Skip to main content

In this interview, we will hear from Dave Selinger, the man behind the future of home security—Deep Sentinel. He’ll discuss how artificial intelligence is disrupting the trillion-dollar security industry and that one step he voluntary took to make it all happen.

—Dave Selinger

Raw and Real Entrepreneurship with Dave Selinger

Topics covered in the interview

David’s first business
Trading Data
Data and convenience
Deep Sentinel

Dave Selinger’s Bio

David Selinger was an early employee at Amazon, working directly under Jeff Bezos. He led the R&D arm of Amazon’s data-mining and personalization team. He co-founded Redfin (now a multi-billion dollar company) and founded RichRelevance, a company that offers personalized shopping experiences for large retail brands, including Macy’s, Barneys New York, Office Depot, and others.

He is now inventing the next BIG thing in home security – Deep Sentinel, an AI-based home protection. The company’s intelligent crime prevention transforms home security from false alarms and ineffective after-the-fact crime alerts to real-time crime prediction and prevention. With Deep Sentinel, Americans can gain a reliable, cost-effective way to protect their homes and stop a burglary, mail theft or driveway break-in before it happens – and feel dramatically safer at home, at work, and on vacation.

Follow Dave Selinger

Know more about Dave Selinger

Show Notes

Read Full Transcript

Dave Selinger 00:00
Welcome to Raw and Real Entrepreneurship, the show that dares to bring the no nonsense insight to those who have the coverage to start, grow and scale a business. Here's your host, entrepreneur, investor and best selling author, Susan Sly.

Susan Sly 00:16
All right, hey, what is up Raw and Real entrepreneurs? I have to admit, I'm fangirling over my guest today, for a whole variety of reasons. It's not just because of our mutual love of all things Star Wars. He was an early employee at Amazon working directly under Jeff Bezos. I don't know if we'll talk about that or not, but he led the R&D arm of Amazon's data mining and personalization team, which is very cool. He co founded Redfin, which is now a multi billion dollar company. So one of the big unicorns out there. And founded RichRelevance, a company that offers personalized shopping experience for large retail brands, including Barney's which I may or may not shop at from time to time. Office depot, which I've shopped at quite often as a CEO of a company. And now he is disrupting a whole new industry, home security. His company Deep Sentinel is an AI based home protection company that is specializing in intelligent crime prevention, which God knows we need, and transforms home security from false alarms and ineffective after the fact crime alerts to real time, crime prediction and prevention. We may get a little geeky today, and I will have show notes. And I'm actually going to have our team link to some different How Tos. So if we go down some rabbit holes with AI, or Star Wars trivia, but you guys can Google that yourselves. I will just disclaim that, because if you're listening to the show in one of our 141 plus countries, what you don't see behind my guest is he's got a whole stack of Star Wars memorabilia, but he also has different cameras in that radius. Of course, we do computer vision, so I immediately latched on to that. But without further ado, my guest today is the one and only disruptive, I don't want to get a F rating here but badass entrepreneur, David Selinger. David, thanks for being on Raw and Real Entrepreneurship.

Susan Sly 02:19
Susan, stoked to be here. Thank you. That was the most embarrassing introduction in history. I'll take all the Star Wars stuff credit for that though, because you know, that's a, that's a good stuff.

Susan Sly 02:28
Well, I have a mutual love of Porgs. They are adorable. So anyway, getting

Dave Selinger 02:34
that is my favorite thing back there by the way. With my Porgs, my kids and I will fight over who gets to hold the Porg when we eat together and stuff. It's, It's so cute.

Susan Sly 02:43
I might have a confession that at our house, Santa, two years in a row brought Lego Porg. And my littlest daughter Emery who's a huge LEGO and Star Wars fan, she was like mama, Santa brought me a Porg last year. And I'm like, Oh, yeah. Okay.

Dave Selinger 03:01
Maybe Santa wanted you to give that one to your mommy. Maybe that was it is. That's sharing.

Susan Sly 03:07
Exactly. Let me ask you this. David, what was your first business? You know, there's this like, are entrepreneurs born, are they made? How old were you when you started your first kind of side hustle?

Dave Selinger 03:20
So when I was, I was young is the answer. And I always, I don't know that I was born with it. But I was always displeased with inefficiencies. I was displeased with the way things were run. So like, in middle school, I had two kind of side hustles that I ran, I should, yeah, two. One was that as a computer kid in the 80s, if you knew how to boot up DOS, and like get it to work, then you were one in a billion. And so I would do consulting for families and teach them how to use their PCs. They, it was just the beginning of the PC era, so people would have bought it. And then there they are in their living room staring at this thing staring back at them, and it doesn't do anything. And so they would hire me and I'll come and I teach them how to use WordPerfect at the time, how to program in basic, how to get it to start at their different programs, how to backup their data, so they didn't lose it. And I thought I was just crushing. I think I was making like 20 bucks an hour as an 11 year

