#896 Rise of the Machines
Edward has type 1 diabetes and is here to talk about advancements in DIY algorithms. Android APS talk.
You can always listen to the Juicebox Podcast here but the cool kids use: Apple Podcasts/iOS - Spotify - Amazon Music - Google Play/Android - iHeart Radio - Radio Public, Amazon Alexa or wherever they get audio.
+ Click for EPISODE TRANSCRIPT
DISCLAIMER: This text is the output of AI based transcribing from an audio recording. Although the transcription is largely accurate, in some cases it is incomplete or inaccurate due to inaudible passages or transcription errors and should not be treated as an authoritative record. Nothing that you read here constitutes advice medical or otherwise. Always consult with a healthcare professional before making changes to a healthcare plan.
Scott Benner 0:00
Friends, this is episode 896 of the Juicebox. Podcast. Welcome to it
today I'm going to be speaking with Edward, he is a type one, he's a father. And he's a computer programmer who uses Android APS to give himself insulin. And he's got a lot of cool thoughts about it. It's a lot to do with where things could be going where they are going. It's incredibly interesting. You're going to be daunted at first by the length of this episode. Don't be. Just settle in. Listen to the conversation unfold, and imagine what could be. While your imagination is running wild, please remember that nothing you hear on the Juicebox Podcast should be considered advice, medical or otherwise, always consult a physician before making any changes to your healthcare plan, or becoming bold with insulin. Here's a couple quick things. T one D exchange.org. Forward slash juicebox. Go complete the survey you're going to help diabetes research if you do that T one D exchange.org. Forward slash juicebox. You can get 10% off your first month of therapy@betterhelp.com forward slash juice box. And you can save 35% on your entire order@cozier.com By using the offer code juice box at checkout.
This episode of The Juicebox Podcast is sponsored by us med get your diabetes supplies the same way we do from us med us med.com forward slash juice box Why would you go there it would be to get your free benefits check. And to get started. Don't like the internet. Call 888-721-1514 Get started today with us med they always provide 90 days worth of supplies. They have fast and free shipping. The podcast is also sponsored today by touched by type one. Now I'm going to be speaking at the next touch by type one big event coming up soon. Check it out, touched by type one.org. Also find them on Facebook and Instagram.
Edward Robinson 2:16
So I am Edward Robinson. I am a father of four children. I am a computer programmer. And I was diagnosed with diabetes at age 18. In 2006
Scott Benner 2:35
How old are you now? 34,634. Okay, for hit? Yes. Is that uh, are you doing that for a lot of work? Where are we? Why are you building an army basketball team? What's going on? Exactly.
Edward Robinson 2:54
So I was one of four. My wife was one of three. And so we always wanted to have a few kids. And it's turned out to be four.
Scott Benner 3:04
Wow. Well, you did it. Are you is there like a hole in your floor somewhere where you just keep pouring money to send the kids to college with later or is that how that works?
Edward Robinson 3:16
No, I just go back to work and work more.
Scott Benner 3:20
Are you prepping them now? Because they're young? They imagine right? What are their ages?
Edward Robinson 3:24
They are young. So the oldest is seven. Then five, then almost three and then one. Or almost
Scott Benner 3:32
I'd be whispering in their ear. You want to be a short order cook. Someone has to pick up the recycling. Billy. Come on, buddy. You don't need college? I don't know, man. That's yeah. Have you ever thought about it?
Edward Robinson 3:49
Yes, yes. We started 529 For them quite early.
Scott Benner 3:54
Well, I bet you did. Yeah. Okay, well, congratulations. No
Edward Robinson 3:57
birthday presents just donations to 529 Oh, we'll just go to the park and save the money. Yeah,
Scott Benner 4:05
I want a lollipop. You take this quarter you run home. You put it the home before it just shut up. Oh, my God. I mean, listen, at a an expensive education at this point. Private education is 65 $70,000 a year. Right?
Edward Robinson 4:24
Yeah. And I can only imagine what it's going to be in.
Scott Benner 4:28
I have to tell you, there is an amount of money that my wife and I imagined when Cole was born. And we were wrong by 130% Maybe. So, I mean, maybe things will change. But seriously, I'd be like there's no reason somebody runs the cash register at Target and I think you'd be great at it. And I would, I would I would just get them a nice, nice blue collar job. Although, what kind of work do you do?
Edward Robinson 5:01
So I'm a software developer, I work for a small consulting firm,
Scott Benner 5:07
do you teach them how to code? Are you like, sit down? I'll show you Python today, like, what do you do with those kids?
Edward Robinson 5:12
No. So like, I actually build the application. So we work with different, different clients. And then they have different problems that they want to solve, and will work with them to solve their problems. It almost always involves building applications. But sometimes it's just like management help, or helping them solve their process problems and things like that. But usually, there's applications involved. And recently, I've been working on a project for artificial intelligence and machine learning.
Scott Benner 5:43
Wow. So far, which is really exciting. That's amazing. Actually, I'm gonna hear more about that in a second. If I asked you this question. And this is not oppression, don't take this the wrong way. If I said to you, Edward, I want to make an app that the people that listen to this podcast can have access to the show through, like, I want to make my own podcast app. And I'd like to be able to send them information through the app and break up the series into groupings for them. Would you like chuckle and be like, that's easy? Scott, I could do that on Saturday afternoon? Or is it a big thing?
