships that you’re thinking about, maybe excited to announce in the near future?
0:37:19 Well, we have a lot of AI pilots in flight. We have the GovAI summit coming up, as I mentioned. So I’ll be
0:37:24 there and listening very carefully to what other cities are doing and what folks in the room who
0:37:32 are driving this innovation are seeing as next on the horizon. I’ll say for myself, when I think about
0:37:40 the efficient operation of City Hall, the two areas where I would most like to see disruption, if you
0:37:48 will, is really permitting and procurement. You know, I think where government often gets a bad name for
0:37:55 being slow, bureaucratic, unresponsive, struggling to move at the speed of business, tend to be in those
0:38:00 two areas, either when we’re making a purchasing decision of some kind, which can have diverse
0:38:08 implications. It can be a contractor for any kind of service we need or land use related permits.
0:38:14 These are two areas where we have objective standards. If you just take the permitting use case,
0:38:21 we have a general plan. We have zoning and land use designations. We have building codes. We have the
0:38:30 rule set, if you will. And then you have applications from typically professional investors, developers,
0:38:34 not always, sometimes they’re small business owners, but you have applications that have structured data
0:38:41 with certain parameters that need to be met. And I think there’s an opportunity here to take a lot of
0:38:51 the confusion, the back and forth, the unnecessary discretion out of the process and go from a process that can
0:38:58 take 12 to sometimes 24 months, depending on the complexity of the project, and bring that down to a
0:39:05 a matter of weeks, potentially. We are, and Jumby mentioned this earlier in our conversation,
0:39:10 we’re doing some piloting around ADUs. One of the things we found is that even though ADUs are very
0:39:17 popular, these are accessory dwelling units or backyard cottages, small, small backyard units that people
0:39:21 live in. Well, they’re very popular. The music studio I always wanted to build, yeah.
0:39:26 Yeah. Well, San Jose is a great place for them. We have a lot of single family home neighborhoods,
0:39:31 a lot of backyards that are underutilized, a great way to have extended family move in or
0:39:35 even generate income, rent it out to a San Jose State student or whomever. Yeah.
0:39:42 You know, even though we do a large volume of these, ADUs now account for 22% of all new homes built in
0:39:48 the city of San Jose each year. And we’ve done a lot of streamlining on our end. We discovered that we still
0:39:59 still return about 90% of initial ADU applications to the applicant to address missing or incomplete
0:40:07 information. And so we are now piloting an AI tool that’s essentially a little built-in co-pilot in the
0:40:17 application process to give people guidance and feedback in real time. So that the first time they submit, it’s a complete application. And we reduce all that unnecessary back and forth to get to
0:40:19 to get to yes faster.
0:40:27 Love it. Jumbi, where do you see potential for AI to transform urban life in short term in the coming few years?
0:40:55 No, absolutely. I think, you know, the mayor hit it on the head. There are these, you know, sort of painful things that citizens experience with cities or startups experience with cities. You know, the two, the one that the mayor addressed, permitting is a huge area. You know, just permitting, licensing, especially for small businesses, especially in a city like San Jose, where there’s such a large variety of businesses and languages that
0:41:03 that people, you know, people, you know, need to communicate with, that’s going to just be a game changer. And the other one, you know, the mayor referred to, you know, the sort of procurement process.
0:41:25 I think the rate at which, you know, technology is moving, especially in the, you know, the world of AI, both in the physical side, as well as the agentic side, making the process by which these new innovative companies can actually start working with cities and government agencies.
0:41:33 If we can figure out a way to make that quicker, it’s going to make a massive difference.
0:41:47 I mean, right now, as I’m sure the mayor well knows, I mean, there are all kinds of contracts and vehicles and requirements for submissions of a zillion, you know, things on paper. It’s just a long convoluted process.
0:41:51 And if we could simplify that, you know, we could take urban innovation to completely another level.
0:42:08 Yeah, completely, completely agree. We have to speed up our cycles for procurement, for learning, our learning cycles. We’ve got to be able to move faster. And I think AI is going to enable us to do that and still maintain quality.
0:42:10 Exactly.
0:42:15 Mayor, in wrapping up here, you kind of just answered my last question for you, but you can expand on it. Maybe I’ll ask anyway.
0:42:34 Last words of advice to other cities. Maybe there’s a mayor of another city out there listening, and they’re in the process, earlier stages in San Jose is, of adopting AI and trying to figure out, okay, we kind of understand what it’s all about, what it’s doing, how we get started, but what do we do now? What do we use with it as practices? Advice from you.
0:43:00 I would encourage folks to visit the GovAI Coalition. It’s hosted on the San Jose City website. There are now over 800 public agencies, and the whole idea is to create a safe, trusted environment for collaboration amongst public agencies that want to advance the frontiers of AI adoption for public sector use cases.
0:43:16 We have a number of documented case studies. We have a vendor registry, policy manual suggestions, just a very robust set of tools and very practical use cases people can pick up from. You don’t have to start from scratch. You’re not alone.
0:43:38 One of the really wonderful things about the public sector is that cities can be much more collaborative than competitive. Sometimes companies are worried about their trade secrets or losing talent to a competitor. And while certainly cities in the long term, the grand scheme of things maybe are competing on some level, more than anything, we can learn from each other, collaborate.
0:44:00 My view has always been, if San Francisco and Oakland are doing well, we’re going to be doing well as well. We’re an ecosystem. We’re not in direct competition. And so it’s kind of an open book. Come to the GovAI Coalition. Identify a use case that you think is relevant to your residents that your elected leadership would get behind and go test something.
0:44:06 You really learn by doing, you really learn by doing and there’s a lot of existing work out there to build on top of. So that’s where I’d start.
0:44:06 Amazing.
0:44:15 What I thought the mayor was going to say, and I’m going to say it for him because he didn’t, is hire some really good people.
0:44:16 Always.
0:44:31 Seriously, I know a bunch of people on the mayor’s team and he has hired absolutely top quality talent. And you can’t start all these great initiatives unless you have that around you. So, you know, kudos to the mayor for doing that to start with.