Susan Sly 04:24
Dude, that was like 200 bucks an hour back in the 80s

Dave Selinger 04:29
It was nuts. And I thought it was so weird. I could charge people this much money for teaching them this basic stuff. And they would do it hand, like can you come back next weekend? Right? I mean, just can you come back, can you come back because I mean for them, they would have spent you know 2500 $3,000 On this computer, and it was completely useless to them. So you know, now that I look back on that it makes perfect sense because here they are with this massive sunk investment that does nothing. And so that was my first one. My second one was I discovered wholesale. And wholesale, you can buy things for cheaper. And as a kid, I'm in middle school, you bought pens and pencils, and you go to the school store and they were like 75 cents. And I go to these, get these wholesale catalogs, you know, this is again, the age of the 800 numbers, you call an 800 number and you request the catalog and they send you a catalog. And then you can buy pens for three cents each, but you have to buy 500 of them. And so I would get together with a bunch of other kids. And we would buy wholesale pens and pencils, and then sell them at half as much as the school store to our, our other classmates. We did that with pens, pencils, and sunglasses, because Oakleys were really hot at the time. And you can buy the knockoff Oakleys. This is really kind of before the whole, like buy stuff on Alibaba Express and sell it to people. But it was kind of the same thing. You know, we buy them for like 12 bucks. I'm sure they really cost like $3 to manufacture but they're 12 bucks. And then we could sell them for 30. And we were making a killing and the kids weren't paying 100 bucks for a pair of Oakleys so, Shazam, Shazam, Shazam.

Susan Sly 06:08
Okay, so that is awesome on so many levels. So I have to digress as a child of the 80s as well. So, by wisdom, the listeners know that I was coding in the 80s. And my dad, you know, that was like, Pac Man. And like all these amazing games, which kids still play today. And so I was like, Dad, can I have some video games, and he's like, my dad is an engineer. Well, he's still alive. But he's not working as an engineer. And David, he's like, Go code them yourself. And I remember the first game I coded it was like, this, essentially like this, this thing that went back and forth across the screen, it was like a tennis ball bouncing up and down. And, you know, I felt like, you know, this is it. But I remember my dad saying, Susan, someday, everyone's going to have a computer. And someday those computers are going to be held in our hands. And that was, you know, circa 1985. Right? So let me ask you this sidebar. What was your favorite 80s video game?

Dave Selinger 07:14
Oh, man, um, I'll tell you the two, two or three that had the biggest impact on me. The first one was the very first one that I truly fell in love with was King's Quest One by Sierra Online. And what I loved about that one was up until King's Quest One, for the most part, games were like you just described like, Pong, right? You do this one thing over and over and over and over again. And King's Quest One was the first game where you explored an entirely new land. And Sierra made their name by creating this, this creative world around really terrible graphics, but you could explore the whole and, and it was the first time again, for the geeks out there, Zork, right, was just all text. And they had just enough graphics to make it more immersive. And you can really imagine that you were this prince kind of walking around in this land. And then the second one that I really, really liked, took that kind of the next level, which combined action and that which was Police Quest, also by Sierra Online. And I was such a Sierra OnLine fan, where they used to have a magazine, and I had every single issue of Sierra Online. And you mentioned my time at Amazon, I actually came to get to know this guy named Frank Kane, who is a little bit older than me. But he was at Sierra in those, those last few years. And it's just really interesting, you know, kind of the attitude of this was such a seminal part of my childhood, of the way that I view the world, of the way that I view computers. And to know that this guy was part of creating that just amazing. We're still friends on Facebook, you know, we only work together for like, I think a year is where we overlap, but I, he is like my favorite face, where everything he posts I look at because he just holds that really unique connection to my origins in terms of the way that I think and I view the world.

Susan Sly 09:21
And I that from that perspective, and looking at how many entrepreneurs who have been able to and you know, confess, you know, we haven't yet at Radius, but be able to become home and be part of teams that create these, these multibillion dollar companies. And there's something about how we thought about tech back then and it was, it was exploratory, and let's create and you know, we could be playing video games and then playing Dungeons and Dragons, right? And just kind of this, the creating these worlds. And before I fast forward time, so one of our engineers here, he has a building that he's built on his property, 3000 square feet. And it's full of completely refurbished 80s video games. Anyone you can make, think of sorry, he's there. You know, he's got Indiana Jones pinball. He's got Donkey Kong, he's got Pac Man, Miss Pac Man. Anyway, so, as a side, he and I are celebrating our 50th birthday parties within 30 days of each other. And we've decided to have 1 80s Video Game Party. So I just wanted

Dave Selinger 10:33
Oh my gosh, so sweet. I'll expect the invite anytime now. And wherever it is I'll fly there.