Edward Robinson 6:19
So I don't do that much. Devote? I've done some Android and iOS development. There are other people on our team who have done a lot more than me and could do it in an afternoon. It would probably take me a weekend.
Scott Benner 6:31
Wow, that's insane. What did you go to college for that? And why are you smart.
Edward Robinson 6:38
I went to college for engineering. And then I within engineering, I studied computer science. So I've been writing programming, or doing programming since 2006.
Scott Benner 6:51
That's crazy. I'm watching my son. Now he just graduated from college. He's got a quantitative econ degree. And, you know, he's going through college, and every once in a while, he'd be like, they make us these these programs we've never seen before, but they don't tell us that much about them. And, and now he's graduated applying for jobs, and in his free time just sitting in his room getting certifications and like, anything he can find because he realized a little too late, that that should have been his minor. He took a mathematics minor, and he's like, I should have minored in this stuff. Because every one of these jobs uses this.
Edward Robinson 7:26
It's amazing how many jobs these days require programming of some degree. Yeah.
Scott Benner 7:32
It's really something and he's, he's picking it up. Actually, he said something incredibly frustrating. The other day, he said, I probably didn't need to go to college, I could have just taken all these courses. And all I could think about was the month. It was like, don't say that to me as like, even if it's true. Never, ever, ever tell me. You know what I mean? Like, please just don't say it.
Edward Robinson 7:59
But something is true, especially around programming. There's just tons of things you can learn online for little to no money. And you can find examples, there are still a lot of like foundational concepts that are a lot easier to learn at a university.
Scott Benner 8:15
Right. I also don't know how you get in the door without some sort of a degree. You know, I to me, that sounds impossible. But I'm sure there's some barrier to entry. Although I'll tell you my brother, my youngest brother is 10 years younger than I am. He was we I never talked about my brothers on the show. But my youngest brother was we used to say either in trouble, or you were about to find out he was in trouble. Like there was no like battlegrounds. You know what I mean? Like he was either doing something he shouldn't be where he was. And you didn't know where the cops were there. So like, you know, there was no middle ground for him. And coming out of high school. He just worked at a pool supply place. And I have to tell you, like my I growing up, my father had a very blue collar job. I don't know that he had any more expectations than that my mom did retail. You know, I think that's probably what he was thinking. And he just met somebody while they were at the job. And they said, you know, we think you could do this entry level thing in it. And now he oversees like a huge department of people who shoot satellites into space. And so that's pretty cool. I don't know how that works. You know what I mean? Like and he sometimes I look at him and I'm like, he's like He got he got away with it. Never went to college, never done anything. And it's making you know, he's making the money just like everybody else. So it's, it's hard to hard decide, but what what leads you to that where you just always inclined or are you super interested?
Edward Robinson 9:48
So I always enjoyed math and science. And so then I went into engineering because I figured it would use those I didn't plan on doing computer science. But I took a couple of classes in college and found that I was pretty good at it. Yeah, there's the first class was like basics of programming. And you had to write programs in, you know, pretty complicated programming languages. These days, there's a lot of like, yeah, click and point programs that you can set up and you can do it with your kids. But this was like hardcore. In order to even like edit files, you needed to know specific commands, and use these like, editors are pretty complicated. And I found that I was pretty good at it and, and enjoyed it. My roommate took the same class with me, and he was really struggling, and he just hated it and want to blow his brains out. No kidding.
Scott Benner 10:46
That's, it's listen. I am a person who in Hold on a second. There's my wife, not I mean, there's she's got two jobs. And one of them stopped the dog from barking.
Her other job is to come up with reasons why we can't have sex started to be so is there a real, you know, stuff she could put on her CV where she could really speak thoughtfully about being adept for sure. If she or he stopped, maybe she stopped. I made it to I was in a very weird situation, we'll get your diabetes in a second. Where I this is so it's embarrassing. They hold on a second. Now she has questions about the dog. But now she doesn't know she needs management. And here we go. She tells me she's I can do it. I can do it. I got pulled out of my kindergarten. But this is going to pull back a curtain on me. I guess I got pulled out of my kindergarten class. And I remember it like right now to this day, my teacher, somebody knocked on the door, my teacher came over to me and said somebody wanted to speak to me outside, I went out in the hallway and this big man, you know who now I realized could have been five, six, I have no idea what was standing over top of me. And he started talking to the teacher and the teacher kind of bent down next to me, I remember her name, it was Mrs. Moon. And she said, Scott, this man thinks you're smarter than the other kids and they want to give you a test. I was like, It's not good for a five year olds ego. I was like, yeah, those other ones do seem dim to me, I get what you're saying, like, let's let's get me separated, you know. And they gave me an IQ test, which I now realize was an IQ test, I realized it because they gave it to me every year, like every year at the beginning of the year. And I'll tell you that I will not tell you my IQ. But I test the same exact number every time I take it. Even if I were to take one online right now I just hit that number, I don't understand what those tests are, how they work. And it was probably the worst thing that ever happened to me. Because in the 70s, they were doing something called a might state state they call it academically talented. And I'm gonna tell you, Edward, I am not academically talented at all. If you just you know, make the list of STEM stuff in your head and assume I'm not good at any of it. Just I don't know what my IQ helps me with. But it is not those things. And so I beat my head against that wall till eighth grade, like in a room with people who were clearly just smarter than I was that, you know, everything we did, they were better at, they understood it more quickly. And when it came to the math, I was really lost. Like really, really lost. They tried to put me in algebra when I was in sixth grade. And I don't know if I've ever told the story or not. But maybe a semester into that that year. I pulled the teacher aside after class and I said, Listen, it's too late for me to drop this class. I'm stuck. You know it and I know it. I don't understand this. I'm never going to understand this. We're in trouble here. Right? I'm making a deal. And he's like, what's this? Imagine I'm in sixth grade.