0:44:53 Jonby, you’re so right. And I, yeah, I, I, I should have mentioned it, but I will just say to your point, it’s not something, if you’re an elected leader and you’re listening to this, it’s not something you can do alone as the elected. You need those champions inside the departments. It’s, it’s really helpful if you can find a way to do this, either in the mayor’s office or the city manager’s office.
0:45:14 I mentioned our chief innovation officer, Stephen Keynes, having that one point person who’s kind of the trusted go-to for use cases, for what’s happening in other places, and being able to delegate that day-to-day, just kind of owning innovation for the agency is incredibly powerful. So thank you for making that point for me. You’re right on.
0:45:31 Absolutely. Fantastic. Jonby Edelburum and Mayor Matt Mahan of San Jose, thank you both so much. It’s been a great conversation. So much that can be done. It’s great to seeing it not only being done in San Jose, but shared in that spirit. We can get those rivalries out between cities on the sports fields.
0:45:45 That’s right. We have lots of avenues for that, but we can collaborate on raising everybody’s quality of life up using AI to help. So gentlemen, thank you again for joining the pod. Best of luck in all that you’re doing. And, you know, have fun at the event tonight.
0:45:46 Awesome. Thank you both.
0:45:46 Thanks, Noah.
0:46:16 Thank you.
0:46:20 Thank you.
0:46:33 Thank you.
0:00:20 now, people have associated the West Coast of the United States with the technology industry.
0:00:26 From Silicon Valley birthing semiconductors and so much computing technology, to Seattle
0:00:30 and San Francisco during the dot-com boom and beyond, many people around the globe hear
0:00:37 technology and think of Northern California, the Western United States, these places. Well,
0:00:41 these days, it’s the city of San Jose, California, that’s grabbing headlines for its innovative
0:00:47 approach to using AI in Silicon Valley. Our guests today, and we have two of them, we’re
0:00:53 going to dive into what it means to be an AI city today in 2025, both at a high level and
0:00:58 in terms of some on-the-ground engagement and projects and initiatives going on right
0:01:05 now in San Jose. So let’s get to it. Matt Mahan is the 66th mayor of San Jose. He’s been serving
0:01:10 since 2023. He’s a former teacher. Matt transitioned out of the classroom and into Silicon Valley
0:01:15 entrepreneurship and then city leadership, and his current administration has positioned San
0:01:20 Jose as a national leader in civic adoption of artificial intelligence. They’ve launched
0:01:26 initiatives including an AI upskilling program for over 1,000 city employees and what I believe is
0:01:33 the nation’s first civic AI grant program. And joining Matt is Jumbi Edelburn. Jumbi oversees
0:01:38 global business development for smart cities and spaces at NVIDIA, where he drives the adoption of
0:01:44 AI solutions for spatial intelligence and urban applications. With a PhD and extensive background
0:01:49 spanning chip design and leadership roles, Jumbi brings deep technical expertise and passion
0:01:56 to using AI to transform city operations and urban life. All of which is a longish way of saying we’ve got
0:02:03 the two best people I can think of here today to talk about what AI means to cities, to the people who
0:02:09 live in the cities, and to where our cities and our communities are headed in the future. So Mayor Matt
0:02:14 Mahan and Jumbi, thank you so much for joining the NVIDIA AI podcast, and welcome.
0:02:18 Thank you, Mel. It’s great to be here with you and Jumbi. Looking forward to our conversation.
0:02:20 Thank you. Glad to be here.
0:02:26 Absolutely. Mayor Mahan, I’ll start with you. Can you talk a little bit about what inspired you,
0:02:31 kind of your journey from the classroom to the tech world? And I’ll throw in real quick that
0:02:37 our stories are kind of parallel to there. I started off as a teacher as well, so a lot of respect for that.
0:02:41 But your journey that, you know, led you to now leading San Jose and how all of that kind of
0:02:47 shapes your vision for the city at large, but also, you know, thinking about and starting to use AI
0:02:53 in daily operations. Yeah, you know, no, it’s definitely a non-linear career path and not one
0:02:58 that I mapped out with any intentionality. I’ve just always been really curious about the world.
0:03:02 I’ve always loved problem solving. I think that’s what drew me to entrepreneurship.
0:03:09 But more than anything, I’ve wanted to think about how we make the world a better place. I grew up in a
0:03:15 little farming town and we had a lot of challenges in my neighborhood and my community. And it was coming
0:03:20 to San Jose when I got a work-study scholarship to a high school here that really changed my life.
0:03:25 I was exposed to the innovation economy, was privileged to get a great education, fell in love
0:03:31 with Silicon Valley. And I’ve just always wanted to solve problems in an entrepreneurial way, but apply
0:03:38 them to the biggest challenges we face of education and health and safety and climate and the housing
0:03:44 costs, the big issues that affect all of us. So I guess you could say that’s what led me here. I’ve
0:03:50 tried to bring a bit of that startup environment, entrepreneurial spirit into City Hall and run the
0:03:55 mayor’s office a little differently than it’s traditionally been done. And we’re having a lot
0:03:59 of fun just getting everybody focused on the outcomes we’re trying to drive and being a little less
0:04:04 worried about risk-taking and making mistakes. So we’re trying to bring a more experimental culture.
0:04:10 Fantastic. Jumby, from your perspective and kind of by way of introduction to the work that you do,
0:04:17 what are some of the biggest opportunities that you see for AI adoption in cities in particular?
0:04:23 Yeah, you know, first of all, I should say I’m super honored to be on the same show with Mayor Mahan
0:04:29 being a resident of San Jose. I can actually see some of the changes that the mayor is implementing
0:04:37 and they are pretty amazing. So, you know, there are just a large plethora of applications in cities for
0:04:45 citizen services, traffic management, public safety, energy optimization, infrastructure maintenance, etc.
0:04:52 But I’ll highlight two major areas where cities are seeing significant impact from AI.
