Jeff Jonas: [00:00:01] For everybody that's had a close call in life, where you're like - Oh, man! That was close one. Every day since then has been an extra day. When you think about life like that, it allows you to just unleash a little bit more and make the most of it.

Harpreet Sahota: [00:00:31] What's up, everyone? Welcome to another episode of the Artists of Data Science. Be sure to follow the show on Instagram @theartistsofdatascience and on Twitter @ArtistsofData. I'll be sharing awesome tips and wisdom on data science as well as clips from the show. Join the Free Open Mastermind Slack Channel by going to Bitly.com/artistofdatascience. We'll keep you updated on biweekly open office hours. I'll be hosting for the community. I'm your host, Harpreet Sahota. Let's ride this beat out into another awesome episode. 

Harpreet Sahota: [00:01:20] Our guest today is an acclaimed data scientist who for over three decades has been at the forefront of solving complex big data problems for companies and governments. He's focused on creating technologies that help solve the world's biggest data challenges while also being an advocate for privacy and civil liberties, tackling many high profile challenges, including identifying potential terrorists, detecting fraudulent behaviors and casinos, connecting loved ones after natural disasters and modernizing voter registration systems. He's a serial entrepreneur and sold one of his companies systems research and development to IBM in 2005. That same company, which famously invented a program called Nora, an acronym for non obvious relationship awareness, which mines data sources to determine relationships between people. He's a former IBM fellow. The highest honor a scientist to IBM can achieve typically bestowed upon less than 10 people a year, a title that puts him in the company of five Nobel Prize winners. In 2016, he founded Senzing based on a one of a kind IBM spin out of the G2 technology and is the first to deliver real time artificial intelligence for entity resolution software that enables organizations of all sizes to gain highly accurate and valuable insights about who is who and who is related to whom. In data, he's a leading creator of entity resolution systems. And today, numerous organizations rely on his systems to extract useful intelligence from tsunamis armies of data.

Harpreet Sahota: [00:02:36] His work has been featured in documentaries airing on networks such as the Discovery Channel, and it's been the subject of prominent chapters in books such as No Place to Hide Safe. The Race to Protect Ourselves in a Newly Dangerous World. The Numerati and the Watchers. The Rise of America's surveillance state is a highly sought after speaker who meets with government leaders, industry executives and think takes around the globe about innovation, national security and privacy. His software has helped Las Vegas casinos identify fraud, increased voter registration, protected single ports, waterways from piracy, and even predicted possible collisions between six hundred thousand asteroids over 25 years to help save the Earth from Armageddon. Thank you. He's also one of only three people in the world who have completed every Ironman triathlon currently on the global circuit. This is especially significant given that he was briefly a quadriplegic in 1988, followed the car accident. So please help me in welcoming our guest today. The man. National Geographic named the Wizard of Big Data, the mastermind madmen and Miracle himself. Jeff Jonas. Jeff, thank you so, so much for taking time out of your schedule today to be here on the show. I really appreciate you being here.

Jeff Jonas: [00:03:46] Thanks, man. Hey, with an opening like that, I'm not sure there's really anything left. I think you just kind of summarize my entire life.

Harpreet Sahota: [00:03:53] Hey man, well it's such an incredible story, man. Like your story and everything you've been through such. such an inspiration for me. I was wondering if you could quickly just talk about your professional journey, how you first heard of data science and machine learning. And what drew you to the field?

Jeff Jonas: [00:04:08] Well, I was like I'd say, 14 years old and my mom said, do you want to go see something called a computer? And I never even heard the word. So I kind of went along with her and she was a lawyer. She was looking at a very early computer called a TRS 80 to do billing for to automate her billing instead of using a typewriter. And they connected this computer over the phone line with those things called modems in the old days and started doing searches for things. And my mom brought another friend with her who is a self-proclaimed inventor. And he said, yeah, I'm inventing. I'm working on copper pipes and moving air through them to create refrigeration without electricity. And he goes, search for that. They came back and said, there's one hundred and twelve other people researching that. And you could just see his facial expression. And right then and there I went. Oh, that's what I do. This is what I do is this is data and finding data. And so it became a purpose. And so I just became obsessed with computers from that age. In high school, I wrote a word processor to do I mean, excuse me in high school, like first thing I did was create a little program to play hangman. But the second thing I did is create, a word processor for another computer. No one's heard of call it that, Commodore. They didn't really exist back then. You couldn't really buy word processors for things like that. My school teacher at high school said, Do you mind if I sell it? It works pretty well. And he went out, sold a couple of copies, including one to the Los Angeles school district. So I got a check. I was like 16 years old and somebody sent me money for writing software. And I thought, this is crazy. And I was hooked. It's just been an obsession my whole life.

Harpreet Sahota: [00:05:48] That's awesome, man. Those are like some really awesome early gigs that you had there in data science.

Harpreet Sahota: [00:05:53] So me coming from where you started from really the early days of computing, where do you see the field of artificial intelligence data science machine learning headed in the next two to five years?

