Conversations on Applied AI

Josh Braaten - Optimizing Systems to Create Better Outcomes Using Machine Learning

Justin Grammens Season 2 Episode 3

The conversation this week is with Josh Braaten. Josh is the CEO of Brandata where he and his team use brand measurement, consumer insights, and growth solutions to turn fast-moving companies into market leaders. Prior to Brand Data, Josh was a Senior Director of Marketing at Lead Pages.  From venture-backed startups to Fortune 500 companies, Josh has developed his skills in high-growth digital marketing environments by working with brands that want to grow fast. As a business consultant, turned technologist turned marketer turned entrepreneur, I can't wait to talk with Josh about his career path and where he sees the future of artificial intelligence going in a variety of industries.

If you are interested in learning about how AI is being applied across multiple industries, be sure to join us at a future AppliedAI Monthly meetup and help support us so we can make future Emerging Technologies North non-profit events!

Resources and Topics Mentioned in this Episode

Enjoy!

Your host,
Justin Grammens


Josh Braaten  0:00  

I think a lot of what you're seeing with machine learning and AI today is just a set of steps and processes that analysts would typically perform by hand. It's just done automatically. And so an analyst back then would say, look at revenue, what are things that would lead to revenue or be predictive of revenue? And then how can we optimize whatever systems are in place? And in that case, it was people systems in workflows and patient care flow. And I think that's exactly what a lot of people are trying to do today look for, well, what is the better outcome trying to find those features that are predictive of those outcomes and then train some sort of a model that will help them understand the path to get there.


AI Announcer  0:44  

Welcome to the conversations on applied AI podcast where Justin Grammens and the team at Emerging Technologies North talk with experts in the fields of artificial intelligence and deep learning. In each episode, we cut through the hype and dive into how these technologies are being applied to real world problems today. We hope that you find this episode educational and applicable to your industry and connect with us to learn more about our organization at applied ai.mn. Enjoy.


Justin Grammens  1:15  

Welcome, everyone to the conversations on applied AI Podcast. Today we're talking with Josh Braaten. Josh is the CEO of Brand Data where he and his team use brand measurement, consumer insights and growth solutions to turn fast moving companies into market leaders. Prior to Brand Data, Josh was a Senior Director of Marketing at Lead Pages from venture backed startups to Fortune 500 companies. Josh has developed his skills in high growth digital marketing environments by working with brands who want to grow fast. As a business consultant, turned technologist turned  marketer turned entrepreneur, I can't wait to talk with Josh about his career path and where he sees the future of artificial intelligence going in a variety of industries. When Josh is not growing brands, he's volunteering for the Minnesota Search Engine Marketing Association, and exploring futuristic technologies like virtual reality or spending time with family most likely cooking. Thanks, Josh, for being on the program today.


Josh Braaten  2:01  

Thanks for having me, Justin. 


Justin Grammens  2:03  

Awesome. Well, great. Well, I gave a maybe a brief description of where you're at today, but maybe you can give a maybe a little bit more detail with the background about you know, who you are, and and sort of what the trajectory of your career has been? 


Josh Braaten  2:14  

Yeah, absolutely. So I kind of had a varied career. But I think everything that I've done is kind of added up to who I am today. So started off working in hospitals as a revenue cycle consultant working on, you know, numbers, based projects to try and figure out how hospitals could get paid by insurance companies more so that the keep their doors open, provide services, from there kind of transition to the IT side of healthcare just because I liked you know, the the data side of things, ended up working on a lot of web projects I in that capacity and ended up getting sucked into marketing, because at that time, SEO, and analytics, and some of these new advertising tactics were just emerging, and found those to be really interesting as well got into those and sped more than a dozen or so years working on digital marketing, in all sorts of different settings. And then for the last about five years, we've been running brand data. It's our brand measurement and growth agency, we've been helping clients big and small. Basically, any brand that wants to be the the the leader in its category, we can help them develop metrics and measurement plans around their brands. And finally, just in the last year or so we've been self funding our own projects, snap foo, which is an AI powered dog poop analyzer that helps dog owners raise healthier, happier, longer living dogs. So that's kind of a long road, but I swear it's all related. It's been a fun one.


Justin Grammens  3:41  

I love it. I love it. I know we'll dig a little bit more at a snafu here further on the conversation. But you know, I was wondering as you were sort of working through your, your career, it sounds like there's a lot of data, you know, you mentioned revenue cycle consultant, it sounds very dry, I guess but you know, you you have a background in math, I guess you're or something with regards to numbers.


Josh Braaten  4:02  

I was an econ major in college. And from there consulting was really appealing to me because I got to go and travel a lot got to impact companies at a higher level and really received a lot of I guess, training that that I felt was a little bit more beyond just an entry level college graduate. So really enjoyed that first role or that first job as a revenue cycle consultant. And it set me on a course of always looking at like, well, what what are we trying to accomplish? So the idea of what's what's the end goal or trying to keep that end in mind and in the for hospitals, it was trying to make sure that they weren't getting their claims denied by insurance companies, because for some reason or other, the hospital didn't follow their procedures correctly that are in their contract. So either the patient didn't have the right type of insurance or they presented without their card or they didn't get the right authorization. All these different things in network out network are reasons why companies are insurance companies don't have to pay hospitals. And so we would help them set up different policies and practices and teams and workflows to streamline those processes to make sure that they just asked that patient for that insurance card, a lot of things that are very routine now, we're not that that way in the early 2000s. And so these best practices were were set because they lead to better outcomes on the back end.