Susan Sly 10:37
You are so invited, you're so invited, and he's got glow lights, so if you weren't neon. So let's fast forward time. We're gonna go through several decades very quickly, because I want to talk about what's happening now. When I was, I just finished studying in MIT. And one of my professors, Dr. Rene Goslin, was talking about this whole concept of how we're willing to trade our data, David, for conveniences. So the convenience issue listed were temporal convenience, for cultural belonging, for monetary convenience. And with the work that you did delivering using AI at Amazon, and you know, what you did in terms of advertising for Macy's, for Barney's, here's the million dollar question. The people who say I don't want to trade my data, is it too late?

Dave Selinger 11:34
Well, I mean, I think it's $100 billion question, right. And I think it's a huge one, right. And I think about it now, a lot. And I thought about it at the time a lot as well. And, you know, I'll start with some of the thinking that if you've watched some of the more recent, like, there's a Netflix series about the way the social, and I forget what's called, the Social Dilemma, right? And I think that's an amazing, amazing film. I think it captures better than anything else that I've seen the mentality of the creators of these companies that take data and use it to engage their users and whether that's use it against them, or use that for them. And the word that you used was creators. And I think that's the right term to capture the mentality. I had the fortunate opportunity to be with a lot of the early founders at Facebook and kind of hear the way that they describe the work that they were doing and why they were doing it. And it wasn't to take advantage of people. Certainly the word, you know, get people addicted to what we're doing, that was around but it wasn't used in a negative connotation. It was because up until then, technology wasn't that engaging. Technology wasn't that interesting, technology wasn't something that you would find an 11 year old girl sitting in front of, using everywhere you went. And so I do want to be cautious in kind of the language that we use in this and the assumptions of intents that we assigned to these people. Because I don't think that there was negative intent. I very, very strongly believe that there was positive intent. And I and you know, again, we talked about the 80s, and then I'll bring it back forward. In the 80s Barbie launched a video game. I don't know if you remember, it was on like five and a quarter inch discs. It was around 1986, 8'7. And it was rejected by the market. And that was shown as proof that girls weren't going to use computers. And here I am talking to you and you're like, Look, I've only got two middle fingers but if I had more I'd share them with right now. And, you know, that was the prevailing mentality and it's fundamentally patriarchal, and sexist, and lame, and limited. And so when we analyze the events at the founding of Facebook, let's remember that that was the state of the world. And that was a prevailing type of thought. And, and as bad as some of these outcomes were, they came from a really good intent that to break through those barriers. That these barriers were huge. The glass ceiling was massive and meaningful and very real. So that's some of the backdrop in terms of my perspective to the answer. That I think a lot of these things were done with very, very, very amazing intent. And but unfortunately, my answer to you is, I think it is pretty freakin late. It's really, really hard. You have to be fully dedicated to tinfoil hat land, in order to keep your data out of the public access. I mean, if you look at what an average person will try to do on a daily basis, and just wake up in the morning, check my email, check my social networks, go to work, and then come home and cook dinner, and then go to sleep. In that context, the amount of data that we're leaving behind is just massive. I in fact, went to a browser, I use Brave Browser for some of my business work just because I don't want to get tracked and followed on on those cookies. I had 25,000 data points on a browser that I clear the cache on every single week. And then I typically browse incognito. So that, those are just like, the minimum of the minimum of the minimum of the minimum. In the past week, I had 25,000 data points. And so if you're going to work to try to get that to zero, the average person is creating what? 10, 11 million data points a week. And to to get that to zero, it's just almost impossible. You'd have to commit your life to doing that. And we're at the point where we've created so many conveniences of our life, that it would be almost impossible to not do the things that do create all this data.

Susan Sly 16:04
I love your view on it. And how you articulated it, and there are some interesting folks that are very committed the tinfoil hat, they're probably not listening to the show. Then there are folks like say, Ben Greenfield as an example, who said, okay, you know, my, if the Houthi building, I think he's going to use Moon wood, which, you know, is akin to sort of coating your house with silver. And then there are different things that he's doing. So he's choosing when to be in technology, choosing not to be in technology. For those of us in technology, I work with some amazing, creative people. And the question is always like, how do we serve the customer? How do we serve the end user? What can we do to deliver Wow? And to your point, this view that there's a bunch of people in Silicon Valley who are cooking up ways to destroy people, there are bad people everywhere, but that's, everyone I've ever met has been honestly a good human with good intent. So if we're leaving this trail of data, and you had such an amazing visual, I imagine it's almost like, you know, imagining a person and it's like this trail of stardust that they're leaving and how companies are choosing to use it is, that is the ethical dilemma that we have. But to your point, if we're willing to trade our data for certain conveniences, then what is the risk at this point? Because we look at, we're Gen X, you and I, and maybe our parents generation, they're not as cognizant of the data trail they're leaving, but we're seeing millennials, we're seeing Gen Z, who are starting to say, hey, we don't want our data everywhere. But we want all the conveniences. Is it possible?