0:04:57 And I think this, these will resonate with the mayor because I’m actually picking up some of the stuff
0:05:04 that they’re doing in San Jose. So, you know, the first big area is making city services more efficient
0:05:13 and accessible. So, AI-driven automation for providing city services, for example, permitting. And here again,
0:05:20 I believe San Jose is using it for streamlining ADU permitting, for example. And then there are other
0:05:28 services such as business licensing, filing forms, emergency relief support, looking for affordable
0:05:35 housing options. You know, AI can make the delivery of all these services, you know, much more streamlined,
0:05:42 comprehensive, comprehensive, personalized, and accessible. So, you know, just to highlight the
0:05:49 accessibility part, you know, the ability of AI to automatically translate the written and spoken
0:05:56 language means that city services can now be provided in a host of languages, which is, you know, of course,
0:06:02 important for a city like San Jose, whose residents speak over 50 languages, I believe. You know,
0:06:09 automated translation is also important, for example, for 9-11 calls. You know, imagine someone calling in
0:06:13 a language on the person on the other side, not even really understanding what that language is,
0:06:20 so they can find a translator. So, AI can automatically translate, understand the emergency,
0:06:23 and route the calls to provide the appropriate response.
0:06:29 You know, as I, not to interrupt, but as I was preparing for the episode, for the recording,
0:06:35 something reminded me of, I don’t know if either of you saw this movie or remember from,
0:06:41 this is 25 years ago, but during the first dot-com boom, there was a movie called Startup.com,
0:06:47 and it was these entrepreneurs trying to build, you know, state-of-the-art, then was web, you know,
0:06:54 web 1.0, right? But trying to build a web-based system where folks could engage with civic services,
0:06:58 and government. And I remember, if I remember correctly, one of the big dramatic points in the
0:07:03 movie was when one of the founders was able to successfully pay a parking ticket online,
0:07:09 you know? And so, I say all that just to say that listening to you speak about, you know,
0:07:15 translation is such a great example to bring that home of just that real-time. Yeah, so it’s amazing
0:07:15 how far we’ve come.
0:07:21 Yeah, it really is. And I’m, John B, I’ll return the fact, I’m honored to be here with you. You have so
0:07:24 much domain expertise here. It’s exciting to have this conversation. And as I noted,
0:07:30 as a San Jose resident, you are one of my one million bosses. So, I’ll be listening carefully
0:07:37 to what you have to say. We’re seeing in just incredible breadth of application for these new
0:07:42 tools. And John B spoke to many of them. Language translation is a great one. In our public meetings
0:07:50 now, without having to go hire dozens of interpreters, we have real-time language translation in dozens of
0:07:56 languages at every public meeting. And it’s two ways. So, I can be speaking in my native language,
0:08:01 English, and anyone who speaks dozens of other languages can see in real-time translation in their
0:08:06 native language. They can speak in their native language, and I can see translated on a screen
0:08:11 in real-time in English what they’re saying. And so, we can seamlessly use our native languages to
0:08:18 communicate in real-time in all public meetings now, again, without having to massively expand our budget
0:08:25 for this service. But to the point of permitting and those basic city services, we’re starting to use it
0:08:30 to automate the detection of potholes, to speed up bus routes. We’re just seeing so many applications
0:08:36 where we can take an existing city service and make it significantly better through automated
0:08:41 detection and service deployment and kind of building these more responsive systems.
0:08:45 Mayor, you said potholes, and I got a little triggered. I’m not one of your constituents,
0:08:50 so it’s not anything to you. But, you know, as we just started to scratch the surface of so many
0:08:56 amazing things that can be and are being done, what are some of the challenges when it comes to adopting
0:09:02 AI in a city? And, you know, this can be either, Jumby, to you sort of broadly from your work, but I don’t know
0:09:08 if Mayor Mahan, you’d be willing to take a stab at, you know, just some of the things that you’ve tried
0:09:12 to work on or would like to work on at the San Jose city level, but are just kind of challenging right now.
0:09:17 You know, I think innovation can be hard in government for a couple of reasons. One is we have a fairly
0:09:23 risk-averse culture because unlike a startup, which is the world I’m coming from, we’re really not allowed
0:09:30 to fail in a big way. First of all, we often have lives at stake, so we have to prevent that kind of risk.
0:09:36 But also, we’re the only game in town. That tax revenue is coming every year and we’re relied on to provide
0:09:41 essential services. If a startup tries something and it doesn’t work and they fail and they run out of funding,
0:09:48 there are often a hundred other startups in the space that are pursuing similar ideas. So we have a risk aversion.
0:09:54 We also have much more need in the community for the essential services we provide than we have resources for.
0:10:01 So it can be hard to carve out time, staff time and budget to try new things when we feel like we’re
0:10:07 underwater just scrambling to do the things that are already identified as core needs and core services.
0:10:13 But I’m lucky to have a great chief innovation officer, Stephen Cain’s on my team. We have very
0:10:18 forward-looking leadership at the city within the city manager’s office. I think we’ve all recognized we need to
0:10:25 carve out that time to pilot new tools because the only way we ultimately make government more productive
0:10:31 and do more with limited resources is through innovation. It’s technology adoption, but it’s also,
0:10:37 and I’m sure we’ll talk more about this, so much of it is about workforce development, giving people
0:10:43 permission to experiment, training them in new tools, giving them space to learn, to grow, to try things,
0:10:49 things to fail. That’s where I think it takes the most effort to break through and actually have that
0:10:51 kind of innovation and better outcomes.
0:10:56 Jovi, do you see similar challenges, similar patterns in the work you’re doing?
0:11:01 Yeah, absolutely. You know, one of the things I’ll highlight is that initially when we started really
0:11:08 digging into the whole sort of gov tech area about five years ago, there was a lot of resistance from
0:11:15 all kinds of constituents around AI. You know, they wanted to have the proper guardrails. They wanted
0:11:22 to make sure it wasn’t biased and all that good stuff. And, you know, Mayor Mahan and the chief
0:11:27 innovation officer, Stephen Keynes, are the ones who actually led the charge on coming up with a
0:11:34 framework. It’s called the Gov AI coalition, which actually was an initial attempt to lay out, you know,
0:11:39 know, some basic guidelines. And that, in my opinion, really gave a boost to a lot of cities
0:11:43 looking at it and saying, look, guys, you know, we have all these guidelines now. We could just, you know,
0:11:49 go for it. Let’s not just hold back. So rather than, you know, waiting for a massive amount of regulation,
0:11:55 you know, even these kinds of basic guidelines that Mayor Mahan drove initially, you know,
0:11:57 I think made a big difference in the adoption.