Justin Grammens  5:22  

Yeah, for sure. I mean, are there some things I'm just sort of like spitballing? Here, but are there some things that now it's whatever it be 15 years in the future, whatever it is now, bringing it back to artificial intelligence or machine learning? It's are there data practices that you've learned over your career, maybe even through the marketing, all the work you're doing with snap foo, stuff like that, that you could maybe see being applied back then. But it was obviously way too early?


Josh Braaten  5:45  

Absolutely. I think a lot of what you're seeing with machine learning and AI today, is just a set of steps and processes that analysts would typically perform by hand. It's just done automatically. And so an analyst back then would say, look at revenue, what are things that would lead to revenue or be predictive of revenue? And then how can we optimize whatever systems are in place? And in that case, it was people systems and in workflows, and, you know, what are the things that would predict those higher revenue outcomes? And, and I think that's exactly what a lot of people are trying to do today look for, well, what is the better outcome, trying to find those features that are predictive of those outcomes? And then train some sort of a model that will help them understand, you know, the path to get there?


Justin Grammens  6:34  

Yeah, yeah. Well, it's kind of dovetails into my typical question that I asked people is, how would you in generally define AI? I know you're using it in some context very much in your startup, but probably also in some of your day to day stuff. You're doing it brand data? You know, if somebody were to ask you on the street, how do you define it?


Josh Braaten  6:52  

Yeah, I think it's, it's very interesting, because it sounds futuristic. But I think it's really just more about the ability to measure and respond to patterns that humans can't fully understand. There are all sorts of data points. And I think we tend to fixate on certain ones, because they end up being more often than not the best metrics to look at like revenue, it's pretty hard to say that revenue is not a good metric, if you're trying to grow sales. But there are certain data signals that can predict whether or not that revenue is going to happen. Or if you lead to more revenue, if you know what to look for. And, and so that's where AI has really helped uncover a lot of value in all sorts of different sectors that that humans didn't didn't see before. You know, there's the classic story. I think it's like over a decade now, where there was a man who received an email from target that said, Congratulations on your pregnancy, and it was, you know, towards his daughter, and he was really offended that target, got it wrong. And turns out if you buy, what was it, prenatal vitamins, cotton swabs, and one other thing, strong correlation with pregnancy. And it turns out that target understood that, that his daughter was pregnant before he did. And he ended up having to like, apologize to her. So I love the idea that technology can do that for us. And it's just a matter of getting the right signals. And being able to tie those to the, to the outcomes secrete those patterns.


Justin Grammens  8:13  

It's like an ocean of data, and how do you find sort of that, you know, that one cup that you're looking for sometimes, but once it does happen, it's pretty amazing. You know, you can start spotting trends and stuff like that. And, you know, one of the things that I like to ask people is, you know, what are some of your greatest strengths and weaknesses? It seems like you are a trend spotter, I guess, of sorts. But yeah, maybe as you take a look back from sort of laying on what you're doing today, but also in the past, what are some things that you find that you're doing well, or some other things maybe you could work on just in general?


Josh Braaten  8:42  

Boy, you know, I think, you know, strengths are definitely looking at being a trendsetter, entrepreneurial, tend to try to maximize whatever things that I'm working on to make them as good as they can possibly be. I think there, there are some downsides to getting attracted to new trends, because oftentimes, you can be early on things. For example, I walked around for a year with Google Glass on because I felt like very confident, I'm still very confident that we will live in a world with a heads up display, but I was completely off. And I misinterpreted what the value of that of that experience was. And also like what the negative kind of externalities of that technology were for other people and how that friction might, might not or might play against the adoption. Right. So so I don't know I think it's it's always interesting to be interested in in the future and seeing where we're gonna go. But I think it's easy to get blindsided by kind of the different forks in the road and what people end up finally doing,


Justin Grammens  9:43  

Sure, preaching to the choir here. I do find myself going off the deep end, especially with new technologies around Internet of Things and in some ways AI and ML for you know, you've been going on it for many years now. And sure, there definitely are some spikes some certain areas that gained a lot of traction but also A lot of areas where we think it can be used anywhere at any time, you know. And so you sort of run through into this Gartner Hype Cycle curve. If you've heard that,


Josh Braaten  10:07  

I feel like I'm always five years early on things and tend to either get to things right before that trough of disillusionment, or stick with it long enough, in certain cases to get to that productive plateau, that folks are looking for as, as new trends become mainstream. So I think, for me, it's just as I get older, I try to temper that enthusiasm for new technology with the pragmatism of Well, how long is this going to take? Do I have the time capital or interest to to actually bring this trend to someplace where I can call it a hobby? Or maybe even monetize it in some way?