Dave Selinger 17:56
You know, I mean, let's look at it from from a purely financial perspective, let's let's just grab Facebook out of the ether and pull them out of the metaverse, right? Let's go to Facebook, and let's look at them. So Facebook. Facebook has been a great place for the average American to stick 100 bucks a month for the last three years and become shareholders of Facebook. And why do I start with shareholders? Well, because Facebook is beholden now as a public company, even though Mark's still the largest shareholder to deliver the greatest value to their shareholders. And so when Facebook is presented with the dilemma of let's allow our users to just turn off all tracking, sweet. So Facebook generates, I don't know, a couple 100 billion dollars of revenue. So let's just cut that in a third. So let's drop 1/3 of our revenue by allowing our users to turn off all their data really easily. And so there's this incredible pressure from the shareholder base to say, Whoa, hey now like, I like, I like privacy, but I really like not losing my entire life savings that I just poured into Facebook, because that is exactly the way that the economics roll here. Right? I mean, there's a great saying, if you're not paying for something, you're the product, right. And so when you say convenience is the first thing that comes to my mind is all these various services that we use from Google Maps to get to work, to get to our deliveries, to Facebook that we use to communicate with our friends and our family. And that costs billions and billions of dollars to maintain these pieces of infrastructure. We either have to choose to pay for them through socialist means, meaning make them governmental, and we pay them with them through our taxes. Not meaning socialist in the, in the sense that the Republican Party likes to use the term or the sense of the Democratic Party likes the term, or we have to pay with them, pay for them directly. We have to pay with our credit cards every time we use them. Or we pay with them for with our privacy. And I think that the way that you started out which is, which one, do we as a society prefer when we have seen consistently as a society, even though there is this up swelling of concern about privacy, whenever it comes down to the ultimate decision, which is, do you want this thing? Of course, I want this thing. I want Google Maps. Okay, cool. Do you want to pay 10 cents every time you use Google Maps? Do you want to make it a socialist thing that is paid for by the Government? Or do you want to give up your privacy? Every single time that choice is presented to Americans. On a ratio of 90 to 10 or more, even in the greatest upswings of privacy revolt, it is, I want it for free. And I will trade my privacy for every single time, all day long, every day. And so as surprising as that may seem, when we look at it purely from a privacy perspective, when you put it back into the context of here are these things that I you know, you call the conveniences, but that I really need now, like I can't drive to my parents house anymore without using Google Maps. Like it just, it just literally doesn't work anymore. I haven't figured out how to drive. You know, we used to have this thing called TomTom. Where we have GPS in our car. And TomTom as a company went up and then bam, right? Nothing. And the reason is, because I would rather have it be online. And I would rather get it real time. And I'd rather get traffic information. And I'd rather rather give up a little bit of privacy, about my location than have to pay for this device that always gets out of date. And that's why in all of our cars today, we now basically have no longer have offline navigation systems. Again, when it really comes down to the economic choice, I just don't see Americans making the decision to recapture their privacy as a society.

Susan Sly 21:53
I don't either. And to that point, I love how you broke it down, David, because when we think about our day to day, there was a stat I read that in the UK, there are 10 security cameras for every one person. We know what's happening in China with computer vision, we know that you know, we think of it as not happening here. And I was writing a paper recently, and looking at how many cameras we are on every day as the average person. And depending if you live in the city, or you live in the country. So if you live in New York City, if you live in Washington, DC, San Francisco, wherever you live, even Phoenix, Arizona, you are on dozens of cameras every single day. CCTV cameras, cameras at your gas station, you are on security cameras at Nordstrom, everywhere. And so when you look at that, we're voluntarily trading data at times and then we're involuntarily trading our data at times. Which leads me to the next question because you are making a concerted effort and not just trying because in entrepreneurship, there's try and do, as Yoda said, there is no try.

Dave Selinger 23:11
There's no try. There's the spoon, and there's no try. But that's a mixing metaphor. Sorry. But yeah,

Susan Sly 23:15
So we had to quote the great Yoda. So just, I promised David before the show, we would get Star Wars in here. So there is no try. You are disrupting an industry and you are doing so from a place of having a career that has been in Silicon Valley, essentially. So considering that we're always on camera, considering that AI ability through vision has the ability to do things like notif, in real time, has the ability to learn essentially, that David and David's kids should be in the house and someone else should not be in the house because that person is not familiar. And this was all technology. We annotated Facebook's data. That's a whole other show episode. We created Facebook, we the people you talk about, you know, the socialism. And I'm not talking about it from a political bent, but we all participated. Is that David with you in the photo, Susan? Yes, it is. Is that David? Yes, it is. Is that your dog, Noah? Yes, it is. We trained all the data. So here we are in this day and age, we have all of these capabilities. And you're saying let's use that to keep ourselves safe. So talk about your current venture and that incorporation of AI and how it's disrupting what is already out there in home security.