Edward Robinson 14:17
I'll give you 20 bucks. How old does
Scott Benner 14:19
that even make you in sixth grade? I mean, I was five and kindergarten. So 67892. I was like 11 years old. Okay, maybe 12. And I'm standing in front. I'm like, listen, here's what we're gonna do. I talked him into allowing me to sleep through his class for three more semesters. And I told him I was going to take a general math class in summer school to make my credit up and he said, Okay, like without even a fight. He just was like, alright, and we kept that. Three quarters of the year. We kept that agreement. I came in I didn't bother him. I failed his class. He failed me just zum zum zum and I went to summer school If I do not when you start putting numbers and letters next to each other and telling me they equal something, I can feel my brain short circuiting. I cannot follow it. Like to save my life. And the reason I tell you that
Edward Robinson 15:11
I'm always impressed by your math skills on the podcast. Sure.
Unknown Speaker 15:15
Sure.
Edward Robinson 15:17
I don't know if you have a calculator ready when you do it, but
Scott Benner 15:21
I can do simple. There's usually pretty good action. But my point is that when I listen to my son talk about math. I think our mailman must be really good at math. Because I can't imagine he's my kid. The way he understands numbers. It's really bizarre. Like, I don't know, it just happens for some people. And, and for others like me, by the way, it did not happen for so
Edward Robinson 15:47
well, on the on the flip side, I really struggled with other things like, you know, reading and history and other classes like that. So yeah, nobody gets everything.
Scott Benner 15:58
You're being delightful. I didn't do any of this. I can't believe I have a high school diploma. My I forget what year it was in middle school, we had this great English teacher, and he's like, we're going to read Romeo and Juliet. And we're going to read it together. And I was just like, Oh, I'm not reading that. And we had my senior year of high school, on my first day of my senior year of high school, the English teacher said, we're going to do a research paper, but don't worry, you have the whole year to do it. And he explained what it was and this whole thing, and I swear to you, I raised my hand. And he's like, yes, your question about the research paper? I said, I do. I said, What happens if I don't do it? Can I still graduate? And he goes, No. And I was like, Yeah, I got a job. I don't think I could do this. And I was just like, it sounded impossible to me what he was saying, and I'm sure if I was back in that room right now, he's probably saying, you know, make yourself acquainted with something and write a four page paper on there. It was not a big, it wasn't a big deal. But he was asking. And to my convictions, I kept good grades in that class that whole year. And I thought, like, he can't fail me for not handling one paper. And I didn't handle the paper, and he didn't fail me. So anyway, his bluff did not work on me. I don't know. So I think that's where my intellect works. It works in. Like, it's why you if you hear me say like, if the zombies come find me. I'll be okay. But we're not gonna do any algebra. Anyway.
Edward Robinson 17:43
So I think, God, I think it's interesting that I think a lot of times, people assume that they can't do things. But I think if they find the right ways to learn it, that you'd be surprised how people can do it. So I always thought that I was terrible at music. When I was growing up, I tried playing the piano, the guitar, it was just a disaster. It sounded horrible. But a few years ago, my wife found a piano for free on Facebook. And we went to go pick it up with a U haul. And I was like, why on earth? Are we getting a piano? This is going to be just a waste of space. But I I tuned it. And I learned a couple of songs on the piano. And I was really surprised how finding different ways to learn it. I mean, just watching YouTube videos, but how I was able to actually learn something that I had always assumed that I was just terrible at and it just wasn't going to work for me.
Scott Benner 18:43
No, that's a great point. And the truth is that I think one of the things that got in my way of expanding when I was younger is that I was adopted. So that nothing about my, the way I saw the world. Did anyone else see it in my house, and those people were in charge of giving me things to do? So they would direct me towards? I mean, no one really way I can say just like more blue collar endeavors like I went to school, because I don't want to paint them as like preppers or something like that. They're not that but I went to school because the state makes you send your kids to school. So that's why they sent me to school. They sent me to school so they wouldn't get in trouble. And it was a great place to park me lives while they were working. Yeah. It was. It was free childcare that no one wants said to me, you know, when you're in high school, you would go to college. Like I didn't take the LSAT. I went to I remember being in school the day I realized my friends were signing up to take the LSAT, and I didn't know what they were doing. Like I had to ask a friend I was like, why is everyone doing this? Is this something I have to do? And and they're like, No, you take this they give you a score and then you send the score like that that was being explained to me like I was like I was from another planet you know, and I'm I was like, Yeah, I'm not doing that. I was like, I can't go to college. I No one's ever mentioned college to me before. And then I did look into it a little bit. And even at like a local community college level, I couldn't afford to do that. And I don't mean like, I couldn't afford college. I mean, I couldn't, but I couldn't afford a car to get me to the college. Like, like, there was nothing about my life that was going to send me to school. And no one ever wants, you know, rigged the board for me to get to the end. And so I just, I went to my, my high school graduation, we came home, my mom gave me an ice cream cake to celebrate, I remember that. And then I went to bed got up at five o'clock in the morning and started working full time in my uncle's sheetmetal shop. And that was it. It's just what I thought I was gonna do. Like, my friends all went on vacation. And they were like, I was like, I gotta go to work. And I just, I graduated from high school and went to work. So I don't I think you're right, I think if someone would have pointed me in a different direction, and given me a couple of pointers, I might have been, I might have been in a different situation. But as it stands, you don't want to ask me about anything particularly technical, except for some reason, diabetes, that I see that I seem to understand. Okay. All right. So do you work with your kids with a DD? The little like, was wiggy? Like, stuff? Do you think they'll be interested? Or do you think you'll be like me one day, would you like these kids have none of my interests at all.