Jeff Jonas: [00:06:07] You know, it's been kind of flat lining, I think, for a while now. Like, you know, it's it's showing a lot of utility against pictures and and and like multimedia sound data, but it's it's it's not really continuing to have the same gains and other kinds of prediction areas, and it tends to flatten out early. You get all these early gains. But then to get to the last mile, it's not... It's been a little tougher. And an example of that is cars. I think it will still be five, more than five, years before we see autonomous cars on the road. Maybe with the exception of long haul, long haul trucking. But some of the challenges that come with this is security and and privacy. I mean, we're especially now with COVID, you know, the amount of instrumentation that the world will see, especially for contact tracing, is going to put things in place that we may have to wonder, did we put the right things in place? I mean, we'll put them in place to save lives, but then will they be repurposed in use later? So that's on the privacy side. On the security side. Some of your listeners may be aware of this and some maybe not. But the number of attacks that are happening against public institutions and private sector with ransomware and phishing attacks is really up. So desperate people do desperate things and other nation states are going to enjoy seeing us in America here down, down in the dumps. So they're trying to make it a little bit harder. So between those two things, I think we have our hands will be full on helping secure our systems. You know, you can't worry about that because you can't take something that's brittle and then or it's a bad idea to take some of this bread alone and stack other brittle things on top of it. So imagine kind of an economy that's teetering a bit and then, you know, my God, if you had a ransomware attack against one of the financial institutions where the ransom was unpayable, like that would just be fragile and fragile and very bad.

Harpreet Sahota: [00:07:55] So in this vision of the future with all these issues that we'll be facing that you just described, what you think is going to separate the great data scientists from the good ones.

Jeff Jonas: [00:08:08] I think that a lot of the big gains to come are not about pointing algorithms at a data set, but converging multiple data sets and getting orthogonal data like secondary data from secondary sources thinking about it like a puzzle. You've got read puzzle pieces, blue, yellow, you know, green, white, black, brown, all these color puzzle pieces. A lot of times what I'm seeing is people taking algorithms and applying them to a set of puzzle pieces like the Red Puzzle piece. We're looking for bad guys. We look at the red puzzle pieces for bad guys. But it really if you weave those diverse data sources together, puzzle pieces into pictures, and I think we're going to see more effective outcomes in our machine learning and A.I., by the way, I don't use those words interchangeably when I say I mean machines that act smart. And when I say machine learning, I mean systems that learn to experience many A.I. systems use machine learning, but not all.

Harpreet Sahota: [00:09:00] Interesting distinction. I like that. Yeah. Very subtle distinction as well. I like that. The point you made about the puzzle pieces, because oftentimes I kind of describe what I'm doing to colleagues when I'm taking these disparate data sources and combining them together, like almost like working with the Rubik's Cube. Right. We've got a cheap is all mixed up and it's kind of our responsibility before we even apply any type of model to it to massage it into the right color schemes, I guess.

Harpreet Sahota: [00:09:29] What's up, artists? Be sure to join the free, open mastermind slack community by going to bitly.com/artistsofdatascience

Harpreet Sahota: [00:09:37] It's a great environment for us to talk all things data science, to learn together, to grow together. And I'll also keep you updated on the open biweekly office RSV hosting for our community. Check out the show on Instagram @theartistsofdatascience. Follow us on Twitter @ArtistsOfData. Look forward to seeing you all there.

Harpreet Sahota: [00:09:59] Get into to some of your history because it's it's quite inspiring some of the things you've been through. Right. So there's there's a time very early in your career when you went bankrupt. You're living out your car. And instead of going and getting yourself a steady job, you just went and decided to go for it again. You know, talk to me about what you're saying to yourself to get you through that. And what did you learn from that? To go on to create something bigger and better than what you had before?

Jeff Jonas: [00:10:27] I dropped out of high school at 11th grade because there were no more computer classes, because that's all I wanted to do. I went to one junior high, excuse me, junior college class for one semester and did a couple of programming class and maybe did a data analysis class. And then I just started my first software company and I ended up hiring a bunch of people. I had like 21 people working for me by the time I was 20 years old, but I'd made promises we couldn't keep. I didn't know how hard it was to build software. Like I agreed to build entire accounting systems, general ledger, accounts payable, accounts receivable for fifteen hundred dollars. Like, that's just insane. I didn't know anything. So that company bankrupted. And so I was sitting there in my office. All my furniture is repossessed. I had all these employees with bounced checks that couldn't, you know, Pay rent to feed their family. I feel pretty much like a schmo, cried some, but, you know, I really wanted to just build software so you could call it my first kind of big life lesson. It was just a full up face plant. And I realized that the reason why we were unsuccessful in building software is we were we didn't have blueprints.

Jeff Jonas: [00:11:29] I didn't really know exactly what they wanted to build before I built it. It's kind of like, why would you try to build a house - you weren't even try to build a small dog house - without a blueprint? You'll end up with the floor crooked and the roof half, you know, messed up. But what if you tried to build a house without blueprints? It'd be a mess you couldn't get your bathtub in because the door was too small or something. So I really focused on doing blueprints first. I went around and now you're a bankrupt 20 year old, couldn't even drink alcohol. And I was showing up trying to get business. And how do you get business if you're 20 years old living in your car? I made outlandish claims. I said, I'm going to do a blueprint. I'm not going to charge you for that. I'm done with the blueprint. I'll tell you how much money I'll charge you - it's six hundred a day. And at the time, that's crazy price, six hundred a day to be a 20 year old homeless person. And I said, but don't pay me till I'm done. And every day I'm late, take one hundred off. Man, you get a lot of people sign up for that. So I do a blueprint.

Jeff Jonas: [00:12:18] I know exactly what they want. One of the companies did some work for it was a philanthropic nonprofit called on the CS Fund, at the time, located in Northern California. They would receive I think it was Ford Foundation money and they would receive grants and they wanted a grant tracking system. And so I designed probably a 50 page blueprint of what that grant tracking would look like so they could promote the best grants and give the most suitable people access to their philanthropic funds. And it would track the status of those grants. But anyway, I gave the blueprint. I said six hundred a day. It said, we've been working on this for a year and a half. We're not paying you 600. Is it all be done in five days? Not on a five day- I'll, I'll -  you can take one hundred dollars a day off the price. They're like, huh? Three grand. OK, do it. And I finished in four. I had to work night and day. Crazy. That's what I did. And I just I did that repeatedly and I really dedicated my work in back when I was programming to Blueprint's First. And that became a very successful way to build stuff.