Dave Selinger 24:36
Sure. So I'll start off just kind of giving the high level what deep Sentinel is. And so at Deep Sentinel, we solve the problem that if you, if you want to protect your business or your home, the majority of crime that's occurring now is outdoor. And so that means that alarm systems are just a thing of the past. They don't solve package theft, catalytic theft, car break ins, that's just not applicable. And so, you know, you've already mentioned cameras, cameras are accessible in those ways. But for anybody who has installed cameras at their business or their home, they also know that oh my gosh, I can either look at every alert 5000 times a day, which I can't do by the way, or I can just look at it after a crime and try to call the police and have them do something but neither of which is really protection. So Deep Sentinel, solves that problem by intervening in real time using two way audio, contacting law enforcement by identifying suspicious behavior using artificial intelligence in real time, and then intervening in that crime. So if you come into my property, the second you step on my property, there's an AI watching you. And you know, again, here you go, right, here's that cookie crumb trail, startdust of data. The second you walk into my private property, my cameras trigger and start tracking you. The second you start doing something suspicious, that AI decides that this requires a guard, and it sends your video to a guard. If for example, you loiter at my front door for more than 30 seconds, a guard will then speak up and say, Hey, this is Deep Sentinel security, I just want to make sure that you have business on this property. And that typically is enough to let a potential criminal know this is probably not the right place for me to be. And they scoot along the way. The tip to go into the technology and then maybe the ethics side of this, the type of technology we're using, obviously, like all AI companies in this day and age, or not all but all the emerging ones is called deep learning. And I don't know if you've talked about you know, your company, or kind of the origins of deep learning, but about six years ago, seven years ago, deep learning emerged from kind of the mire of all the research happening in artificial intelligence, and started winning competitions, that started winning, becoming the best at recognizing images of cat versus not cat, dog versus cat, hot dog, not a hot dog. And, and for those of you that watch Silicon Valley, I hope you got that joke. If you didn't, it wasn't a funny joke in the first place. So then, what it did, though, kind of fast forward to today, the computer vision and artificial intelligence technology that I can, as a private individual, go onto the internet and download in open source form is better what I can buy or get for free, not buy, get for free, install on my laptop running in my home, and use for facial recognition, video analysis, text analysis is better today for free than what the NSA was using 10 years ago on hundreds of millions of dollars of equipment. That's the kind of size of this technology revolution. And that's a true step function in technology. And that, like you said, we use it every day, we tag our friends and Facebook, Facebook's able to figure out who our friends are. Amazingly enough, if us Google Photos, one of my favorite features on Google Photos, which is kind of creepy, is I can find the pictures of my daughter, who's now 13, all the way back through her life, I can identify Oh, is this Abigail at two and a half years old? Wow. That's not just Abigail that was like, you know, the intersection of those two things. And that's an amazing step forward. For those of you that haven't followed AI, other than like, in the colloquial news, you probably know that in the 90s, we were talking about AI being able to take over the world and reaching sentience and passing the Turing test and, and tricking people into being a real person. For those of you that own Alexa, you know, that's not going to happen anytime soon. Because, hey, Alexa, turn off the lights. And you know, your entire house turns off. Hey, Alexa, add this to my, my, you know, my cart, and it figures that out, it's really good. It just repeats the same lines back to you over and over and over again, meaning we still have quite a long ways to go. But the key is that the technology at the center moved up more than 100 fold in terms of its quality in this incredibly short period of time. And so at Deep Sentinel, again, kind of to play this back to our company, what we do is we use that technology to make our guards faster and reduce the cost of that type of a service which used to be only available to folks like Jeff Bezos, folks that are celebrities and can afford $100,000 a year for security. Now that's affordable by pretty much everybody.

Susan Sly 29:29
There's so many things to unpack there, the open source stuff, right. So at Radius we're strategic partners with Nvidia and, and you know, Metropolis lab and all this cool stuff that we have access to. Our models are custom, but your point yes, there's ,if someone who wasn't even coming from a technological background, wanted to like say, you know, for the listeners, if David and I suddenly had no income, no companies, and you put us in a conference room and we're like, come up with a company in 12 hours, not only could we do it, we would be able to go and get that beginning tech for our MVP, or minimum viable product, we'd be able to get stuff from open source, you have no code stuff, you have all sorts of things out there that make it readily available for entrepreneurs. And I don't want to give away your secret sauce. But like my big, my big question like, is really run GPUs. But we don't have to go down. How many, how many cameras are, so let's say, I get Deep Sentinel for my house, like, where are the cameras located? Where, how many per house, like

Dave Selinger 30:44
Our average, our average residential house has between three and five cameras, you know, for, you know, 2000 square foot to 3000 square foot property. And it scales up and down, right. And it scales up and down based on how big is the property as well as how concerned about security to someone. So if you have a 2000 square foot property, but you're a judge, or a police officer, or, you know, we have a couple of police departments that are, you know, 3000, 4000 square feet, but they'll have 15 cameras, because they are serious security targets. And so it's like anything in security, it's really a function of the target, the parameters of the target, and then the concern around that. What's the actual security risk? And so, again, our average person now has three or four cameras. And, you know, for, for me, obviously, I have a lot more than that, because I, you know, kind of consider myself to be a little bit more of a target. But for me, I think that I would, in an average home, feel very comfortable with four or five.