Edward Robinson 21:30
Um, they actually already have it in kindergarten, they've got some like, code ish. Like games and things like that, for kindergarteners, which is really amazing. I think really just getting them to enjoy math and science and other things, but also exposing them to everything as much as we can. It's what we really like to do. My son, we're actually super proud. He just had a birthday party. And we were making up pin the tail on the donkey games. And so it's like, you know, pin the arm on the robot or things like that. And he wants to do pin the science on the stem. Look at that. And
Scott Benner 22:16
yeah, let's get that get a good job on that. I like that. I think I was actually inclined to it. Because if you've ever heard me talk about it, I bought the first RadioShack computer. Like I had it. And it was all about programming. And I had programming books, I couldn't figure the first thing I tried didn't work, I gave it up. And I realized now the reason I gave up so quickly, it was financial. Like, I took me so long to save up the money to buy this computer, that when it didn't do what I thought it was going to do, which really just meant that I did something wrong. I took it back right away, because I couldn't afford to own something that that didn't that didn't do something for me. So it's interesting that how that stuff impacts you. Anyway, sorry, there's a long way to go. I was just very interested. So you're 18 and 2006. You're on your way to some sort of a dork Olympics. It sounds like to me. You get diabetes. Is there? Yeah. Is there anything that would have made you think prior to that day that type one was a possibility for you?
Edward Robinson 23:26
I knew nothing about type one. Didn't couldn't even tell you what it was.
Scott Benner 23:34
Nobody? Nobody, nothing. Yeah.
Edward Robinson 23:37
Nobody in the family. My grandmother had some thyroid problems, but nothing
Scott Benner 23:47
was linked to yourself. You wouldn't look at your grandmother and be like grandma's hair falls out a lot. I'll probably get diabetes one day. Yeah.
Edward Robinson 23:53
Yeah. So I started losing weight. I played a lot of sports growing up in high school. played soccer, lacrosse, hockey, skiing, snowboarding, tennis, all sorts of things. But I wasn't playing sports at the time. And I started losing weight. And my parents were concerned and they, they actually told me that I needed to put on weight and forced me kind of start drinking milk. Which is like the worst thing that could have happened. I'm sure my blood sugar just went straight through the roof for hours on end. And one day I went on the treadmill and afterwards, I felt amazing. I was like I'm gonna take up more running. Now like no, you can't run because you're losing too much weight.
Scott Benner 24:49
Keep drinking the liquified fan. It'll be okay. Yeah,
Edward Robinson 24:53
this is helping you. The I mean, they're they're absolutely wonderful, but it was just kind of ironic. because it worked out. And then. And then finally they're like, Okay, we gotta go see a doctor. And so I wanted to school to take a test. And then afterwards, they took me to, to the doctor. And there, they tested my blood sugar. And it just read as high. So it must have been, yeah, over 400 or some something crazy high. Yeah, to be off the meter. So then we drove to Yale, pediatrics for diabetes. And on the way there, I fell asleep or passed out in the car. And the next thing I knew, I was in the hospital, and they're trying to get me into this bed, and I woke up and I was like, trying to push people away. And they, and then I guess I fell asleep again. Okay, and you're sort of woke up? Do you think at some point later, the K
Scott Benner 25:57
or did they ever say? Uh, maybe.
Edward Robinson 26:01
I mean, I, part of me wants to call them up and see if they have my records from, you know, I mean, ages ago, but I don't know if they would, because
Scott Benner 26:09
I'm actually I'm looking at you today. And I can only see you from the shoulders up. But there's something about your frame that makes me think you're tall. Is that fair? Not particularly No, just five 510? Oh, well. All right. Listen, still, I'm trying to imagine them getting an 18 year old view from the cart into the hospital that you noticing like you had to have been out of it.
Edward Robinson 26:29
Yeah, my mom said that she tried to wake me up. And I wouldn't wake up in the car. They had, the doctor had said you could call the ambulance. But by the time the ambulance gets here to then take you to the hospital, it'll just be faster to just drive me straight there. So I went there. And so I was diagnosed, I started on MDI, so it's two shots a day with, I don't remember what the insulin was called. But it's the one that's cloudy, and you had to like roll it before taking that, and then Novolog. And so I do a shot in the morning and shot in the evening. And every day I would call in to the to the doctor's office and give them my numbers for the day. And they're telling me what shots to do and how much to eat for like the next day.