Justin Grammens  10:43  

Yeah. So you've been very much into marketing, it sounds like and obviously new trends, Google Glasses, that sort of that sort of out there, for sure. As well, I guess, talk me through how you've been looking to use AI and ML at Brand data. Let's talk a little bit about your startup as well. Some of the ways in which you're sort of seeing artificial intelligence or machine learning being applied to businesses.


Josh Braaten  11:04  

Yeah. So on the brand data side with our clients, it's, it's fascinating because advertising right now is, is in complete upheaval, you have these competing dynamics, where privacy is actually taking away a lot of data signals that people have been using to optimize their advertising efforts. So for example, today is January 19. It's important because today's the first day where Facebook will no longer allow advertisers to use certain targeting characteristics like health interests, political interests, and other sensitive topics that could be deemed, you know, violations of privacy. Up until now, advertisers were able to if you wanted to find somebody who was interested in certain political beliefs, you've just typed it in and said, I want to target people with these beliefs or people with a certain health condition, I want to I want to advertise to these folks, very, very easy. And that's been driven by, you know, a decade and a half of performance marketing and, and everybody is seeing ROI on these types of ads. But now, it's, you know, there's that backlash when it comes to privacy. And this is happening at the same time where the platforms are, are getting so good at optimization, that they might not even need some of the signals that we traditionally have had. So for example, I've talked to a lot of different brands, performance marketing teams, agencies that work for brands been talking a lot about this. There are a lot of folks that are saying, you know, even the best first party data systems where you have all of the downstream data on a customer all their purchases, all their channels, all their touchpoints, a lot of different demographics, psychographic, you know behavioral data, all those different things, even if you have all that and you feed that to the platforms, typically, we're in a lot of cases, now the platform's themselves, they'll do better if they if you just give them a list of your customers. And because what they're saying is that we have more signals on these folks than you could even dream of capturing. So that just tell us who they are, we'll find people just like them. And so it's interesting, because we're seeing this title Riptide of data going out with all this loss of privacy data, but the question is, will it matter? Will we see performance gains long term? Or will the platform's catch up and, and help us advertisers continue to see the types of performance that we've seen for the last, you know, couple years?


Justin Grammens  13:24  

Wow, that's fascinating. Yeah, I haven't been tracking that in terms of like, when that was going to happen. And so So can I, I mean, if it comes down to like race, and gender and age, can I not do that anymore, like Target a white person, age 32, you know, in Minnesota anymore, or


Josh Braaten  13:41  

Those are just basic demographic characteristics that still are allowed to be targeted. We're talking about more granular political beliefs, health related data, things that are considered a little bit more sensitive. And I think the idea there is, is that they want to make it a little bit more safer experience for people to browse on the internet. So that if you Google something that's very personal, that all of a sudden, like you don't end up with 17 different Facebook ads that are talking about something that maybe you feel is private or a violation of that trust that that you granted the advertisers that, you know, that are on the internet.


Justin Grammens  14:20  

I'm not sure if it's just, you know, like luck of the draw, or what have you. But I have seen sometimes it feels like I've Googled something, and then I've gotten an email related to something I just searched for, you know, so it's like, off its tip. There's a lot of something going on behind the scenes, some sort of tie back to that. 


Josh Braaten  14:35  

A lot of those things are not coincidence. There are, you know, explicit advertising, checkboxes and tabs that you can optimize to make sure that people see things, you know, in those circumstances.


Justin Grammens  14:47  

Sure, sure. Well, they have a treasure trove of data. I don't know I've always sort of thought as long as it's being used for good and consumers find value in it. It's it's not, you know, harm to me, per se, but it certainly gets a little distance concerning how granular that data gets, and how much they know about you, and then obviously, once that data gets out, it's leaked, it's probably more of a larger concern.


Josh Braaten  15:08  

I think everybody has their own personal philosophy about this, I, I personally understand how much data they actually have about, you know, us and all these different things. So I opt into a lot of things because I want to use them, but then I turn all the features on, because I know they have it on the back end. So why don't I see what they have on the front end. So for example, location sharing on Google Maps, it's a very easy way to tell, you know, friends and family that you know, I'm out running errands or I'm safe. When I'm on long road trips, that kind of stuff. Just turn on location sharing, invite a person for 12 hours or so while you're on your cross country road trip. And you don't have to worry about the phone calls and texts you when you get there. Like all that kind of stuff. It's your new version of that right. And, and so that extra kind of transparency creates the extra utility, if you know how to look for it.


Justin Grammens  15:56  

Yeah, I still there's a use right? for it. And in your mind, you don't care that Google knows where you're at, because they're going to know where you're at anyways, why not actually, you know, derive a benefit out of it. To share it. 


Josh Braaten  16:07  

They're getting benefit out of it, why can't I?


Justin Grammens  16:10  

That's true, it goes back to like you are the product, right? In a lot of ways. Let's talk about poop tracking for a minute or so what's going on with your startup?