Harpreet Sahota: [00:13:15] It's something that, you know, I often hear very pretty commonly said about machine learning and AI projects and initiatives is that eighty five percent of them fail.

Harpreet Sahota: [00:13:25] And I think the reason 85 percent of them fail is because nobody has a solid blueprint to follow. So fast forward just a couple of years later. Now that you've re-established your software company, you're 23 years old, waking up in a hospital, completely paralyzed after terrible accident, you know, suffering the same injury that left Christopher Reeve paralyzed. I think most people at that point would have given up and had kind of the woe is me attitude. But instead you say to yourself, I think I could still use my nose and attach a pencil to it, and program. I mean, life literally knocked you down. You got right back up, in a literal sense. So talk to me about your mindset or the self talk you had during that time. What was going on in your head? And then how did you overcome that challenge?

Jeff Jonas: [00:14:11] Well, first, you just put a little color on that incident as I was test driving a BMW and the salesman was driving and they were showing off the car from the dealership. They said they were trained on the track, and they lost control on a road.

Jeff Jonas: [00:14:25] I think it was 63 miles per hour. And we hit a dirt embankment and it snapped my neck at C2. And so, yeah, I was a full quadriplegic. I'm laying in a hospital, but I never asked if I would walk again. I kind of just believed that would like I didn't even ask. I just went, I'm going to be fine. I had know basis to know that or think that other than just positive thinking. And I just to me, software that's just been kind of like my hobby. And I imagine I said, well, I can at least do a program like I'm attaching pencil to my nose. I might be slower, but I could still program. Yay!

Jeff Jonas: [00:14:59] Eight days later, a toe wiggled and then I worked my way through a walker and a wheelchair walker, make it cane for a few hours. And I walked out of the hospital 18 days later, my left leg drag. At the time, I would've told you I had a bunch of lessons. I wore something called a halo vest, by the way. Some people have seen these.

Jeff Jonas: [00:15:16] I had metal bar that was around my head, a ring, and it had bar because it would come down and attached to a chest plate and a back and would hold your head in suspension like a cast without all of a white cast material. Had to wear that for four months. And I had a lot of lessons from that. Like if you would have asked me back then I'd be like Oh, I learned this and that. But all of those faded away. There's only one lesson that remains from that time of my life, and that is every day is an extra day for everybody that's had a close call in life. Oh, man, that was close one. Every day since then has been an extra day. When you think about life like that, it allows you to just unleash a little bit more and make the most of it and freak out a little bit less about death because who wants death.

Harpreet Sahota: [00:16:03] Oh, that's really powerful. I think I would adopt that that that new motto every day is an extra day.

Harpreet Sahota: [00:16:10] It reminds me of I just recently, a couple days ago, was watching Impact Theory. And there's a guy on there called Hal Elrod, I'm not sure if you're familiar with him, but he he suffered a horrific accident as well. And he  formulated what he called the miracle equation. And it's really simple, like to get yourself through anything. All you need is unwavering faith and extraordinary effort. So yeah to draw some parallels from that. And just out of curiosity, back then, what was the programming language that you were working with?

Jeff Jonas: [00:16:45] I was using something called dBase Three and Four and Clipper and FoxPro is a class. There's a language class. All those are different products that had a basically the same language. It was called a fourth generation database language.

Harpreet Sahota: [00:17:01] So you've been involved in some pretty amazing entrepreneurial initiatives. You've been able to kind of pick yourself up and push it to push yourself through some devastating setbacks. Do you have any advice or tips for anyone who's been toying with the idea of entrepreneurship that maybe has been kind of lulled into inaction by the comfort of a monthly salary?

Jeff Jonas: [00:17:24] Well, it's funny now, the COVID era, you know, there's going to I think 20 per cent of my friends are going to be unemployed and like the next six weeks or at least three months, including a lot of data scientists. The question is, what are these people going to do and how are we gonna get how are we to reboot the economy? I think I think in terms of new blooming data scientists that are really early in their career, I think getting your hands on data and touching it and experimenting with it and bouncing one data set up against another is the thing to do. It's just gives you real world experience. So I'm right now trying to compile a list of open data sources that I think are useful for that. My field is entity resolution, which is understanding when two people are the same, even though they were described differently. One record says Elizabeth the other one says Liz, one's got the maiden name one's got the married name, the month and days are transposed because a lot of times in a lot of data, months and days are transposed and dates of birth are they  really the same people or not? This plagues so many companies in so many areas from marketing and fraud and customer service.

Jeff Jonas: [00:18:35] So that's kind of become my obsession area. And for people interested in that for free, they can just download my software and without even giving us an email address and they can run it on their own address book or their company, Salesforce dot com, and find all the duplicates in a heartbeat. But that's if you if you if you can't figure out who is the same as who, whether it's a person or a company downstream analytics downstream. Machine learning doesn't learn so well. If you think it's five different customers with five transactions, then it's really one with five. And that's just another area where I think data scientists there. They've had more work to do to get the data in order. It's kind of like your Rubik's Cube example. It's like that's about like Dataprep. It's like getting the data in order so that you can make good analysis. So anyway, that's been one area where I think there's gonna be a lot of opportunity, Is it entity resolution. Up till now, it's just been really hard and only available to PhD mathematicians and companies will have 10, 20, 50, I know some companies that have 100 people working on it. And so it's been it's been tough. Anyway, I'm not sure I answered your question, did I answer your question yet.

Harpreet Sahota: [00:19:42] Partly second, second, half of the question was just Justino tips for anyone who's been toying with that idea of entrepreneurship because, you know, I think it was Nassim Taleb that said to the most dangerous things in life are heroin and a monthly salary. So tips for people who have kind of been lulled into inaction by that monthly salary, who have been toying with that idea for entrepreneurship but haven't followed through.