Scott Benner 27:21
How long did that go on for?
Edward Robinson 27:24
So that went on for six months to a year. And then I got it on a Medtronic pump.
Scott Benner 27:31
Wow, wait for for a year, every day you call a number, like you were.
Edward Robinson 27:39
I think that after a few months, I had to call in, like once a week or
Scott Benner 27:44
something. Okay, still, that's amazing. Even if it was just a few days.
Edward Robinson 27:48
I remember this was in like the time of like, beepers and stuff where you call a number to send a text message to a beeper, and things like that. So
Scott Benner 27:57
that must have felt like it must have felt like you were spot looking like just a disembodied voice on the other end, you're like 118, she's, she gives you coordinates back and you're like, yes, yes. Okay. And? Well, that's fascinating. That would have been a much cooler way to think about it. Yeah, you probably should have had a little of my brain that would have helped you not be sad. I'm assuming you were. And you're also missed this call. This is going to college age two. Right. So this happened to you. Sounds like in high school, because you said you still went to a PDF.
Edward Robinson 28:28
And it's actually right before spring break. And I still went on spring break to the where'd I go? I don't know, somewhere in Mexico or something for spring break. Which was a lot of fun. But you went to a probably not the best idea.
Scott Benner 28:45
But did it? See? It's interesting, isn't it? Because, you know, I interviewed a woman yesterday 72 years old, said diabetes for 50 years. She doesn't have one thing wrong with her. It's just fascinating. Like, you don't I mean, like you just It's interesting how, I mean, she describes her care in the past and you think, like, how does she not die? 40 years ago, you know what I mean? Like, like, eat and it was in a time when the expectation was I've done a few older people's interviews this week. And a gentleman who was older as well told me, you know, when as soon as I was diagnosed, he's like, they you know, it was a shortened lifespan. It was you're not going to live as long. Yeah, don't even worry about it. Like you said his care was even centered around the idea that you know, we're just trying to keep you going for as long as we can. And if you were getting cloudy, then I mean, that was still that that was that kind of care. Did you ever have that feeling? Did anyone ever say something like that to you?
Edward Robinson 29:49
It's so the doctors I remember them putting a really positive spin on it. Being like, you know, there are these basketball players who have diabetes or like these other famous People there, lots of people have diabetes, it's gonna be totally fine. And I remember thinking to myself like, Okay, I've diabetes, whatever, I'm just gonna live my life with it. And things will be fine. Little did I know. Yeah. What all that is gonna be? Yeah.
Scott Benner 30:17
I think they were probably telling you about Chris Dudley if it was 2006 He's been on the show, maybe? I don't know. I mean, he's the only one I can think of. So, you see, you're a little naive, right? went off on your trip, and then you headed off to school. But they gave you a pump your freshman year in college. Yeah.
Edward Robinson 30:37
So I started on a I forget what the maker was, but it was before Medtronic. So I started on a pump. And then a few years later, I switched to a Medtronic pump. And I was on those for a while.
Scott Benner 30:53
So did you go to human log? At that point? Do you remember?
Edward Robinson 30:57
I think it was Nova
Scott Benner 30:59
Nova log. So Nova log in a pump. And that changes your company. So you went from like calling people to doing everything yourself?
Edward Robinson 31:08
II? Yes. I didn't do it well, but I did it myself.
Scott Benner 31:13
That's what I'm trying to get. So what were your outcomes, like in college?
Edward Robinson 31:18
I don't know. But they weren't really great. I remember having lots of lows. And but when I had lows, I was kind of funny. These days when I get low, I just still like normal and can think, but I would like say ridiculous things or you know, do funny things. I'd get really stubborn sometimes. My girlfriend at the time, who's now my wife, she would tell me all these like stories where she would be like, you have low blood sugar, you need to have something to eat. And I'd be like, no and aren't get argumentative. And she'd be like, you know, what, if, if you don't have something to eat, I'm gonna call your mom. And then I put my hands over my mouth.
Scott Benner 32:04
Like, sort of like home alone. It's like, oh, my god like that. Like, no, like,
Edward Robinson 32:09
You can't force me, I'm just gonna cover my mouth.
Scott Benner 32:14
Like you're four years old at work.
Edward Robinson 32:18
Yeah, and then another time, I had low blood sugar, but I thought it had to be high blood sugar. And I didn't have my meter with me. And she was like, I'm just gonna eat these cookies, and put one here next to you. And of course, I snatched it up and ate it.
Scott Benner 32:35
I just had a moment with Arden the other night where I was like, you have to do something about your low blood sugar. And she's like, I will. I will. I said art. And now like, if you do it now, you're not actually going to get low. Like I can see it coming like this is going to be like, This is a 50 If you don't do something, and she's like, she's like, I will I will. I will I'm not hungry. I don't want to have a juice. No, do this. No, no, get out of my room. I'll take care of it. I'm like, okay, so I mean, you know, I walked out of her room and sat like up the hall waiting because her blood sugar was, you know that I saw it go the other way. And I was like, okay, she got it. But I know I know that thing. Like you lose your you lose your cognitive ability. You just, you know, and you think you're making a ton of sense to, which is Yeah, it's really interesting. And your friends loved it. Right? Because you turn it into a goofball. Right, exactly. Yeah. You're lucky they didn't figure out how to give you insulin just to make you a low just to pass the time. Someone give Edward a unit. So he says something stupid. She's and your girlfriend who becomes your wife afterwards? She saw all that in college and married you. Wow.