0:12:02 Well, thanks for mentioning that, Jambi. I’ll take the opportunity to note that our
0:12:11 second annual Gov AI Summit is in San Jose, November 5th and 6th. This coalition, if you will,
0:12:17 has taken off like a startup. We didn’t really know what we were starting, but I’m very proud of our team
0:12:24 at the city for convening other public agencies. And as Jambi was saying, we needed to create a permission
0:12:30 structure and a set of guardrails and guidelines to give people confidence to take risks.
0:12:37 And so this entity has grown into a platform for over 800 public agencies all over the country at
0:12:43 every level of government. It has AI policy manuals where people can take something off the shelf and
0:12:49 adapt it for their own agency, but they have those guidelines in place. It has tools like an incident
0:12:54 response plan. If something goes wrong, how do you dissect it and do a retrospective and learn from
0:13:01 it? It has a vendor registry where we can collectively initiate procurement by essentially understanding
0:13:06 what are the use cases? Who are the vendors? What are the, you might consider it like the nutrition
0:13:12 facts on your box of cereal. What do we know about these vendors? And then we have a library of actual
0:13:17 deployments of use cases where we say, well, here are all of the pilots we’ve run and here’s how far we
0:13:21 got. Here’s what we did. Here’s, here was the data. Here’s what we learned. And so it’s just
0:13:28 become this incredible platform for collaboration and it’s allowing public agencies to leapfrog off of
0:13:32 one another. We can’t test everything all at once in San Jose, even though we are a bigger city in
0:13:36 Silicon Valley. So we, we look to other cities and say, oh, somebody else is doing something really
0:13:41 interesting around procurement or permitting. Let’s go try that. We can jump off from wherever they,
0:13:44 wherever they are today versus starting back, you know, at zero.
0:13:48 Open source government in a, in a way, exactly. Yeah.
0:13:54 How are the city workers taking to using the new tools? And, you know, I think you mentioned earlier,
0:13:58 well, you mentioned the upskilling program in particular, and, um, I think you, you’ve trained
0:14:01 over a thousand city staff at this point, which is fantastic.
0:14:04 That’s the goal. We’re in process toward a thousand.
0:14:06 That’s the goal. Okay. Got it. Got it. Okay.
0:14:11 You also mentioned, I think, or maybe it just always pops into my mind, but you know, the notion of,
0:14:17 you know, AI comes up and you still have people who respond in different ways and often, you know,
0:14:21 maybe one of two, one of three ways, but one of them is, I’m not so sure about that. Yeah.
0:14:25 What’s it been like rolling this out to city employees and kind of the range of reactions
0:14:29 that you get and, and just kind of how’s the whole process going with the people?
0:14:36 Yeah. I think Jumbi would probably agree with me that the innovation isn’t just about new tools and
0:14:43 technologies. We often get very focused on the new app or the new, the new device, the new gadget,
0:14:51 and really the, the, the greatest unlock in terms of innovation really comes from humans wanting to
0:14:56 to use it, wanting to experiment, being creative in the application of new tools and finding new
0:15:03 applications with, of, of these capabilities. AI is so, so robust. The, the number of applications is,
0:15:11 is basically limitless. And so what we’ve tried to do is essentially create a, a system through which
0:15:17 our workforce can become more comfortable with and adept in applying AI and then turn them loose to
0:15:24 go apply it in their daily work. We have an MOU with Nvidia that includes a pillar around workforce
0:15:29 development. We also have a close relationship with San Jose State University, and it’s right here next
0:15:37 door to city hall. And we’ve created this curriculum for AI upskilling. And rather than do this in a top
0:15:42 down way where we came to our city workers and said, AI is coming. It’s a big deal. You should
0:15:47 be scared for your jobs and go take the course or you, you know, you’ve got to, you got to quickly
0:15:54 learn this thing. We did the opposite. We said, we see AI as not a labor replacement so much as labor
0:15:59 enhancement. And we want to empower you. We want to give you the training and tools you need to keep
0:16:04 growing, to be more productive and to have frankly, a more interesting job. Let’s take some of the drudgery
0:16:11 out of your day-to-day work by automating the things that are kind of rote and routine. And we saw a
0:16:18 great response. We’ve already graduated two cohorts of employees who voluntarily signed up for a 10 week
0:16:24 AI upskilling program. We’re in our third cohort. As you mentioned, the goal is to have a thousand
0:16:30 employees by the end of the year who sign up voluntarily, go through a 10 week upskilling
0:16:35 course. And then they go back into their departments. We have over 20 different departments. We like to
0:16:41 say we do everything from A to Z. We literally run an airport and we have a, we have a small zoo out at
0:16:47 Happy Hollow. We, we want them to go out to their departments and start building those custom GPTs and
0:16:52 finding different applications. And we’re seeing that happen. And it’s incredible. We’ve already saved
0:16:59 over 10,000 hours of staff time this year, just with those first cohorts beginning to
0:17:03 create these new workflows using, using these new tools.
0:17:09 Fantastic. I will, Jumbi, before I get to you, I just want to do a quick shout out promotional drop.