Josh Braaten  16:19  

For the last year, so we've been working on Snappoo, which is the world's first AI powered duck poop analyzer, it works because ultimately, there's so many different things about your dog's health that can be determined through looking at a stool sample. If you think about it, the first time your dog's not feeling well, you call up your vet, you're saying, hey, what do I do with my dog, and they're saying we'll bring in a stool sample, they'll look at the consistency, they'll look at the color, they'll look at a lot of the different potential dangers that could be presented in a dog like zoonotic parasites, which are like worms, Pyro virus, dehydration, like all sorts of different things, visually, they'll be able to take a look at it and say, yep, this is what's wrong with your dog. So we figured we could employ, you know, image classification tactics, and then have a level of synthesis where you kind of figure out, like, what's happening there. And then ultimately, let people know what's going on with their dogs poop, and overtime, coach them into raising a healthier dog through different suggestions and coaching and tips. So we got the idea when we were out walking our dog  one day, we talk actually a lot about IBD on the human side. And so we're always talking about poop and, and then we were walking our dog one day, and it just kind of clicked, we're like, we should totally do an ML dog, or an ML slash AI app for dogs.


Justin Grammens  17:39  

That's great. And so you said you guys are in sort of a private, private beta right now.


Josh Braaten  17:43  

Yeah, so we're finishing up our private beta. We launched that private beta last fall. And we've got about 50 folks or so that have been invited to join and have been exploring the product, and helping us calibrate the technology to make sure that we're kind of sorting out the right results. There's some sensitivity that we're training into right now, because certain things show up more frequently than they should. And so we're kind of fine tuning the sensitivity to our model right now.


Justin Grammens  18:12  

Interesting. Is this publicly available? I mean, is there is there a website stuff like that, that people can go to?


Josh Braaten  18:17  

Oh, yeah, being the marketer, entrepreneur, we have a website, there's a an email list, there's all sorts of different ways that you can engage with us. Snappoo.com is where you can go to learn more about the product, you can request to join the beta. And once you do, we'll be able to send you a  link to be able to download the latest version, it's a, it's in an APK. It's not on the actual Google Play Store or the iTunes store yet, it will be we're shooting to try to get that out in the first quarter or second quarter. But we're building some new features based on feedback that we want to introduce as part of launch.


Justin Grammens  18:48  

That's awesome. Very, very cool. I guess dogs could be one thing, but could you do this with cats, or hamsters, or guinea pigs or other animals?


Josh Braaten  19:00  

You know, we thought about it in theory, you can write because with image classification, you just need a large enough data set, you need to have some features to train it on. And then, you know, figure out a way to create some utility from it. But there are challenges for each type of poop. We actually had the idea for humans first, because, you know, we're a family that is very, very close to ulcerative colitis. And so we're, we're always talking about poop, and we're thinking about doing it for humans, but there's all sorts of barriers to that. There's privacy barriers, there's an just inertia type, habit type barriers, a lot of people do their business a very specific way and taking a picture afterwards might not might not work. And then lastly, there's all sorts of just sensibilities you know, privacy concerns, like taking your camera in the bathroom with you is generally not something that people want to do. Yeah, sure. Sure. So we ruled out humans cats was also another one that we initially thought of, but if you think about cats, they tend to cover their poop with cat litter, and so You end up you know, obfuscating a lot of the different signals that you'd need to utilize. And so for dogs, it's perfect because you're outside with them in a lot of cases, you're taking them on walks, you know, a lot of times you have to pick it up anyway. Yeah. But I was thinking you're usually right there. We did a survey. You know, we we asked 1000 plus people via Twitter, all sorts of questions about their behaviors with dogs and taking care of their dogs and smartphones. And over 90% of dog owners have already taken at least one picture of dog poop in their life for some reason or another. And so this is not something that would come as a huge shock for people, I'm sure. And so we thought, we thought for all those reasons, dogs was definitely the right market to start in.


Justin Grammens  20:46  

Y eah, that's great. Well, we have sort of liner notes for each one of the episodes. So I'll be sure to mention it. And, you know, put links to Snappoo.com in those notes, so people can check them out. You mentioned about going out and walking dog. So, you know, obviously, you you mentioned before, as you know, before we started recording, you're working from home, I mean, what's what's a day in the life of a person in your role?


Josh Braaten  21:07  

Sure, well, being an entrepreneur that runs an agency, and has a lot of different clients. And then also trying to sell fund, you know, an AI app, there's a lot of different hat switching. So we're we're constantly working from different brand measurement projects, different brand growth projects, learning different things about the ad platforms, different survey platforms, and delivering results, all sorts of different things. And and then what whatever time that we have leftover, we are dedicating to getting snap, who further along and so we're getting feedback from customers interpreting how we need to translate that into features, building those features, getting feedback. On the product side, on the tech side, we're learning so much about so many different parts of the actual tech stack of building an app. And then on the marketing side, it's just trying to get people to do stuff that they had didn't think that they need. So for new tech products, it's kind of hard to know, when something's gonna sell right away, like hotcakes, right? It's a term that I like to think about a lot. It's just does the cell like the expression hotcakes. Originally, when the car came out, it sold right away because it was it was amazing. And people knew immediately that they wanted to have it. Production was very cheap. And it sold like hotcakes, essentially, because it was a great, great product market fit. And then some are hit like a dud. Like Google Glass, right. And so trying to tell the story of this new technology and and get people to see what you see as a founder is kind of a huge challenge. But But still, it's a lot of fun with whatever time that we have leftover in a given day.