Edward Robinson 33:50
She did. She did. And my mom was even like, you know, he's got diabetes. It's gonna be a lot of work. Are you sure you want to marry him? And
Scott Benner 33:58
Oh, thanks, Mom. You want things? It's easy to get? There is it was a totally fair thing to say. Well, yeah, but someone got to be on your team. I pulled her aside. I've been like, Listen, I have no other prospects. This is not something that wasn't okay. Well, she comes
Edward Robinson 34:18
from a family of allergies. And so she had allergies to milk and eggs growing up. Like if milk touched her skin, she'd start throwing up. Oh, and her sister has like a peanut allergy. And our kids have allergies. Some of them have been they've outgrown some of them. But between like diabetes and allergies, our kids potentially hit like the Yeah. The genetics jackpot. Yeah, well,
Scott Benner 34:50
I think of it's funny. I don't know how science thinks of it. And I'm sure we could Google it and find out but i i I think of allergies as autoimmune Like you don't I mean, like it's your body, right having a ramped up response to something. So that's something else. Geez, yeah, forget college, you better start saving for medical insurance and stuff. Wait a minute. Like I, if I flick milk, get your wife and it hits her she vomits
Edward Robinson 35:22
Not anymore. Not anymore. But when she was very young, she did. Yeah. So, so there's one time when they they wanted what was it? They had to go somewhere. And they had to like change their flights or something. And they're like, sorry, you can't change your flights unless you've got a doctor and get like a refund unless you've got a doctor's note. They're like, Huh. And they gave her a little bit of milk and took her to the doctor so that they could get their flights changed and get a refund.
Scott Benner 35:57
Okay, hold on. So her family made her vomit, to get a doctor's note to get a refund on an airline ticket. Yeah, like I like this.
Edward Robinson 36:06
So it's not a for her. It's not a paid dangerous overreaction to just throw up and take Benadryl and then she'd be fine. It's not like a peanut allergy, where you have like respiratory issues and can't breathe.
Scott Benner 36:22
They explain this to her. Was she a co conspirator in this thing?
Edward Robinson 36:27
Um, maybe
Scott Benner 36:30
I don't know what the thing I don't know if I like the story better if they just are like, drink this thing in front of the doctor. Or if they were if they pulled her aside like a little like thug, and they were like, here's what we're gonna do.
Edward Robinson 36:42
The best part was that it was in Canada. So seeing the doctor is free too. So they didn't have to pay for that either.
Scott Benner 36:46
Oh, my goodness. Are you Canadian?
Edward Robinson 36:50
She is Canadian. He is your man. Her father is Canadian.
Scott Benner 36:53
Okay. Yeah. That's something where do you live? You don't live?
Edward Robinson 36:58
No, no, we live in Pennsylvania. Okay. outside of Philadelphia. Oh,
Scott Benner 37:03
we're very close to each other.
Edward Robinson 37:05
We are we are when I signed up for this. I was like, hi. I wonder if I should just go to Scott's house and you'd have a nice microphone. We could record it in person.
Scott Benner 37:12
I do have one right here. We could have done. I've never done I've only ever sat in a room with Arden and made a podcast. The only person I've ever been right in front of oh, we're done. We're gonna have to find out where you are? Because it's possible. Yeah, we're up the street from each other almost. Okay, so you've got diabetes, they start you on a pump? Not particularly well, a lot of lows. Are you going to the doctor with any regularity through college getting a one CS or anything like that?
Edward Robinson 37:41
No, the the pediatric type back office where I was going, they stopped writing the prescriptions in like, sophomore year or something. They're like, you haven't been here in two years. We can't keep writing new prescriptions. So then I went to see a doctor once or twice, so not very often.
Scott Benner 38:02
Right? So you were just you just went back to another doctor to get a different prescription? Wait, yeah. Or do you? I was gonna say when did you start seeing a doctor regularly, but maybe you don't.
Edward Robinson 38:15
So it was after college? I did. And for me, the promise was always in technology. I started on the Medtronic 670 G. Or maybe it's 630 G the one that the first one that had the automated mode. Okay. And so I went I went back I actually looked it up my A onesies were in like the sevens. And then with the 630 G, it brought my A onesies down to 6.1 6.2. So that was really amazing. For me, I found special ways to make it work. And I even found that in the settings, if you changed the insulin activity to be you could turn it down all the way to like two hours or something like that, which is obviously not how long insulin is in your body for but by doing that it made the algorithm more aggressive, because it would see that you have almost no insulin left on board and therefore it's going to give you more to bring your blood sugar down faster.
Scott Benner 39:25
So did that pump have a glucose monitor attached to it at that point? It did, right. Yeah,
Edward Robinson 39:30
so that's the one with the integrated glucose. There's one before that had a glucose monitor. But it wasn't very good. And it had this like harpoon to get it into you and it was super uncomfortable. And it was like probably half an hour to 45 minutes delayed in what your blood sugar actually was. And so I didn't really find it helpful at all. The endocrinologist wanted me to wear it. So I'd wear it for like a week before going to see them which in hindsight now I see how it could show you the trends and things like that to help you better adjust your basil, but were you focused in settings, but
Scott Benner 40:08
we're focused then on that it was such the data was so late from when it was measured that you're just like, well, how is this valuable to me, but you could have macro to it and seen? Seeing Right, exactly. Yeah, I understand they didn't even I don't even think back then the doctors probably knew enough about it to even give you that insight to you know, and this is when, what's the year for that? Do you know about?