0:17:15 We had Dr. Teniente Mattson, president of San Jose State on the podcast. The episode came out
0:17:22 late September, wherever you get your podcasts. But yes, a very, she, a lot of similar themes talked
0:17:27 about. Definitely worth a listen there. Glad you mentioned her mayor. But Jumbi, I was going to ask,
0:17:32 you know, kind of along these lines, what do you see as some of the biggest opportunities
0:17:38 to help public sector employees drive civic innovation through the use of AI? You know,
0:17:42 and the mayor was talking about this. And this is kind of how I got into it myself in my own work and
0:17:48 learned really what works for my workflows and then how to share and learn from colleagues was just
0:17:52 kind of getting a sense of the guard, guardrails, right? What, what should I just not do in the work
0:17:58 environment? And then just playing with it and seeing how it could work in my daily flows. Jumbi,
0:18:03 what are you saying? So, you know, I’ll absolutely echo what Mayor Mahan just said that, you know, in the
0:18:10 public sector, I like to think of AI more as assistive intelligence. You know, there really is an artificial,
0:18:16 anything artificial about it. It’s real stuff that people can use in their real work to make it, you know,
0:18:23 more interesting to remove some of the sort of repetitive stuff that, you know, AI is now good
0:18:30 at doing. So I think the two broad areas of opportunity I see in this kind of innovation is
0:18:36 using AI to unlock, you know, potential of large amounts of data that cities have
0:18:44 to streamline services and provide sort of more holistic insights. And, you know, in the end, make
0:18:51 better informed decisions. You know, the ability to take large amounts of multimodal data, and that could
0:18:59 be, you know, data, audio files, videos, etc., consolidate it, analyze it, is where, you know,
0:19:05 AI can be incredibly useful. And, you know, creating these kinds of applications, as, you know,
0:19:12 Mayor Mahan was saying, is now, you know, a lot easier to do than it used to be even a couple of
0:19:17 years ago. In fact, you know, the amazing work that they’ve done in upskilling the city’s workforce
0:19:24 and giving them the training to create new tools, you know, with the amazing amount of data that cities
0:19:29 have, they can really truly become, you know, engines of civic innovation.
0:19:29 Yeah.
0:19:36 And, you know, one thing I’d add is, you know, once cities start unlocking the value in their data,
0:19:42 they may actually find opportunities to even monetize some of the data and the analytics
0:19:46 to boost revenues, which is, of course, you know, something all cities are looking to do.
0:19:51 Along those lines, Mayor Mahan, are there, you know, you mentioned potholes earlier,
0:19:56 and, you know, connecting the virtual dots, if you will, kind of from something like AI and talking
0:20:01 about data and, you know, online services are something that are made real, right? When a
0:20:07 citizen can log into the public works website or whatever the interface and have something happen.
0:20:12 But connecting it back to the physical world of core public services, services, you know,
0:20:16 filling in potholes, cleaning up waste, public transit issues, these kinds of things.
0:20:21 What impact have you seen in these core services at the street level,
0:20:24 directly resulting from AI-powered initiatives?
0:20:32 Yeah, I think that the availability of relatively low-cost cameras, other types of sensors that we
0:20:39 can put on vehicles or otherwise distribute out through the built environment, and then the incredible
0:20:45 data generation, real-time insight, data and insight that can be generated from that is really
0:20:50 powerful and we’re just scratching the surface. My favorite example in San Jose is bus route
0:20:58 optimization. We partnered with a local startup, Light AI, to put these sensors on buses and then
0:21:04 controllers on red lights to create signal prioritization. Signal prioritization for public
0:21:10 vehicles is not a new concept. People are probably familiar that many fire departments around the country
0:21:17 will give that signal prioritization to a fire engine when it’s going lights blaring down a road,
0:21:23 but it has been a very blunt instrument and one that has had downsides. So you’ll see a sea of green
0:21:29 lights for miles ahead as the fire engine goes down the road and it’s holding everybody else up,
0:21:36 creating congestion and creating all these negative externalities. This new AI-enabled system
0:21:44 is dynamic. It doesn’t guarantee a green light for every public bus, but it tracks the velocity of
0:21:51 the bus, tracks what’s happening more broadly with the signals and can dynamically lengthen or shorten the
0:21:58 duration of a red light so that cross traffic is still going, but you’re reducing the probability that a
0:22:03 bus ever hits a red light. And so what we’ve been able to do without massively disrupting traffic
0:22:09 flows in the city is reduce the likelihood that a public bus in San Jose hits a red light by over 50%,
0:22:17 which speeds up their overall bus route time by over 20%, which means on a route that’s a loop,
0:22:22 you can have better bus service by 20%. You not only have a faster commute, but you can increase the
0:22:27 frequency of buses stopping by a bus stop by 20% without having to buy any more buses, hire any
0:22:35 more drivers. We can literally just make your commute faster, have more bus passes and optimize the whole
0:22:41 system in a way that encourages public transit use, which has a further spillover effect of fewer cars
0:22:47 on the road and less congestion. And all because we’re dynamically monitoring velocities of buses and
0:22:52 slightly modifying the duration of signals. And it’s, it’s just a really cool, it’s like kind of a neat,
0:22:57 neatly packaged use case that has really meaningful real world effects for everybody.
0:23:02 Now that’s beautiful. And the down, the downstream effects you mentioned, you know, just fantastic.
0:23:08 Uh, Jumbi, you know, yeah, I was just going to add, you know, this is an amazing application and just
0:23:16 imagine now sort of extending this application across the city to not just prioritize bus routes, which is,
0:23:20 you know, of course important for public transportation, but really to orchestrate,
0:23:26 you know, all the lights in the city using AI. Uh, one city that actually deployed a solution
0:23:32 orchestrating literally all the lights across the city to optimize for traffic patterns, et cetera,
0:23:40 saw 50% reduction in overall congestion, uh, which means, you know, big reduction in times that cars have to
0:23:45 wait at lights, which also has a positive environmental impact because, you know,
0:23:50 as they’re waiting at the lights, you know, they’re, uh, have a bunch of emissions.
0:23:57 So, you know, intelligence signaling, uh, can, you know, like the mayor said, be prioritized
0:24:03 public transport or to make, you know, passage of law enforcement and emergency management vehicles
0:24:09 quicker and safer, but applied more broadly, it can actually have huge impacts and congestion
0:24:10 within the city.
0:24:14 I love that. You’re going to have to share that use case with us, Jumbi.