Jeff Jonas: [00:20:19] Yeah, well, you know, it's one thing to start a company when you're 20 living your car, you got nothing to lose. If you had a family and you don't have like a bank account, it's like bursting at the seams where you're going to need some money tomorrow. It's a lot harder. And right now, in this COVID climate, post-COVID era, there's really going to be only - there's not that many ways to go and start a business that I can think of. One would be anything that you can do and work from home productivity is probably a good business. So if you're entrepreneurial in that area, there's probably a lot of opportunity there. Every cost, almost every company is having a lot of top line revenue pressure like their gross revenues are dropping. So when that happens, you have to take costs out. So if you're an entrepreneur and want to be successful in the next year or two year, if you're not building a worker productivity tools you better be able to do fast cost takeout. Like a return on investment. In under nine months, if you're if you're creating something and it doesn't give somebody a really fast return on investment. I just it's going to be very, very hard to sell that because companies right now are scrambling to reduce their costs.

Jeff Jonas: [00:21:26] And then I think fraud, waste and abuse are really gonna hit an all time high because desperate people do desperate things. After Katrina happened, the government pumped like 50 billion, 50 something billion dollars in to help them help the people in the company and the companies. At that same time, about six billion dollars was estimated to be stolen because they just opened up the pipes and pumped 50 billion through it and six billion disappeared. You know, it was stolen. So imagine it pumping two trillion out through the pipes on the same statistic. I think it's two hundred and forty billion of fraud and opportunities are going to be going after banks and going after health care and insurance. So any type of fraud, waste or abuse detection product, which is a great place for data scientists and machine learning, is to find patterns in the data to quantify fraud and stop it. I think that's going to be that's going to come in the second wave. You know, first wave will be cost take out and second wave will be, wow, everybody's trying to rip us off. So those are the areas I would be focusing on if I were trying to be an entrepreneur.

Harpreet Sahota: [00:22:33] Interesting point you made about them. You know, these these data scientists now that are unfortunately, you know, becoming unemployed due to the COVID crisis. And a lot of people who are breaking into the field now, they tend to focus primarily on hard technical skills and they think that that's what's going to separate them from the rest of the crowd.

Harpreet Sahota: [00:22:50] But now, you know, we have we have this great equalizer of the Internet. We can look everything up that we need to, right. So what would you say are going to be some of the soft skills that candidates are missing that are really going to separate them from their competition?

Jeff Jonas: [00:23:04] Do not underestimate curiosity. It is such a great soft skill. You know, just to be funny. Have you heard you can't sneeze with your eyes open. Have you heard that?

Harpreet Sahota: [00:23:13] I've heard of that phrase. Yeah.

Jeff Jonas: [00:23:16] Have you ever tried? Have you ever tried to hold your eyes open. Okay. I'm curious. I gotta try. I'm telling you, man, I don't recommend anybody do this because it was a very, very painful thing. I held my eyes open with my fingers and I went to sneeze. And you if you if you have a small, petite sneeze, I think you can do it. But if you get a big, gnarly, manly sneeze, it turns out to be exceptionally painful. Like I literally almost forked up my throat. It was really just everything except to sneeze. Anyway, creativity can get you hurt. But, um, but in data science, being creative, knowing where data it, knowing where, where the data is and knowing how it's structured, knowing how it flows and knowing how to combine it, knowing how to find the red, green and blue yellow puzzle pieces and how to stitch them together to go from piles of puzzle pieces to pictures I think is are some softer skills

Harpreet Sahota: [00:24:08] I agree that that creativity and curiosity are definitely two things that you can't learn in any textbook.

Harpreet Sahota: [00:24:16] And you can cultivated just by working on a bunch of different problems. Right. But it's it's something you have to do that you have to cultivate it. Right. So that's awesome. So if we can get in to the the Iron Man. Right. So. So you've crushed a ton of these things.

Harpreet Sahota: [00:24:32] So what compelled you to come to complete every Iron Man on the planet? And can you share some of the many, many accomplishments that you've had in that space?

Jeff Jonas: [00:24:42] Wow. Well, Really accomplished. But I'm more of a hacker. I was 31. I'd never done any outdoor sports ever. My mom says to me, Do you want to run a marathon? And I say, Sure. It was five weeks away. So I had to train from no running effort. I'd never run a race. Not a 5K ever. And I had five weeks, so I trained as much as I could. At 20 miles. I had to ask my mom to walk. And then every mile after that She's like are you ready to run Jeffrey? Are you ready? I'm like, No, Mom, not yet. I mean, she could carry me over the finish line. So that kind of got me started. And then and then I started doing some mountain biking and then did a triathlon and did a bigger one, and a bigger one, and a longer one. And then I started doing these Iron Mans and then my buddy we did my buddy and I went did like the Ironman in France and then the next year he said do you want to do one in France again and I go, no, let's do Switzerland. He goes why? I go, I'd rather do them all once instead of the same one over and over. I didn't really mean it, but it turns out that's what happened. And I said it for years. I go, Yeah, I'm just going to slowly work my way through. But I never thought it was possible. There's millions of people to do these triathlons and to do every single Ironman turns out to be really, really hard to do because they delete them and they add them

Jeff Jonas: [00:25:56] And so it's like a whack a mole thing. And some years I had to do seven in one year. And then Once you get in that club where you were one of three people, there's five now, but when you're one of three people on the in like 2014, I think it was 2013 somewhere around there. There was nobody that had done them all. And on that day in Copenhagen, three of us came across the finish line. And so we enacted something called the club. And yet it's it's hard to get it. Then every year you have to watch the Iron Man organization, by the way, you know, if you have a listener that doesn't know it's a 2.4 Mile swim, which is 3.8 Kilometers and it's 112 mile bike ride, which is 180 kilometers. And then it's a full marathon. You just do them all in a row. So it's a long day, but now we're in this club. And so every year Ironman announces new races. And it's a bit nerve racking because, A, you still have to get in and sign up before other people and they fill up. And then sometimes they they're not trying to optimize them. So the three of us in the club can do them. One year, they put four races on four continents in two weeks.