Justin Grammens  22:47  

Totally agree with you. In my seat. As with regards to multiple hats, right? I am the founder and CEO of Lab651. But also building technology and running a number of different startups doing this podcast and having that by the AI meetup and, and running a nonprofit. There's just a lot of different things. It's it's an exciting time, though, right? I think to be in technology,


Josh Braaten  23:05  

It really is. I think some people will say, Oh, Justin, you should focus, you know, you should. And this is where you'll see a lot of especially VCs and advisors talking to startup founders, they'll say you should really focus on this one thing that you're doing. And yet the prevailing wisdom for VCs is that you need to diversify as much as possible. And so I think as entrepreneurs, the idea of diversifying your professional interests and things that you focus your attention on is actually unconventional, but very positive paths to go down. Because that's how you can become somewhat of a renaissance person, in the sense that you're a painter, you might be an inventor, you might be a philosopher, that kind of thing. Only today, it's you might be an entrepreneur, you might be a startup founder, you might be a podcast host, and you're not having to be defined by one thing. And I don't know, I think it helps me end up having more energy throughout the day, because I don't end up having to do things that I don't like to do so much. I can delegate or partner with people for those things and end up applying my efforts to the things that really appeal to me.


Justin Grammens  24:15  

Couldn't agree more well, as you've been working in in AI ML space, I like to ask people like, So what sort of projects have you seen where there's some interesting things that you know, and doesn't even even have to be things that you have personally done? But as you're like, kind of like looking through the news, things come across your desk? What are some cool things you're seeing going on out there?


Josh Braaten  24:32  

Yeah, that's a great question. You know, it's always interesting, because when you see something, it's not always obvious that it is AI sometimes. Sometimes it's very, very clearly like touted as AI. I think one of the things that I think was most profound to me recently was Google's Alpha Fold technology that they've been working on for a number of years now. But just last last year, they ended up making a huge breakthrough prior to Google Alpha Fold. which is the protein folding, artificial intelligence that is supposed to predict all these infinite, it's like 10, to the 30th power, number of different ways that these proteins can unfold, then, and knowing that is the key to understanding so much about, you know, ourselves and health in the future and that sort of thing. So, prior to this 17% of protein structures known to mankind or personkind, were plotted out all the different 3D structures of these proteins. And now, we've got nearly 99%, predicted with high degrees of accuracy. And this was just a giant orders of magnitude leap on a problem that geneticists have been thinking about for decades. And this was just AI, right, this is, it represents kind of the extreme positive side of how the technology can help our, our species and just us as people, you know, benefit in the world. So I think that's always great. You always hear about the negative sides, too. But this is one of the things I think shines through is like this, this might actually save us versus turning into the Terminator type scenarios that you hear about too.


Justin Grammens  26:12  

Sure. Well, you mentioned the negative side, I guess, what are some issues that you could see with artificial intelligence going into the future?


Josh Braaten  26:20  

I mean, there's all sorts of issues, I think some of them in a range aligns to some of them are very practical, such as in business today, there are all sorts of biases that are introduced into API's that prevent access to certain people. So for example, in recruiting software's certain populations are underrepresented in candidate pools, because AI's are created by those who are enforcing current systems. And so it's difficult to to kind of use AI in that scenario, without understanding how biases are, are impacting, you know, things like candidate flow. So those are like, current challenges. I think, in the future, you have a lot of today's more powerful figures, who I'm imagining have a lot more time access and money invested into these areas, and you hear them saying, well, we need to legislate around this immediately, we need to draw some guardrails on the world in terms of what we want to allow AI to do and what we want to have be off limits. Because obviously, we could do a lot with this type of thing. And I think if you've seen shows like Black Mirror, it's not unthinkable to think of, you know, swarms of micro drones that are programmed to look for people and, and that sort of thing. So I think those are the really scary things that nobody really wants to think about. But yeah, so I tried to think about the positive things like, you know, being able to map all the protein structures, and use those in medical science. 


Justin Grammens  27:49  

You know it was funny, I was just you were talking about sort of people thinking it's AI, but it's not really AI. And there, there have been a lot of startups that I've known of that like, it's actually human behind the scenes, right, they use something like Mechanical Turk, to actually do the work. And then I thought about bad things. And I don't know if you've been following sort of the Elizabeth Holmes and Theranos, and all that type of stuff, where, you know, as an entrepreneur, she promised all these things. And at the end of the day, well, a it didn't work. But then B, if you read the book about it, they were actually just funneling the data off to these other much more expensive machines, laboratories, where people were manually doing all the work, and then sending the information back to the machine. So it was like, tada, look at this machine. It's doing all these things with a drop of blood. And it's a no, there's actually a whole team of scientists behind, you know, the curtain, that you have no idea that actually, you know, so you're for sort of false promising a lot of these things. Is that so wrong? I mean, from my standpoint, if I'm an end user, and it gets the job done, whether it's, you know, using AI or ML, or it doesn't the technology, somebody doesn't matter, as long as it's providing a service in some way, that you know, you will eventually, you know, automate. Is that so bad?