0:24:22 Absolutely. Our guests today are Matt Mahan, the 66th mayor of San Jose. Matt has been serving since
0:24:29 2023 coming from a background in entrepreneurship and the tech industry. And we are speaking with him
0:24:36 about AI adoption in San Jose and beyond, uh, more broadly in cities all over that can benefit from AI
0:24:41 from what San Jose has been doing, sharing out some of the work they’ve been doing, um, as we talked
0:24:48 about. And we’re also speaking with Jumbi Edelburum. Jumbi is an NVIDian who oversees global business
0:24:54 development for smart cities and spaces at NVIDIA. And we’ve been talking about some of the ways that
0:25:01 San Jose, other cities are starting to use, have been using, uh, and getting measurable results out
0:25:07 of bringing AI into day-to-day civic life planning, everything that goes into running and governing a
0:25:11 city. And, you know, this is no surprise to you guys, given your work, but as we’ve been talking,
0:25:16 as I’ve been listening to the two of you talk, just been thinking about the parallels between
0:25:22 what you’re talking about doing with AI in civic operations and overseeing communities. Guests we
0:25:28 have from enterprise, uh, running large factories and, and all different situations, medical research,
0:25:33 what have you. So many parallels, so many things kind of boil down to these, you know,
0:25:39 getting started with the data and the guardrails and making sure that the people using the system sort of
0:25:44 understand, you know, well, what does this do? What does it can’t do? What can I try? And,
0:25:49 and, you know, what do we have to really get, you know, more governance and policies and kind of
0:25:55 production ready mindset into before we get into, and just fascinating thing about all the things that
0:26:01 go day-to-day into, you know, being the mayor of a city, let alone one on the size of San Jose.
0:26:07 Mayor Mahan, San Jose just unveiled the winners of your first ever civic AI grant program. So along the
0:26:11 lines of all these things that go day-to-day into running a city and innovating for the future,
0:26:16 what kind of problems, what kind of, uh, issues are these startups solving and working on?
0:26:22 And what is a program like this? What does the investment mean for, for the city of San Jose
0:26:25 and the notion of AI innovation in cities nationwide?
0:26:32 Yeah, we, we, we, this was a first of its kind program. As far as we know, we’re, we’re trying to
0:26:40 create an AI ecosystem in San Jose that’s broad, diverse and touches every sector. And I, I think
0:26:47 downtown is a particularly good place to see the diversity of that ecosystem. We’ve talked a lot about
0:26:52 what we’re doing inside city hall. We mentioned our partnership with our wonderful public university,
0:26:59 San Jose State University across the street, which has 40,000 students in downtown and graduates more
0:27:03 students with technical degrees every year than any other university in the state of California.
0:27:07 Mayor, forgive me, but for people who, listeners who may not know the scale,
0:27:09 you mentioned the size of San Jose State. Yeah.
0:27:15 San Jose, around a million population, third largest city in California by population. Is that right?
0:27:19 That’s right. Yeah. We’re the largest city in Northern California with a million people,
0:27:26 only LA and San Diego and Southern California are larger in the state. And I believe we’re now 12th
0:27:32 largest in the country. Okay. So it’s, it’s a big city and we have big challenges, but it also means we
0:27:37 have the resources to be more innovative. And it’s particularly helpful that we’re here in the heart
0:27:45 of Silicon Valley with just an incredible workforce, incredible universities and the, the AI startup
0:27:52 incentive, the grant program that we created was a program through which we wanted to get to know who are
0:27:59 all the local startups working in AI that might want to work on civic and social issues, might want to
0:28:06 have a presence in our downtown next to our large public university. As we try to create this ecosystem,
0:28:11 we recruited plug and play to open an AI center of excellence downtown. The university is doing
0:28:18 incredible work. We have the MOU with Nvidia that I mentioned. And the other piece is making sure we are
0:28:24 more conducive to startup formation. I come out of that world. I help build two startups.
0:28:31 Barriers to entry are high. Housing costs are high. We have a lot of regulatory barriers in California,
0:28:37 and we just want to clear the way. And so we, we move forward to incentive programs. One was a very
0:28:43 simple office leasing incentive that’s open to anybody downtown, which says, if you lease new office space
0:28:49 downtown, you get, you have no business tax for two years and free parking in city garages for two years.
0:28:57 So that’s just kind of a baseline. But the incentive program or the grant program for startups solicited
0:29:05 applications for AI startups that wanted to be in San Jose, ideally work on a civic or social problem.
0:29:14 And we would give grants, which we did. We had, we were blown away. We had 170 applications. We chose our
0:29:20 four winners that are working on a diverse range of issues from maternal health to food waste reduction
0:29:28 to a satellite company and offered them the cash incentives, but equally importantly,
0:29:35 essentially a network and a community that they’re now part of. And we connected them with VCs,
0:29:44 with professional services firms. We had CBRE, the real estate brokers and, and then a venture
0:29:53 company, J2 Ventures offer pro bono professional services for the applicants and, and for the winners.
0:30:01 And, uh, and then NVIDIA very generously offered the ability for the startups to, to, uh, actually apply
0:30:07 for these, the inception program, which Joomby can probably explain better than I can, but we think
0:30:13 that’s a great opportunity. But this ecosystem is how do you bring all these resources to bear in one
0:30:17 dense environment to spark more innovation and entrepreneurship? And Joomby, do you want to
0:30:20 share a little more about inception and how, how the, that works?
0:30:26 Yeah, absolutely. You know, NVIDIA has the program, the inception program for startups.
0:30:31 And essentially it’s exactly what you said. We want to build a community of startups around
0:30:38 particular industry segments. It could be, you know, urban technology or, you know, smart city
0:30:45 technologies. And then of course we have healthcare and retail and manufacturing, et cetera. So, you know,
0:30:50 once these startups in the program, they get a host of benefits, you know, free cloud credits to test
0:30:55 their applications, et cetera. And of course, get connected to all the experts within NVIDIA
0:31:01 that can actually guide them in terms of product development. And all this from the NVIDIA side
0:31:07 is completely pro bono. So it actually makes for an extremely attractive package of stuff,
0:31:08 may or mayhem that you guys are putting together.
0:31:16 I think it’s great. By the way, I’ll just say, I think what we see when these ecosystems start to scale
0:31:24 and generate the next crop of great companies is you need mature companies like NVIDIA to help
0:31:29 facilitate. It doesn’t, these, they don’t just pop up by accident. They tend to be near large,
0:31:37 companies like NVIDIA and large universities like San Jose State or Stanford. And so we’re really
0:31:44 starting to see a lot of green shoots, if you will, here in downtown San Jose and in North San Jose.
0:31:51 Between the two, we have nearly 100 AI startups with venture backing, between them over $5 billion
0:31:57 worth of venture backing. So we’re seeing that next crop of startups come online, which is pretty exciting.