Susan Sly 31:45
It's interesting too, because as you said, you know, we're looking at this whole real time alerts. And for people who are not technical, that requires a certain kind of compute. And we know this at Radius, because we do real time for retail and healthcare. We were doing, you know, work with hospital sectors, pre screening patients for COVID. And, you know, I won't say which hospitals. So there's a lot of compute that goes into that. There's also the last mile problem, David. So we're talking about vision, and we're talking about security. Radius is asked in the commercial sector. Can you do weapons detection? You know, yes, we can. And there all sorts of things. But here's the question in the last mile, it's the ethical question, what happens? So we are seeing in, you know, major cities, people running in David and grabbing, as long as it's not over $1,000 worth of items, the police aren't going to do anything. So yes, just because it's detected doesn't mean that something is going to happen. And so, you know, my last question for you is, how is Deep Sentinel solving the last mile problem when it comes to now we have the real time alert? We, you know, know it's not a false positive. But, you know, now we're passing it off to law enforcement. So how do you solve for that?

Dave Selinger 33:09
Yeah. So I mean, I think there's just a ton of different dilemmas in here, right. So you know, you mentioned the amount of compute that's going into these, these AI models. And you know, and you mentioned Nvidia. I'll give an opportunity to to just do a quick shout out to Nvidia which is a local company down here. The CEO there has been involved in a bunch of local nonprofits and you know, it's a company so whatever. But every time I have met him, he has just been amazing. And he, he would spend the time, they were really little, I mean, not really a little bit like, you know, very smaller, much smaller company would show up and donate everything they could, he puts his heart into these things. So anyway, it's easy to kind of get distant with people. Think of everyone as being as horrible as Mark Zuckerberg. But not every CEO is as bad as Zuck.

Susan Sly 34:01
Let's give some love to Jensen. Yes, sorry. This is another sidebar. For those of you listening, you're like, why are they talking about Nvidia? So I love a nonce over there. I do a lot of business with him. I just had dinner with Azita Martin, who's the head of retail AI for Nvidia. We have a regular cadence meeting with Nvidia. We love Nvidia. So, okay, so

Dave Selinger 34:25
Great. They're a great, great, great, great company. And they've enabled this entire revolution of AI because they made it in a very high level affordable for professors to do research in artificial intelligence. Google, as amazing as Google is, made it to the core to AI research was search. And search costs about $100 million to start doing researching, but Nvidia made it so that you could do research on a desktop that cost $3,000 and could do cutting edge research and machine learning again. And that was just phenomenal. So just quick shout out to them on the side. What happened because of that, though, is that we then started feeding these new problems. Images, and videos, and language into these models. And the ACLU, the NAACP said, Wait, hold up here, I'm noticing two things. I'm noticing that our facial recognition models are generally trained on white people, our suspicious behavior models, our airing towards people with darker skin, specifically African Americans and black Americans. And then we're giving this to institutions that have shown historical propensity to hire people that may or may not have these, these internal prejudices. So you have this kind of stack of behaviors that are new, and those stack of behaviors because they're new, have the potential to assume the biases of the human beings that created. And so that sentence that I just said there immediately became political, righ? It gets pulled on the left towards look, AI is racist, great. On the right to look, the left is being ridiculous. Both of them are right, right. And both of them have a very, very clear and important role to play in this. And so what we did at Deep Sentinel, as you can probably tell, like, I acknowledge the political leanings of individuals, and then I try to find my solution. I think that the solution will not be found in either of those political leanings, but that you have to embrace the existence of both of them in order to find a solution. So what did I do? I started a company that works with police departments, oh, zero, Republican. And then I reached out to the NAACP in the ACLU, and I said, Can I have a meeting with you and present you what I'm going to do and get your feedback? Oh, you're a liberal? No, I did both of those things. And, and I think that's the only way that we're going to be able to solve these problems in a sustainable way for our kids, right. So if your goal is just to make a pretty penny today, you don't got to do what I did. But if your goal is to build a sustaining enterprise that changes the world, changes the way that we view the world, change the way that we interact with people around us and the world around us, and then lasts for generations. I think that is the right way. So I work with police. I work directly with police, I send white people and I make calls for service against criminals who are white, against criminals who are Hispanic, and kind of criminals who are African American and black. I also, like I said, met with the ACLU and NAACP and I let them tear apart my technology and solution for eight hours, top to bottom. Answered every single question that I had about what we did at every step, and then built a roadmap to say how do you then remove, or at the very minimum, reduce potential for the biased human beings that are building this, to influence that process? And I took copious notes. And then what we did was voluntarily, not because we were regulated, not because the ACLU followed up with us, not because we got to, voluntarily took every single of those actions, and we follow them to the tee every single day. And they were things that, you know, frankly, I don't know that I would have agreed with going into that meeting.