Jeff Jonas: [00:27:01] And that included on one weekend there was an Iron Man in Mallorca, Spain on the island on Saturday. And then they had an Iron Man, a brand new Ironman that we had to go do. That was in Louisville, Kentucky, the next day. No one has ever even attempted to Iron Mans on two continents in two days. And we thought it was impossible. And then and we figured it out. It's very it was a big, big logistics challenge to finish one race, then get on a plane and get all the way across the continent and then get to another race. We got to the second race and to the starting line 30 minutes before the gun went off. I mean, like, literally just dashed across the globe. And then 30 minutes later, bang and another Iron Man back to back. That was that was a long weekend and took a lot of perseverance. I walked almost all of the marathon on the second one because I didn't know I wanted to finish for sure. And I did not know how close I was to actually falling over and being in the hospital. And so to just be totally sure, I got across the finish line. I just I walked probably 40 of the 42 kilometers just to be sure I wasn't going to like have I don't know, have a heart issue. But that was that was maybe the most Herculean thing I've done is pretty awesome.

Harpreet Sahota: [00:28:25] That's pretty awesome. So, I mean, so I'm getting from you. It's not even the actual physical activities are difficult. It's just the logistics that are challenging.

Jeff Jonas: [00:28:35] These were really hard.

Harpreet Sahota: [00:28:38]  So tell us about that. Taste like mango story

Jeff Jonas: [00:28:42] I was in South Africa doing The Iron Man down in Port Elizabeth. It's around the corner there from the capital. I finished the swim. I finished the bike. And now I'm on the run. And as I'm running, I'm catching up somebody that's walking. And I notice that they've got a bunch of brown diarrhea in the back of their pants, you know, right where you would expect that to be. Clearly, it's diarrhea. And I come next to this guy and I just look at him. I go, look, I don't know you. You don't know me, but this is the kind of friend I am. You know, you've had some diarrhea in your pants. He looks at me with his New Zealand accent and says, You mean it looks like I shat my pants? And it kind of. It's hard to imagine that you wouldn't know. But it was there's an there's an edge case where maybe your sharted and you didn't know. I don't know, man. So I said, yeah. And in fact and he looks at me perplexed and he scrapes his bum with his right hand. And now he's got the oozy brown stuff in his hand. And it I can't tell you when I'm processing because my brain's not you're already kind of exhausted. You're not thinking clear, but he looks confused too. He's startled that it actually looks like it's diarrhea. He looks at it looks at me perplexed. Looks at it. Sniffs It looks at me. Looks at it. Perplexed and I kid you not sticks his tongue on it. And by the way, this is related to my work.

Jeff Jonas: [00:30:02] Like, you might not seem like it, but this is act actually related my work. So he sticks his tongue on it and looks at me and I'm out of my mind now. And he just goes, it tastes like mango. Excuse me. Tastes like mango. He had a gel replacement pack, you know, a sports pack chomp down and it ruptured. So it wasn't diarrhea its  just his sports gel pack kind of popped right in that spot. And, you know, I don't know, he put some water on, He goes is it gone. He puts more water on and rubbed his hand on it, is it gone". And finally I say, yeah that's gone. Shake my hand and thank me and I'm almost done. You almost don't want to shake his hand. So sure it was. You know, it's kind of get it. But but here's the thing is I wouldn't bet you a million dollars that was diarrhea. When new evidence emerged here, I realized it wasn't. And that's something that we've spent millions doing in our own software, as you have to let new observations reverse. Earlier assertions. This is something that's missing in most machine learning. The only way to introduce new knowledge is you have to rerun all the models and it's a big batch process. But real time learning. We're right in the moment. You say, oh, now that I know that. Had I known that the beginning. And you have to redo your decisions. I decided it was diarrhea and undecided when I was given more evidence.

Harpreet Sahota: [00:31:25] So a new feature in the software will be a diarreah detection module. There's a lot of people out there who were trying to to break into data science. And maybe they don't feel like they feel like they don't belong or they don't know enough. They aren't smart enough or whatever. Do you have any words of encouragement for them?

Jeff Jonas: [00:31:44] I would tell them to download some data and start. If you're there, if they're bored and have free time on their hands, I'd find out. Maybe a charity that they like and ask them if they need any help analyzing some of their data and just start. I would literally just work on stuff, even if it's free, just to get your chops up, just to get your hands on real, real problems and real data.

Harpreet Sahota: [00:32:04] Awesome advice, yeah. So last kind of formal question here before a jump into Lightning Round.

Harpreet Sahota: [00:32:17] Are you an aspiring data scientist struggling to break into the field? Well, then check out DSDJ.co/artists To reserve your spot for a free informational webinar on how you can break into the field. That's going to be filled with amazing tips that are specifically designed to help you land your first job. Check it out. DSDJ.co/artists

Harpreet Sahota: [00:32:41] What's the one thing you want people to learn from your story?

Jeff Jonas: [00:32:44] You know, I got this quote from this lady that I've done a lot of Iron Man's with. She said it to me on a race course when we were both suffering. There's times where you just really want to quit. And she said to me, if you quit, there's no chance a miracle will happen. And that really stuck with me.

Harpreet Sahota: [00:33:01] That's powerful. I like that.