0:32:02 And by the way, Noah, if I could just add one thing, the mayor didn’t mention this, I’ll mention it.
0:32:09 But every month they actually organize a San Jose innovators meetup. In fact, there’s one this
0:32:14 evening and I’ve been going to those and kind of the networking goes both ways. I’ve actually met
0:32:19 companies there that we didn’t even know about and that weren’t in our inception program.
0:32:24 And we kind of introduced them to NVIDIA and how we can help them. But that also is a great program
0:32:29 where, you know, companies, again, build their networks within the city and they get exposed to,
0:32:35 you know, much larger bunch of companies, other collaborators, et cetera, in the region.
0:32:39 So that itself is an excellent initiative, mayor.
0:32:42 Well, thanks for mentioning it. I’ll see you there tonight.
0:32:42 Absolutely.
0:32:43 Absolutely.
0:32:49 So bringing people together on the AI podcast, that’s what we do. That community, that ecosystem
0:32:53 that you’ve both been talking about, though, is so important, you know, for the individual companies,
0:32:58 the startups getting involved. But then, as you said, it just it grows, right? You plant seeds and
0:33:03 you harvest them and you get, you know, you get more than the way the sum greater than the individual
0:33:09 parts, all those cliches. But it’s really true. It’s fantastic to hear about. Along those lines,
0:33:15 Jumbi, I wanted to ask you, technology, AI powered technologies, new technologies being used in
0:33:21 urban settings, government, civic, you know, urban communities. Are you coming across other
0:33:27 innovations, other things that you’re excited about that you’re seeing maybe in other cities or just
0:33:34 kind of in the startup world? Sure. You know, what I find particularly exciting right now is the
0:33:43 potential of the so-called digital twins and AI based simulation technology. You know, digital twins are
0:33:52 essentially digital representations of the physical world. But now, AI makes it possible to run all kinds of
0:33:59 of interesting simulations of what if scenarios that could be used to model, for example, the impacts of
0:34:07 urban development or extreme weather events or fires or accidents. You know, for example, if you see the
0:34:16 impact on traffic of a new development, you know, rather than giving citizens a 100 page traffic impact
0:34:23 analysis report, you can actually simulate the changes and show people visually, which is a lot more
0:34:30 compelling and understandable. And running these kinds of simulations can give, you know, public agencies,
0:34:39 get them much better prepared for development or emergency scenarios. And more importantly, be able to share
0:34:46 information with citizens in a compelling way so they too can be better prepared. So applications around those,
0:34:52 you know, across in cities across the world, I find is a particularly exciting area in terms of urban innovation.
0:34:59 To mention her again, I believe Dr. Tiente Mattson from San Jose State was talking about a digital twin project
0:35:05 they’re doing. And yeah, the possibilities for modeling and the implications of that are fascinating.
0:35:11 Yeah, we’re really excited about it. It’s an emerging technology for us. We’re not as familiar,
0:35:19 but starting to test the waters. And we’ve spoken with some of your colleagues at NVIDIA, as well as the
0:35:27 university. You know, the LA fires really caught our attention when it comes to better modeling things like
0:35:36 fuel load out there, the vegetation risk, which can be modeled, evacuation routes, and just really
0:35:42 getting a better handle on how do we both prevent and prepare for and then respond to disasters,
0:35:48 natural disasters when they happen. So we see, whether it’s that or climate smart or traffic
0:35:52 patterns, as you mentioned, Jumbi, we’re really excited about this emerging area.
0:35:57 Yeah. And, you know, going back, Mayor, to the comment you made about economic development,
0:36:04 of course, attracting businesses to the city, you know, they can even be used to visually show the
0:36:10 businesses, the areas within the city, you know, potential sites for locating their businesses.
0:36:17 And think about applications around tourism. You know, you can give tourists a much richer experience
0:36:25 through things like 3D wayfinding and giving them timely and relevant information on their surroundings,
0:36:31 you know, as they walk around the city. So yeah, just a plethora of applications. That’s a super exciting area.
0:36:36 I love that. Jumbi, if you ever want to leave NVIDIA, we’ve got a roll open for you over here at the city.
0:36:42 So you beat me to it, Mayor. I was going to offer him a spinoff podcast where we’re like,
0:36:47 we take, you know, leaders from different sectors and walk them through all the possibilities.
0:36:53 Yeah. Mayor, on more of a serious note, perhaps, what is next for San Jose?
0:36:59 When you’re thinking about AI and San Jose AI journey and already, you know, as we’ve said,
0:37:05 kind of taking a really bold forward approach, what are some of the projects, maybe technologies
0:37:10 or partnerships that you’re thinking about, maybe excited to announce in the near future?
0:37:19 Well, we have a lot of AI pilots in flight. We have the GovAI summit coming up, as I mentioned. So I’ll be
0:37:24 there and listening very carefully to what other cities are doing and what folks in the room who
0:37:32 are driving this innovation are seeing as next on the horizon. I’ll say for myself, when I think about
0:37:40 the efficient operation of City Hall, the two areas where I would most like to see disruption, if you
0:37:48 will, is really permitting and procurement. You know, I think where government often gets a bad name for
0:37:55 being slow, bureaucratic, unresponsive, struggling to move at the speed of business, tend to be in those
0:38:00 two areas, either when we’re making a purchasing decision of some kind, which can have diverse
0:38:08 implications. It can be a contractor for any kind of service we need or land use related permits.
0:38:14 These are two areas where we have objective standards. If you just take the permitting use case,
0:38:21 we have a general plan. We have zoning and land use designations. We have building codes. We have the
0:38:30 rule set, if you will. And then you have applications from typically professional investors, developers,
0:38:34 not always, sometimes they’re small business owners, but you have applications that have structured data
0:38:41 with certain parameters that need to be met. And I think there’s an opportunity here to take a lot of
0:38:51 the confusion, the back and forth, the unnecessary discretion out of the process and go from a process that can
0:38:58 take 12 to sometimes 24 months, depending on the complexity of the project, and bring that down to a
0:39:05 a matter of weeks, potentially. We are, and Jumby mentioned this earlier in our conversation,
0:39:10 we’re doing some piloting around ADUs. One of the things we found is that even though ADUs are very
0:39:17 popular, these are accessory dwelling units or backyard cottages, small, small backyard units that people
0:39:21 live in. Well, they’re very popular. The music studio I always wanted to build, yeah.