Dave Selinger 38:19
They were some of them were things that I was just surprised at. Like for example, one of the attorneys that showed up to our meeting said, we would really like to make sure that every single time you call the police, you document what the ethnicity and race is of the guard that called the police, as well as the race and ethnicity of all the potential suspects. And I was like, Yeah, but if I do that, then that could go back into a machine learning algorithm. And then, you know, we can start predicting that we should call police more frequently on Asian Americans, for example. And they said, Yes, we'd ask you to continue to segregate that. But if you don't do that, you can't enforce that. And I had literally been doing the exact opposite. I've been doing kind of the typical technical, let's just be race blind, and not tracking that. But that, in that meeting, they walked me through the the best practices and procedures to make sure that we weren't creating bias. And then in the event that we had hired an employee that was being biased, we'd be able to identify that, track it and then remove it from our systems. And so there were a lot of things like that. But what it took was, what it took to get to that solution was being willing to listen to something that I found very disagreeable at the beginning, to give it 100% of a chance, and then to just do it. I mean, I'm very proud of what came out the backside. Let me, let me put it that way. I'm very proud that we met with those groups. I'm very proud that we were able to survive that meeting. It was a long, hard meeting. It did not start out well. But unless you're willing to hear the opinion of people that you're completely in disagreement with, you're never going to find a solution. And so, at Deep Sentinel, that's been our way of doing it, we get feedback from police, I get feedback from judges. One of the other neat things that I'll share with you that we, that came out the other side is to make sure that law enforcement agencies that we work with, we partner with, in a way that enforces the freakin rules. This is a nation of laws. And while I am not a law enforcement agency, making sure that we are holding those that we interact with when we're, when we believe we're following laws accountable. So very specifically, we will receive a subpoena or a warrant for some video footage. A subpoena can technically be signed or physically be signed by anyone, right? Like I can just write, Judge, you know, Clarence Thomas on something and then sign it and send it to you. And you may, if you don't verify that, in the same way that in technology, we use two factor authentication, we use ACH X to take care of things, a lot of people, when they receive that, will just ship the information back. And that's actually what's required. That's the minimum standard that's required by law. What we do then is we will call the judge's office, we will say I need to confirm that this judge signed this warrant. Okay, this is his clerk or clerk, and we're confirming it. Okay, great, excellent. If you can do that, I'm sending you a form that I then need you to sign. I'm sending that form to your official email address, you then need to respond from your official email address, with that form signed, confirming that this is an official warrant. And I won't name names, but I will tell you that there are times when it does not come back confirmed, and that police officers have certainly taken the liberty of sending us a warrant that was not fully signed by a police officer, I'm sorry by a judge. And so and again, you know, this is one of those things that I was sitting there in that meeting. And they said, Oh, it's going to happen. And I was like, There's no way that's going to happen. Sure enough, within six months, we got our first like, not confirmed search warrant.

Susan Sly 42:13
Yeah. And it's that last mile piece, I love how your willingness to say, I'm going to, the question always in technology is who is the end user. And sometimes, people get it wrong in tech who the actual end user is, right? So it'd be very easy to assume, if I was looking at, you know, a pitch deck that the end user is actually the customer who purchases Deep Sentinel, but the end user here, going even beyond that, could be the police, could be the judge. And they're, you know that as we peel back the layers of the onion. And that value in that, David, being so willing to listen to the NAACP, and the police. And in insecurity, one of the things we, a time and again, so a million years ago, in a land far away, I was in law enforcement. And because Radius deals with enterprise level customers, traditionally, the head of store security is former law enforcement. And one of the things I said to our team is, we will ultimately have to get their feedback, because if we're going to share camera views, streams within locations, we have to have those relationships which have been valuable getting their feedback, answering their questions, to your point, as hard as it is. Even, we're doing, you know, doing things to help the employees that want to be there using AI. And one of my partners, and I sat in a room for two hours, could talk about the creepiness of AI and answer the questions and to get the feedback. And so what you're doing is so incredibly, not just valuable, but in my opinion, it is the way the next evolution of tech companies will evolve from. That willingness to be transparent to get that open dialogue. Because just like you and I, were children of the 80s. And I have one final question for you, not even related to this conversation, just that creativity and how we came out and problem solving and things like that. Those of us who are seeing that there are companies like Meta and what's happening on Instagram. I've had so many fake accounts. And they started to direct message my children and I just posted on Instagram about its child endangerment. They're messaging my children as me .aAd my youngest child is 13. She doesn't have Instagram, but my next one is 16 and they're DMing them as me, their mom. And so those companies and their unwillingness to be transparent to get the feedback of the users, they will be forced to evolve. And that's why I love what you're doing because you're getting out in front of it. And that next wave. So my final question for you, you ready?