Edward Robinson 40:33
So I started on the 6/3, the one with auto mode in 2016. Okay, so the previous version would have been 2012 to 2016. Gotcha. All right.
Scott Benner 40:47
It's so I have this problem all the time. I look at 2006 written in front of me when you're diagnosed, and it feels like it was I don't think of it as long ago. I don't know. It's just my age, like anything with a two in front of us like, yeah, that just happened. You know, and then we start telling the story, and I keep thinking it's it's 2022, you know? Yeah. Just, I mean, I don't know, there's something about those numbers. They mess with me. So you're doing this, but what you really like about it, and 39 minutes into this, Edward? It's why you're on the show. I mean, they're great at this are really bad at it. I don't know, I can't even tell the difference anymore. But but the technology part you said really attracted you to it. So when did you like you built your own algorithm? Right? Correct back. So yeah, tell me about it.
Edward Robinson 41:38
So yeah, so I was on the 630 G. And actually before being on the 630 G, the prior one with like the harpoon CGM, that wasn't very accurate. I, I found online somebody who had managed to kind of hack into the pump and extract the data. And so I worked with him and was actually able to decode the data for the CGM and load the CGM data onto a computer. And so I initially tried to create my own closed loop system. Using that, and I did it like on a laptop computer, and then I was going to put it onto like a small Raspberry Pi computer that could fit in my pocket. But between like that computer and a battery for that computer, and the pump and the transmitter to communicate with the pump, I would have to carry like a fanny pack around. And I was like, Yeah, I'm not gonna do this.
Scott Benner 42:43
Even good health,
Edward Robinson 42:44
I'll just wait, I'll just wait for Medtronic to release the 630 G and use that right. But after using it, I kind of got frustrated with it. Because it being the first one, it was very conservative. And it would bring your blood sugar down to like 150 and then slowly bring it down to like 120 but it wouldn't go. It would try to keep you above 120 all the time, and I want to start getting better control. And so when I was done with that I was looking around to see what to switch to next. And that was when I came across the loop community and the Android APS app and started using that and since then, so when I started using that my a one C dropped from
Scott Benner 43:45
I'm going to tell you about one of the better decisions I made last year, I switched Arden's delivery of her diabetes supplies from where we were getting them to us Med and US med is more than edging out the service that we were getting from that previous company. right from the comfort of your home or office, you can join over 1 million satisfied customers who rely on us med for courteous, knowledgeable and trained customer care and their representatives are going to keep you up to date with your medical and diabetic supplies. All delivered right to your door. Us med.com forward slash juice box or call 888-721-1514 To get your free benefits check right now. US med features a litany of things that you're going to love. How about an A plus rating with the Better Business Bureau. They accept Medicare nationwide and over 800 private insurers they carry everything from insulin pumps and diabetes testing supplies to the latest CGM is like the FreeStyle Libre three, the Dexcom G six and a little bird told me the Dexcom g7 coming very soon. They always provide you with 90 days worth of supplies, and fast and free shipping. better service and better care is what you're going to get when you You go to us med.com forward slash juice box. On top of all of this US med is now dispensing Novolog insulin aspart and human log insulin lispro through their pharmacy benefits. What are you waiting for us? med.com forward slash juice box 8887 to 1151 for us med is the number one distributor for FreeStyle Libre systems nationwide. They are the number one specialty distributor for Omni pod dash. They are the place we got our hands on the pod fives from us med provides Arden with her Dexcom supplies. And they're the number one fastest growing tanam distributor nationwide. I mean, I guess I could say it again. But are you just already online? Getting it done? Are you even listening to me anymore? Or have you already called 888-721-1514? Don't like the phone? Us? med.com forward slash juicebox. The other day, I got an email from us Med and it said are you You want some more supplies? I guess it was time. And I said yes. Click the button. And then they just showed up. You want to do it like that. It's pretty damn easy. Us med.com forward slash juice box.
Edward Robinson 46:18
6.1 to 5.6, right, which was really amazing for me. But the best part was really the alerts and alarms. So for me, I'm a very heavy sleeper. And I would like the alarms would be going off on my pump. And it would be beeping away and I'd be sleeping through it. And my wife would kick me because it was disturbing her. And I would wake up in like a half, half asleep and I would just sort of dismiss the alarm go back to sleep. And 20 minutes later she'd kick me again. And I'd wake up and dismiss it and go back to sleep. And sometimes, and then eventually I would get up and yeah, have some sugar or things like that. But in my like, half asleep state I would sometimes do the wrong thing. So there are even times when I would give myself insulin when I was low. And my wife would look at me and she'd be like, What on earth are you doing? You just gave yourself insulin and you have low blood sugar. And like, after like a few minutes of talking to me, I would then be like really cognizant and
Scott Benner 47:26
having a meal, I need to
Edward Robinson 47:30
realize that I would need to go and chug a bunch of orange juice because I just given myself three units of insulin right when I already had low blood sugar.