0:39:26 Yeah. Well, San Jose is a great place for them. We have a lot of single family home neighborhoods,
0:39:31 a lot of backyards that are underutilized, a great way to have extended family move in or
0:39:35 even generate income, rent it out to a San Jose State student or whomever. Yeah.
0:39:42 You know, even though we do a large volume of these, ADUs now account for 22% of all new homes built in
0:39:48 the city of San Jose each year. And we’ve done a lot of streamlining on our end. We discovered that we still
0:39:59 still return about 90% of initial ADU applications to the applicant to address missing or incomplete
0:40:07 information. And so we are now piloting an AI tool that’s essentially a little built-in co-pilot in the
0:40:17 application process to give people guidance and feedback in real time. So that the first time they submit, it’s a complete application. And we reduce all that unnecessary back and forth to get to
0:40:19 to get to yes faster.
0:40:27 Love it. Jumbi, where do you see potential for AI to transform urban life in short term in the coming few years?
0:40:55 No, absolutely. I think, you know, the mayor hit it on the head. There are these, you know, sort of painful things that citizens experience with cities or startups experience with cities. You know, the two, the one that the mayor addressed, permitting is a huge area. You know, just permitting, licensing, especially for small businesses, especially in a city like San Jose, where there’s such a large variety of businesses and languages that
0:41:03 that people, you know, people, you know, need to communicate with, that’s going to just be a game changer. And the other one, you know, the mayor referred to, you know, the sort of procurement process.
0:41:25 I think the rate at which, you know, technology is moving, especially in the, you know, the world of AI, both in the physical side, as well as the agentic side, making the process by which these new innovative companies can actually start working with cities and government agencies.
0:41:33 If we can figure out a way to make that quicker, it’s going to make a massive difference.
0:41:47 I mean, right now, as I’m sure the mayor well knows, I mean, there are all kinds of contracts and vehicles and requirements for submissions of a zillion, you know, things on paper. It’s just a long convoluted process.
0:41:51 And if we could simplify that, you know, we could take urban innovation to completely another level.
0:42:08 Yeah, completely, completely agree. We have to speed up our cycles for procurement, for learning, our learning cycles. We’ve got to be able to move faster. And I think AI is going to enable us to do that and still maintain quality.
0:42:10 Exactly.
0:42:15 Mayor, in wrapping up here, you kind of just answered my last question for you, but you can expand on it. Maybe I’ll ask anyway.
0:42:34 Last words of advice to other cities. Maybe there’s a mayor of another city out there listening, and they’re in the process, earlier stages in San Jose is, of adopting AI and trying to figure out, okay, we kind of understand what it’s all about, what it’s doing, how we get started, but what do we do now? What do we use with it as practices? Advice from you.
0:43:00 I would encourage folks to visit the GovAI Coalition. It’s hosted on the San Jose City website. There are now over 800 public agencies, and the whole idea is to create a safe, trusted environment for collaboration amongst public agencies that want to advance the frontiers of AI adoption for public sector use cases.
0:43:16 We have a number of documented case studies. We have a vendor registry, policy manual suggestions, just a very robust set of tools and very practical use cases people can pick up from. You don’t have to start from scratch. You’re not alone.
0:43:38 One of the really wonderful things about the public sector is that cities can be much more collaborative than competitive. Sometimes companies are worried about their trade secrets or losing talent to a competitor. And while certainly cities in the long term, the grand scheme of things maybe are competing on some level, more than anything, we can learn from each other, collaborate.
0:44:00 My view has always been, if San Francisco and Oakland are doing well, we’re going to be doing well as well. We’re an ecosystem. We’re not in direct competition. And so it’s kind of an open book. Come to the GovAI Coalition. Identify a use case that you think is relevant to your residents that your elected leadership would get behind and go test something.
0:44:06 You really learn by doing, you really learn by doing and there’s a lot of existing work out there to build on top of. So that’s where I’d start.
0:44:06 Amazing.
0:44:15 What I thought the mayor was going to say, and I’m going to say it for him because he didn’t, is hire some really good people.
0:44:16 Always.
0:44:31 Seriously, I know a bunch of people on the mayor’s team and he has hired absolutely top quality talent. And you can’t start all these great initiatives unless you have that around you. So, you know, kudos to the mayor for doing that to start with.
0:44:53 Jonby, you’re so right. And I, yeah, I, I, I should have mentioned it, but I will just say to your point, it’s not something, if you’re an elected leader and you’re listening to this, it’s not something you can do alone as the elected. You need those champions inside the departments. It’s, it’s really helpful if you can find a way to do this, either in the mayor’s office or the city manager’s office.
0:45:14 I mentioned our chief innovation officer, Stephen Keynes, having that one point person who’s kind of the trusted go-to for use cases, for what’s happening in other places, and being able to delegate that day-to-day, just kind of owning innovation for the agency is incredibly powerful. So thank you for making that point for me. You’re right on.
0:45:31 Absolutely. Fantastic. Jonby Edelburum and Mayor Matt Mahan of San Jose, thank you both so much. It’s been a great conversation. So much that can be done. It’s great to seeing it not only being done in San Jose, but shared in that spirit. We can get those rivalries out between cities on the sports fields.
0:45:45 That’s right. We have lots of avenues for that, but we can collaborate on raising everybody’s quality of life up using AI to help. So gentlemen, thank you again for joining the pod. Best of luck in all that you’re doing. And, you know, have fun at the event tonight.
0:45:46 Awesome. Thank you both.
0:45:46 Thanks, Noah.
0:46:16 Thank you.
0:46:20 Thank you.
0:46:33 Thank you.
Mayor Matt Mahan and NVIDIA’s Jumbi Edulbehram reveal how AI is making San Jose smarter—optimizing transit, translating meetings in real time, upskilling city staff, and powering pioneering civic programs. Learn how AI is helping streamline services and create a more responsive city for everyone.
Browse the entire AI Podcast catalog: ai-podcast.nvidia.com
Leave a Reply
You must be logged in to post a comment.