Harpreet Sahota: [00:33:03] Yeah, you definitely embody that. You know, through through everything you've been through and everything you've been able to accomplish.

Harpreet Sahota: [00:33:09] So that's that's powerful. Jumping in to the lightning round here. So what's the number one book, fiction or nonfiction that you would recommend for our audience to read and who are most impactful take away from that?

Jeff Jonas: [00:33:21] I really I've only read maybe 10 books in my adult life. Like, I really don't read, and I don't watch TV. Like, I actually personally learned through talking to people and doing real projects and real data. But that said, though, one book I have read in the last 10 years was a book that Lee Kuan Yew, the founder of Singapore, wrote called One Man's View of the World. Now, I can't compare it to any other books that I read of the books. That was it really struck me on what it means to have smart leadership. And I think a lot of places in the world, the way that the democratic process can work. Lee Kuan Yew made a point about how when you get two parties that bicker so much, it's hard to get really smart people because who wants to get that much mud thrown in their face. So if it reduces your - I'm not, this is not a political statement on the right or the left, by the way, this is different to that - It's just you aren't really smart policy in place. So that was my takeaway from that book. And then because I my Senzing company was actually not much of a start. It's more of a reincarnation because I spun it out of IBM. But I read many of the chapters in a book called Zero to One by Peter Thiel. And there's a bunch of chapters in there that I was just like, oh, that's very interesting. And one of them was, if you don't have something that's like 10 times better and high margins, then you can't innovate and you end up having to just kind of struggle slowly to the bottom. Competing with others so you don't have something is truly break away. Then you can't continue innovating. And Apple was an example of that. But we strive for that anyway. There's the answer to that.

Harpreet Sahota: [00:34:50] So if you could somehow get a magical telephone that allowed you to contact 18 year old Jeff, what would you tell him?

Jeff Jonas: [00:34:58] Think twice before getting married. I have three ex-wives, man. How's that for more.

Jeff Jonas: [00:35:08] More record setting? Man, I'm an overachiever. I just hate to think. I just hate saying no. It's like you marry me. I'm like, yeah, sure. I hate letting people down. I don't know. That's all I got on that list. Wait, wait, wait. Maybe. Maybe a better one is I would have said I would have said to myself, stay focused on things that are useful and sustainable. I create things that are useful and sustainable. I did spend some time creating things that might have been intellectually curious, but they weren't and they didn't have the same utility and sustainability.

Harpreet Sahota: [00:35:47] Would you care to elaborate on on any of those projects?

Jeff Jonas: [00:35:50] You know, I wrote some software for the North American Llama Society to write, to do llama birth certificates. Like they needed it. And it made I made a few dollars doing it, but it wasn't as impactful as my work in modernizing voter registration in America or as impactful as reuniting loved ones after Katrina, that couldn't find each other because they got dispersed across the relocation of facilities. Like those, those are just more impactful. So I wasn't careful. It's more like a locus has no directional control. I'm told if you create a line of fire and you put Locus right next to it and the locusts are going, hey, this is hot. I got to get out of here. They don't know where they're jumping. They're just jumping. So half jump in the fire. And so I was somewhat directionless. I would just take the next piece of work that came along without much thought about things that would it be more useful and sustainable. So I might have had a bigger lifetime impact. Had I'd been a little more cautious. Now, I'm very cautious around that.

Harpreet Sahota: [00:36:48] Even so, like you still got, you know, over one hundred inventions that that are credited to you, I think that that having that kind of attitude, that leads to being really prolific. Right. So of these hundred inventions, you know, obviously the the llama birth registration is low on the list, but what would you say is your favorite one? 

Jeff Jonas: [00:37:12] Well, voter registration modernization. We Americans are highly mobile society where it was at least two months ago. But, you know, if you registered to vote in one state and then you move to another state, you'd forget to unregister. And so you end up with more people on the electoral rolls than even live there. And so we built a system that's now running over half the country. So and we did it in a privacy enhancing way. And so we're really quite proud of that. I did some work with some astronomers at the Institute of Astronomy at the University of Honolulu, Hawaii. Whatever it is, though, I think and and that was pretty exciting because knowing that asteroids hit each other and then you don't know where they're going and they didn't really have a way to forecast that. And I invented a a method to compute it. They didn't think there's enough computers. Really needed 10 million computer hours to try to forecast when asteroids hit asteroids. And I figured out a really simple way to do it very fast. And let astronomers be able to watch asteroids get close to each other and possibly hit each other like anticipate when and where in the sky night sky. Six hundred thousand asteroids might cross each other's paths really close. So that's pretty exciting.

Harpreet Sahota: [00:38:23] Wow. Would you mind getting that going a little bit more deeper into that into that project is kind of, you know.

Jeff Jonas: [00:38:30] Yeah. Well, I was I was visiting with these these astronomer types. And because I don't really have any smarts in the area, I had to ask them a bunch of very stupid questions. But but along the journey, they explained one of their challenges. They they said, you know, we we have all these asteroids. We know they don't hit Earth because we can check that. We check them. When we when we register asteroid, we look at its journey and make sure doesn't hit Earth. And I said, well, you know, the asteroids always when you look at the night sky, you can always kind of and predict where the asteroid is going to be. And we can't always predict where it's going to be. Cause sometimes they've rotated and the sun's in the opposite direction and you can't see them. Other times we had the orbit calculation a little bit off. So it's not kind of where we thought it was. And sometimes they hit other asteroids and I don't even know where they're going. And they said, we've only seen it twice. The Hubble telescope was taking a deep space picture. And in the middle of the picture is a giant X. Like what? And it was the aftermath of two asteroids that hit each other. So it looked like an X and they said there's only between one at the time.