Josh Braaten  28:53  

That's a great question. I think, as marketers, we're faced with that question all the time. I think answering that question can happen in two ways, or it should anyway, number one is technically correct or incorrect? And that's where is like, is the technology actually ml AI? In that situation? Probably not, you know? And then philosophically, is it AI? Does it do the same sort of things that that maybe the technology could accomplish? And I think that's really important to make a distinction, too, because oftentimes, there are a lot of different startups and a lot of different startup founders. That will, you'll have the startups that are trying to do something AI related. And you'll have startup founders that are plagued with imposter syndrome. And they're like, Well, my dozen mind's not AI quite yet, but it's going to be great. I think that's an important distinction. Because if it's on the path or on the product roadmap to actually make the machines do something then then being able to market that vision ahead of time as you know, it's going to be an AI technology. I think that's okay. Aspirational is different than being deceptive. If it's not AI, and there's no intention of being AI and you know, it's not AI and you're just using it to say scan people. That's not okay.


Justin Grammens  30:02  

Exactly. And I guess some of it is around efficiency. Some of the things that I like to think about is, you know, how does this impact the future work of humans. And, you know, some people think about all look at all these jobs are going to be gone. But you know, if you've listened to me in prior podcasts, I'm usually more on the optimistic side where it's usually the jobs that people don't want to do, that I that I sort of see the greatest benefit coming to us in the next 5, 10, 15, 20 years with this new technology. Is that, does that resonate with you? Do you sort of agree with that or have a different view?


Josh Braaten  30:34  

You know, it's always so difficult to understand whether new technology is going to create more, more demand for new types of jobs, or whether it's going to replace jobs in a way that's net negative, some of the conversations that I've been hearing in business circles, is suggesting there might be a short term net negative workforce reduction. So ultimately, maybe losing some jobs in the in the near near term. But it's hard to say if that's because they're going to need to retool at some point in the future for some of those newer jobs, or if they're just going to go to different parts of the economy, or, or something else, right, or if it just gonna be gone forever. So I think it's too early to tell. I do think we are in a sector, or we're in an economy right now, where large parts of our workforce are, are built upon sectors that are ripe for displacement, and automation. So retail service serve at the service sector, you know, there's a lot of those types of jobs that could be subject to automation. And anytime you have those large changes, that can't be quickly offset by gains in innovation, that's where you might have some short term or even medium or long term, I guess, employment declines because of how automation has offset that that demand for labor.


Justin Grammens  31:57  

Andrew Yang has a book out called The War on Normal People. And his philosophy is, is that he thinks that this is actually a different type of revolution. So the industrial revolution happened, right. But then people left the farm and they went to the factories to build these machines, right. And so there's always sort of been this next path, where it's just like, you know, I can move into this next industry where I'm actually doing something. And, you know, he fundamentally thinks that now this intelligence, and whereas the humans that maybe don't have the intelligence, right, no one can become a software engineer, like not everyone's gonna become a software engineer, right, there's a maximum with regards to what people need to do to build out there. And so it can be scary, but I don't know, I always believe that, you know, people got worried when, when all sorts of things, you know, what would change within the economy, a new, you know, tools and advancements would happen, and people are like, looks like I'm out of a job. But there's always something else. I feel like there's always something else around the corner.


Josh Braaten  32:51  

Yeah, I think it's really interesting, I think, you know, figures like Andrew Yang to they're talking about, they're drawing parallels to different points in history, that actually, where there were giant shifts in different types of curves, rather than just fluctuations. And so, for example, during the early 20th century, you had Industrial Revolution, and you had that large kind of change in labor that Yang was talking about. And that led to a lot of social change, too. Because as you had people moving from the countryside, into the cities, they were so little regulated at the time that the industrial barons of the time, could set the terms of how those people came into the cities, how they lived, what type of wages they were able to get, what type of protections were available to them, all those different things. And it wasn't until people felt like the technology, they had far exceeded their their own standard of living, that there were large changes in society actually led to, you know, a couple of different major economic transitions within the 20th century. But ultimately, you've seen major shifts in society because of those technology surges. And the question is, Is this one of those major technology surges? Is this just a new technology like, you know, TV, the Internet where, or you know, telephone and TV, the internet, where now it's just a lot more media channels, and it's create a lot more value and that sort of thing? Or is it moving from the country into the city where there needs to be, you know, some guardrails and guidelines put around technologies like AI, ml, and automation, just because if we don't potentially we could end up with a workforce that or a society where, you know, we don't have enough people or jobs to go around for the people who, who need them, potentially, you know,


Justin Grammens  34:42  

Right, right. Yeah, I guess I guess that is that is the the $10,000 question. Yeah. How big of a shift is this? He has this whole thing around basically automation of self driving trucks, right? And then all of a sudden now all these people, their livelihoods were sort of dependent upon driving these trucks, you know, hundreds and 1000s of miles. But obviously that's going to be gone. And in some ways, you know, I can see the point of it that we're, we're talking about physical world changes, you know, the the change from phone and television to the internet was just a largely digital transformation that was really only stayed in the world of essentially zeros and ones. But now like this technology is reaching out and actually moving things and doing things that a potentially a human could have done in the past. So it feels like it's a little bit more of a monumental shift.