Dave Selinger 45:14
Yeah. Okay, entertainer interjection, I actually met with these groups the same week that Mark Zuckerberg was testifying before Congress. And so my first hour was getting yelled at as if I were Mark. And I finally just stood and said, Look, guys, I get it. Like, I'm not a huge fan. I think Mark's done a lot of amazing things. But like, I'm not him. I'm here voluntarily. I called you here. You didn't call me here. And so let's use this time together, as if I'm actually wanting to listen to what your feedback is. Let's just pretend that's the case. Because if you just yell at me, we'll be done with this eight hour session and we're not going to have gotten anywhere. And maybe you'll feel better but I'm not going to change and I'm not going to grow. I'm here to grow. So if you want to do that, I'm here. And they're like, Well, honestly, no one's ever done that for us. And so thank you. And it took him about half an hour to kind of change gears. But I do, I do hope that you're right. I don't know if I yet believe it. But I really do hope because technology has to be part of our future as a society and it can't just be businesses or private interests, or social interests. All of those things have to come together at a certain point. And if we allow them to continue pulling us apart, we are the ones that are pulling ourselves apart, we will lose and have to make the decision to stop that. So sorry. Good. Final question. You you make it happen.

Susan Sly 46:38
Oh my gosh. And then that leads me to the whole piece around Elon and Twitter. And I'm celebrating that because of all of the fake accounts and the value and the fact that Twitter isn't willing to admit how many bot accounts and fake accounts actually are there. Mark hasn't admitted how many bot accounts and fake accounts are on Instagram. I mean, the list goes on. That's a whole different show. So my final question, David, and honestly, I just, I will say this in front of like, everyone. I love what you're doing and I am so excited that I get to meet you and I'm so excited for everyone listening. Please share the show everywhere. Tag us, let's hashtag David is not mark, that can be the show tag.

Dave Selinger 47:24
Thank you.

Susan Sly 47:25
All right, here we go. Favorite 80s Rock Band, favorite 80s Rock Band.

Dave Selinger 47:36
Oh, man, so I'm gonna have to, this is not going to be super popular. I'm gonna have to go with Gin Blossoms. For the simple reason that that was the first concert I went to with my friends. And it made me feel kind of part of the world. It's not my, it's not, it's not the music that I will listen to the most but it's the music that connects me the most with my friends from that time period. I got everyone together. I convinced everyone to buy a ticket and we all went together. So I'll go Gin Blossoms.

Susan Sly 48:11
Gin Blossoms. Awesome. Mine, for the concert reason would be in excess. My non rock favorite is Run DMC. I love Rev, Run. Like it's kind of like, shameless.

Dave Selinger 48:30
That's him right next to me in that picture literally right behind me.

Susan Sly 48:35
I had, I had, David, I had Jesse Itzler on the show. I had Jesse Itzler on the show and you know there was a hot minute where Jesse been doing some managing for Run DMC. So Jesse's on the show and at the end, I challenged him to a rap battle because he was on the first white rapper and he wouldn't do it. So my claim to fame is I beat Jesse Itzler in a rap battle because if I challenge you and you don't it, then I'm winning. And I did young MC. This here's a jam for all the fellas, tryin' to do what those ladies tell us. Anyway. So I won't do the whole thing.

Dave Selinger 49:11
I don't think I will challenge you to at anthying. I think that's just a bad move.

Susan Sly 49:17
So we've talked Star Wars, we've talked 80s, we have talked AI, we've talked technology. So I encourage everyone to go check out David's company, Deep Sentinel. And follow David on Twitter on LinkedIn. Instagram is not my favorite. I am not endorsing that platform, right. So with that, thanks, David so much for being here. I truly appreciate you and I wish you well and I can't wait to see Deep Sentinel on the latest CrunchBase list of you know, Ultra unicorns. And with that, everyone, thanks so much for being here. God bless. Go rock your day. This has been another episode of Raw and Real Entrepreneurship.

Follow Susan Sly

Check on previous episodes

Susan Sly

Author Susan Sly

Susan Sly is considered a thought leader in AI, award winning entrepreneur, keynote speaker, best-selling author, and tech investor. Susan has been featured on CNN, CNBC, Fox, Lifetime, ABC Family, and quoted in Forbes Online, Marketwatch, Yahoo Finance, and more. She is the mother of four and has been working in human potential for over two decades.

More posts by Susan Sly