Jeff Jonas: [00:39:31] And then because I don't have any background in math. I asked him a stupid question. I go, well, if you can check the orbit of Earth and check the orbit of each asteroid to make sure doesn't hit Earth, why don't you just check all the asteroid orbits against each other to see if they're going to hit each other. And they looked at me like, you fool. They said it's multi body orbit math, which is kind of very expensive form of math. We have more than one gravity pulling thing, pulling on an object, and there's six hundred thousand asteroids. So it's an N squared problems. So it's an N squared problem with multi body orbit math. It's not like astronomers have that much compute. And later, an estimation was done as you need 10 million hours. And then I said, yeah, but why would you do it that way? Why wouldn't you predict when asteroids were going to hit each other by doing it this other way? And then, like, you know, they looked at me whimsically and I said, well, let me explain. Why don't we use...is this too much detail, by the way?

Harpreet Sahota: [00:40:26] This is good. I want to hear.

Jeff Jonas: [00:40:27] OK. I said OK. So I said out of those six hundred thousand asteroids, where to take asteroid number one and we're gonna use your fancy math, which turns out it's Fortran, which is a little embarrassing. We're gonna use your fancy math and we're gonna ask it, where's that asteroid gonna be tomorrow at noon? And your fancy math is going to come back with an exquisite point in space, they call it "R.A.N dec", it's like latitude, longitude on earth. It's a 360 degree system. So we're going to take the asteroid, we're going to take one asteroid, and ask it tomorrow where it's going to be at noon. And the fancy math is gonna come back and say "Oh! It's gonna be RIGHT HERE!" Very precise. And I said what we're going to do is we're gonna put it into what I call a space time box. We're going to fuzz it up to a big unit of space and time. Going to put it in there. It's kind of like saying, what zip code are you gonna be OK? And then when you go to the same asteroid, we're to say, hey, we're going to be the day after tomorrow at noon. And it comes back with a "Oh! Right! Here!" And then we go, oh, yay, yay, yay, yay.

Jeff Jonas: [00:41:19] What zip code is that?. And so we take the first asteroid and go twenty five years forward at noon and ask it where it's going to be. But in each case, is it up to a zip code? And then take the next asteroid and go through each asteroid once one once a day for twenty five years and get those points, but fuzz them up to zip codes. It turns out if you do that on any given day, there's only two thousand asteroids on average per day in any zip code. Well, you narrowed the problem down. So then you go back to just those asteroids to go back to one space time box. It's pretty big. It's like a zip code. There's two thousand asteroids in there at the same time. So just to those two thousand, you go to them and you say, hey, where are you going to be at 1:00 a.m., 2:00 a.m., 3:00 a.m. just by the hour. And then you put it in a smaller spacetime box. It's kind of like saying, hey, what street are you gonna be on? Not what zip code, what street? So you do two passes the number of asteroids that are on the same small space time box. Like what street in the same hour is so small.

Jeff Jonas: [00:42:12] And then you can run their heavy math on it. And so in under sixteen hundred hours with some doing some compute in parallel over over a few weeks we gave them a twenty five year forecast of every asteroid's proximity to every other asteroid. And since then there's been a new science papers that have come out because of this. Even if the asteroid get near each other but don't hit because they both have mass. They change each other's orbit a little bit. And you can't really estimate the mass of these asteroids. They're just shiny dots, a light. You don't know if it's like a soft cotton candy cloud or if it's iron. But when you can see how they change each other's orbit, then you can understand their mass. And so I got a cool email from them it said it's the first time in the history of astronomy. We knew where to look, and when.

Harpreet Sahota: [00:43:00] That's awesome man. 

Jeff Jonas: [00:43:01] Yeah watch Two asteroids glaze each other. Yeah. Much of my work is used if I were to generalize that across all of my work. It's my work is often about helping humans focus their finite resources.

Harpreet Sahota: [00:43:13] Wow, that is awesome. Sounds like like a little bit of a on the surface like Divide and conquer and recursion techniques that you apply to that problem.

Jeff Jonas: [00:43:23] Yes. You can call it smart clustering. I know you can call it coming up with the candidates. My work in entity resolution, if you'd have a billion records loaded and you get a new record, you're not going to go table scan the whole database to try to find the record. So you have to go find candidates fast. So the idea that I came up with the asteroids was how to take every how to find all the candidates and then only do heavy math on the candidates.

Harpreet Sahota: [00:43:45] That's awesome. Well, thank you for saving all six billion of us. I think...

Jeff Jonas: [00:43:51]  You're going to owe me. You are gonna owe me.

Harpreet Sahota: [00:43:54] Checks in the mail. So which which of your inventions do you think is most relevant now to the current times?

Jeff Jonas: [00:44:01] My new company, Senzing is democratizing this entity resolution problem. Everybody struggles with it. Everybody's spending too much money on it. Nonprofits have duplicates in their mailing lists. I let him just download and clean up their mandolin list for free. We're letting anybody that's doing humanitarian non-profit work for COVID. We'll give those big licenses that they need. But it's in the epicenter of a lot of from CRM to antifraud to contact tracing. Understanding who is who is related to who has been really expensive and hard. And we just made it easy and part of the invention. It's a self tuning, self learning, self-correcting the past, learning it all while in real time, no reloading, no batches. You have to pre train the data sets. It's a very... Wired ran a story called "AI Needs Common Sense." And I realized after reading it, it's the best way to describe what we do. Have you seen these AIs where you showed a picture of a bus? But you change a few pixels in the picture of the bus. It still looks like a bus to you And I, but the machine then says that's an ostrich.

Jeff Jonas: [00:45:06] So then you take a picture of a bridge and you change a couple pictures, pixels, and then it comes back with the same AI goes "That is an ostrich". And you and I are going "That's a bridge!"