Josh Braaten  35:27  

Yeah, and we're the early days too. I mean, like, 20 years ago, Amazon was a bookstore, and now it's bigger than Walmart. 20 years ago, Walmart was the was the bad guy. And now people are like, well, the deal, just drive to Walmart. So I don't have to give my money to Amazon. I've heard people say that


Justin Grammens  35:42  

Sure all how things change. You know, whenever you're the monopoly, you're, you're, you're you're the bad guy.


Josh Braaten  35:47  

Yeah, I think that's a really interesting thing is that we do have this this potential natural monopoly in, in technology at some point where you've got four or five different brands at the top, they're the ones that have the infrastructure and the platforms. And you know, it's, it's interesting as a marketer, and as somebody who's a technologist and somebody who is an entrepreneur who is building on top of a lot of these platforms, I feel like, hopefully, they continue to be seen as forces have good in our economy so that people can continue to create value from them versus being seen as a negative force in that needs to be reined in, or, you know, done away with.


Justin Grammens  36:25  

Yeah, for sure. Well, let's change gears a little bit, maybe talk about some other things you're doing, I guess, I guess, outside of AI and machine learning. Are there any sort of like books that you're reading right now, or any sort of hobbies, other things that you do?


Josh Braaten  36:40  

Well, we do a lot of cooking. So, you know, if you if you ever see my Instagram page, you'll see a lot of breads, you know, noodle dishes, will make meats, all sorts of different things. But ever since the pandemic started, that was kind of our, our hobby. And it's just something that we like to do quite a bit, we read somewhat, I read a lot of business books, they're very consumable, because everybody wants to write a book. And so they're all given different takes on very similar topics. But I think I think some of the timeless ones really enjoy Good to Great by Jim Collins, really been a big fan of just the strength mentality throughout my professional career, how brands grow is a very amazing book by Byron sharp, it really talks a lot about the importance of brand in marketing. And it talks about how brand metrics like awareness, consideration, preference, how those actually do impact sales and how brands should be thinking about them in as a part of their marketing mix. Reading from other parts of the world, we've been trying to just do our part and learn as much as we can in today's society. So a good book is, So You Want to Talk About Race by Ijeoma Oluo. And, you know, if you want to talk about race with friends or family in a way that's respectful or doesn't end in shouting matches over dinner tables these days, it's just, I don't know, it's a good book in something that we found to be very helpful.


Justin Grammens  38:10  

That's awesome. Yeah, for sure. When it comes to cooking, you know, my wife and I just started this, I probably was around the pandemic time is just ordering from one of those one of these companies that brings you the food and then you have to cook it. So we use a thing called HelloFresh. I mean, the food tastes great, actually, I'm really, really happy with the quality of it. But a lot of it is just the time together, right. And we have two young kids, and so we're always running around and trying to feed them and everything like that. But you know, at least two nights a week, we have a meal that we make together for 45 minutes, and we have a lot of fun with it. So that's awesome. I mean, it's a really, really fun sort of way to relax, I think with your significant others.


Josh Braaten  38:47  

Yeah, I couldn't agree more. And I think there have different camps, right? There's the it's kind of like, do you shop at stores? Or are you a Stitch Fix person? Right? The HelloFresh deal is great, because if you're wanting to cook but you're not 100% Sure, you don't have your 50 classic recipe standbys, where your pantry full of things that you know, of your staples, then, then they're just going to bring you all of the things that you need, you've got your instructions, you've got your fresh ingredients, and you can do that, you know, no shopping needed. I like that, too. It's a good concept,


Justin Grammens  39:21  

for sure. Well, you know, we're gonna wind down here in the next couple minutes. But one of the things I like to do is ask people, you know, as you think about back on your career, because you've been doing a lot of a lot of different stuff here, Josh, you know, are there any any sort of advice that you would give people maybe that are getting started, specifically some classes that they might take or meetup groups or anything? I mean, how have you seen success that you could maybe pass on to people just getting started?


Josh Braaten  39:43  

I think people overestimate how long it takes to learn something. And so they're always looking for the one class or they're always looking for the one conference. So the one person that they have to meet. I think the two or three most helpful things that you can do are just number one, trying to find some Good, high level education on whatever topic that you're looking into. So for example, with AI, just simple things like learn with Google's AI, they've got some great kind of basic videos and education topics that tell you the basic principles. And then they're also really good, full on courses like Coursera, is course by Andrew Ng. He's, you know, the founder of Coursera. And was Google brains lead at one point, really famous and helpful person just teaching the subject directly for free, in Coursera. So it's not that the education isn't out there, it's just a matter of people, I guess, being intimidated by where to start, and how long it's actually going to take to get there. So I think, you know, finding simple free courses like that to get started. And then using some of the tools out there that are available to everybody, Google Cloud platforms, AI tools, we use vision ml for a lot of the snap who project but there's AI tables, speech to text recommendations, natural language processing, they've got all sorts of different tools out there. And I feel like it's not as prohibitive as people might think if they've got some data, and they've got a business case, or some sort of use case for AI, a little bit education, and some, you know, rolling up your sleeves and just digging into some of the tools. That's the best way to learn in my opinion.


Justin Grammens  41:21  

Yeah, for sure. For sure. I'll be sure and put links to those classes that you mentioned here in our liner notes, along with the books that stuff that you mentioned, as well, you know, I was just thinking back to so you want to learn about art. So you want to talk about race. I've heard that name, there's another one out there called cast by Isabel Wilkerson that I was just recommended. And I run a fair amount, I still listen to a lot of audiobooks. And that's another really, really good one about race. If you haven't checked that one out. It's all about caste systems, and how we actually are in a caste system in the United States through this.


Josh Braaten  41:52  

It's really interesting, I put that on my list for sure. Because I think if you were to suggest to some people that we are in a caste system, I think people might think that you're being just hyperbolic, or you might be overstating something. But if you do expose yourself to some of these books, and other, I guess resources that will help highlight those true parallels. It's not actually that surprising.


Justin Grammens  42:14  

Right, right. I guess when it comes to time. I mean, the audio book is probably 18 hours long to listen to some of these things. And I'm probably like, a couple hours into it. But yeah, she's really good. It's read by the author, and I kind of came into it with some of the same feelings that that you thought, you know, as well. So like, Well, really, but even within the first couple chapters, she lays it out really, really well. So highly recommend it. I will drop a link to that. In what speed


Josh Braaten  42:38  

Do you listen to your books at Justin? Yeah, 1.5 1.5. Yeah, then I could do 1.75 something. I can't do the 2.0. That's just we keep that


Justin Grammens  42:46  

No, no, no, no, for sure. I it has to process in my brain. And like I say, I'm usually doing something while I'm listening to it. Right. So I have to keep my feet running and moving. Most of the time. And by the way, I'm also listening to this book as well. So yeah, I kind of start to miss things. I think if I'm listening at two, we're sure


Josh Braaten  43:04  

I think when you read audio books, you can say that you read the book about 80%. Right? Something like that. So yeah, percent of five books. That's four bucks. Good job. Yeah.


Justin Grammens  43:15  

You know, I, for me, it's also the material just to kind of offer a little bit of a tangent, like, there's another audio book that I'm listening to right now that I'm you know, I'm probably about 75% of the way done. It's called 1000 Brains. And really, really fascinating book, I forget the name of the author, but it's the guy who founded Paul, basically, it's very much about neural networks. And it's basically about how chemistry happens in your brain to essentially, kind of do these predictions that everyone thinks about, you know, the classical deep learning neural net, he has a lot of data to show that's actually he's spot on, it's actually spot on what's what's what's happening, what's going on in your brain. And as you're reaching out to pick something up, for example, there's a lot of predictions that are going on with regards to am I there yet? Is my hand there yet. Okay, I've touched it, you know, and there's a lot of reinforcement learning that's going on. But to get back to what I was saying is, it's a deep subject, right? And I find myself as I'm listening to is like, gosh, it would be so much I would absorb it a lot more if I was actually reading a book and like taking notes, there's something about coming in through your mind and then having heavy regurgitate it through writing. And I would definitely pick it up a lot more. So some certain books that are more storytelling. I feel like I actually pick up a lot better than ones that are like hardcore, deep science stuff that I really want to you know, absorb and then have to regurgitate a lot.


Josh Braaten  44:25  

Yeah, I wouldn't be surprised if there's some truth to that, for sure.


Justin Grammens  44:28  

So me so yeah. How do people reach out and connect with you?


Josh Braaten  44:31  

Yeah, absolutely. You can always find me at brand data.com or snap who.com. You can see me on Twitter @JLBraaten or drop me an email. I think our emails available on our website. So that's a good way too


Justin Grammens  44:46  

Sounds good. Josh. Was there any other thing you wanted to mention? I guess on the way out for we we close this out?


Josh Braaten  44:51  

I think the only thing I want to say is thanks so much for having me on the podcast. It's been a real treat.


Justin Grammens  44:56  

Awesome. Well, great, Josh. I appreciate your time and we look forward to keeping in touch for you as you continue on your career and keep trying new things out new technologies.


Josh Braaten  45:04  

Thanks so much. Take care.


AI Announcer  45:07  

You've listened to another episode of the Conversations on Applied AI podcast. We hope you're eager to learn more about applying artificial intelligence and deep learning within your organization. You can visit us at applied ai.mn To keep up to date on our events and connect with our amazing community. Please don't hesitate to reach out to Justin at applied ai.mn If you are interested in participating in a future episode. Thank you for listening



People on this episode