Jeff Jonas: [00:45:14] So Wired did a story on common sense, And so that's what I've created with tens of millions of dollars is a real time learning A.I. that starts with common sense. And then it real time learns to improve its decisions forward and backwards, like changing its mind about the past. Mango style. That's exciting. And I'm trying to make it. I'm trying to take something that was tens of millions to build and make it super affordable to everybody.

Harpreet Sahota: [00:45:39] So cool. Do you want to get into that a little bit? Or are you allowed to - I'm not sure.

Jeff Jonas: [00:45:43] Well, yeah, whatever, whatever. What do you want - the one minute version? I'll tell you what, why should a machine or somebody trying to entity resolve data with learning systems have to learn? Get Richard Dick, Ricky, Rick, Ricardo are all the same. Like, why? Why should they have to relearned that? Why should they have to relearned it? Mohammed is but one way in Arabic, but over one hundred ways in English. And Elizabeth, Beth, Liz, Lizzie. So we have a machine learned library of all those named transliterations across cultures. It's just built in. We got it from our IBM relationship. And it's already had the investment. That's common sentiment. Well, you have to relearnt. Same with addresses. Addresses are messy. It comes out of the box with a machine learn address parser that was trained off of open street maps with machine learning. So. So those are common. Those are common sense aspects. And then another common sense item is if the name and passport are the same. That's a pretty good indicator as the same person. What Real-Time Learning would be if if the passport fields got some garbage in it, like it's just got one, two, three. You have 10 million records and the passport numbers is 123. Well, passport numbers might be usually good, but not that one.

Jeff Jonas: [00:46:57] So in real time, that wasn't part of common sense. Common sense as passports are great, but in real time learning it goes. Wow. Fifty people have a passport number called one, two, three. That can't be. So then it gets smarter going forward. Then it goes well, now that I know that I changed pass and passed.

Jeff Jonas: [00:47:12] And so that's a very simple example of interweaving common sense with real time.

Harpreet Sahota: [00:47:20] Continuing with the lightning round - wow, that was great.

Jeff Jonas: [00:47:23] It don't feel like a lightning round anymore.

Jeff Jonas: [00:47:25] Tell me about your Asteroid Hunter Award.

Harpreet Sahota: [00:47:29] So what's the best advice you ever received?

Jeff Jonas: [00:47:32] You know, maybe I don't know...I received it from from the feedback loop of the world. But when I went bankrupt, I realized that I had made a lot of promises and I wasn't able to keep them. And if you want to do really high quality work and build a reputation, you've got to really deliver on what you promise. And so that was kind of a gift from the planet about just be really careful about making commitments you keep and then keep them.

Harpreet Sahota: [00:47:57] Awesome. Good advice. Do you have a favorite Iron Man event?

Jeff Jonas: [00:48:02] Oh, I like the one... You mean Swim, bike, or run or a location

Harpreet Sahota: [00:48:06] Both. Both. Yeah.

Jeff Jonas: [00:48:09] The cycling comes easiest for me. You can coast. That's my favorite. Ironman is the Ironman in Austria.

Harpreet Sahota: [00:48:20] Beautiful country - Was that because there's a lot of downhill coasting you can do on the bikes?

Jeff Jonas: [00:48:25] No, that race was pretty flat. That was what we would call a flat course.

Harpreet Sahota: [00:48:31] So what motivates you?

Jeff Jonas: [00:48:32] Right now, I'm just trying to make a difference, you know, especially in this COVID world. I'm trying to make sure my all my team is doing well and their families are doing well and trying to make sure my network is doing well. Those things motivate me. In data, I love producing really great outcomes and I'm becomes super focused. And it's a survival point is, is if you can't...in my business, if you can't figure out how to help somebody reduce their costs and pay for my software within weeks or you know month, a few months or maybe weeks, there's no business to be had. So you just have to find to find where the wind is and get in the wind. And so, yeah, I'm focused on helping people around me. Yeah.

Harpreet Sahota: [00:49:10] So how can people connect with you? When can they find you?

Jeff Jonas: [00:49:13] I'm Jeff@Senzing.com. I answer every e-mail I get from everybody.

Harpreet Sahota: [00:49:19] This is true. I've tried it.

Jeff Jonas: [00:49:24] I'm also on LinkedIn. I tend to only connect with people that I kind of know or you know, somebody commented that they listen to your podcast then they're in. But I'm pretty cautious. That's because you know how LinkedIn recommends happy birthdays and congratulations for new jobs.

Harpreet Sahota: [00:49:42] Yeah. Yeah, it'll flood your inbox.

Jeff Jonas: [00:49:45] I actually do all of those and I customize every single message.

Harpreet Sahota: [00:49:49] Wow. Yeah, that is that is awesome, man.

Jeff Jonas: [00:49:52] Ten thousand people in my LinkedIn.

Harpreet Sahota: [00:49:55] That's so cool, Man

Jeff Jonas: [00:49:57] It's important to be acceptable. You know, I learn a lot by connecting with others and it's created a lot of really goodwill. So I'm highly accessible. Just if anybody wants to chat, shoot me. It doesn't matter if it's about your mom's health. I don't I'm not an expert at that. I guess that's no, I'm not a doctor. But anyway. Yeah, accessible.

Harpreet Sahota: [00:50:16] Definitely. I know I learned a lot from you today. And, you know, whoever's listening to this podcast, definitely there's a lot so much to take away.

Harpreet Sahota: [00:50:23] Thank you so much for your time, Jeff. I appreciate you taking time out your schedule to be here. It really means a lot to me. So thank you.

Jeff Jonas: [00:50:30] Thanks, man. I enjoyed it.