Scott Benner 47:37
You know, there's a story floating around on the internet right now of a gentleman that just did that and killed himself. So it's such a crazy tightrope act, you know?
Edward Robinson 47:49
Yeah, it's so it's like Android APS, I love it because I set up alerts were ill send off an alarm and the alarm will go off every five minutes. If my blood sugar is dropping below, below 75 Or if it's dropping below 80 and dropping quickly at night. And so the the first one I usually just dismiss and go back to sleep, but by the third or fourth alarm, I've woken up and had something to eat and and then that saves me a lot from having nighttime lows. I have fewer nighttime lows, but they still happen occasionally.
Scott Benner 48:27
You keep stuff by your bedside. Yeah, looks like a little Kimmy. It's sort of a Wawa, I guess. Finally I can say yeah,
Edward Robinson 48:35
and then sometimes I come down to the kitchen in the morning and I've just like eaten everything. There'll be like candy wrappers and gummy wrappers all over the place and my blood sugar will still be coming down from like, 200 I'll be like, Oh, God. That was horrible. I
Scott Benner 48:50
overdid it. Yeah. What's that feel like in the morning after you've been through that? Is it a? Like I know people say like lows overnight, make them feel hungover. What a highs overnight. Make you feel like?
Edward Robinson 49:03
Yeah, I think it's kind of similar. I feel like dehydrated and hungover. And then having young kids there's usually one of the kids up at night at least once during the night. So then you just feel even more tired, right?
Scott Benner 49:16
Do you ever feel achy? Is that ever a thing that happens after highs just body aches or
Edward Robinson 49:22
so I've started to notice that sometimes my like tendons and stuff hurt a little bit, or they're not hurt, but they're just like more sensitive. And I can feel that, like in my hands or my ankles. They're like tighter. And stuff, especially when I have high blood sugar or extended high blood sugars.
Scott Benner 49:48
It's interesting because Arden's ribs get sore sometimes. And, you know, lace cartilage and connective tissue and stuff like that. It's interesting. Wow. Oh, okay, so you built is is that? Um, Dana Lewis? Who, who I had on the show years ago about Android APS like, is that a name you recognize?
Edward Robinson 50:10
Yes, she's one of the contributors. And there are a bunch of other contributors to Android APS as well. A lot of them are abroad.
Scott Benner 50:22
Okay, what made you go Android? It was available prior to loop. Right. Android was first Android. APS was before loop. Am I right?
Edward Robinson 50:30
I don't know which one was first one, either.
Scott Benner 50:33
You're just an Android person.
Edward Robinson 50:36
But yeah, just an Android person. Yeah. So I run with that. And so then, for that, it's, it's really interesting, it's got this really cool feature where when you start using it, it doesn't unlock all the features until you've gone through certain milestones. So it starts you out in manual mode. And you have to do that for like, a couple of weeks. And you need to answer some like, kind of test questions to make sure that you're, that you understand how it's working. And then it will do a mode where it will prompt you whenever it wants to change your Basal rate. And you do that for a few weeks and learn how that works. And then it starts to do that automatically. And then a few weeks after that, you can learn how to use the, what they call small micro boluses, which is basically where I'll give you bonuses rather than adjusting your Basal rate.
Scott Benner 51:35
I think that's, by the way, I think that's brilliant. If you listen to the show, I you've heard that I've advocated it to companies all the time, and no one seems Yeah, no one seems to listen to me. But for all different kinds of applications, I keep saying to them, like why not have a beginner an intermediate and a an expert level that you can unlock by? I mean, that's a great way even having your doctor sign off on you sliding up some things. Like imagine if you could, I mean, you know, right now what only pod fives target is one tend to control like US targets 112 and a half, like what if you could unlock a blower target or something like that?
Edward Robinson 52:16
Yeah. And you can even do even more stuff. Like, if you have lows, or if you have an extreme low, then you know, it's going to kick you out of that lower target for 24 hours to make sure that you don't get another extreme low. Things like that. Some of it is like, you know, the first couple of hours before any of the Dexcom. Sometimes it can trend lower than it really is like it says it's 46. But you test your blood sugar and your 86. But eventually, it sort of catches up, which some of that could be frustrating. But I think there's so many things that you can do. And there's tons of stuff like that built into Android APS, which I really love.
Scott Benner 52:55
I never know how to answer people's questions, because inevitably, someone says on these automated systems. What if my CGM is way off? What do I do? And I don't know how to say like, I don't know. It's never just not usually an issue. And, and when it is like you just like Arden switch to CGM day and a half ago. And for the first 24 hours, it was reading lower than she was. But her blood sugar's were higher. So she had like a like an impact from from hormones. So she was in this more like 170 range. And then I don't I don't know how to put this other than she was too steady, too high for too long. And it made me think, Why is her blood sugar not coming down with the amount of insulin we gave her? Oh, her blood sugar must actually be higher than this says. And there's something about that collection of circumstances that make me go test your blood sugar, you're higher than this says you are. And then as soon as I prove that that's correct, then you can make a large enough Bolus to actually impact it, and then boom, you're okay. But yeah, I don't know how you explain that to somebody on day one, you know?
Edward Robinson 54:05
Yeah, yeah. Like some usually when I first start the sensors, I'll just stay in manual mode, and just let my regular Basal rate run until I know that the sensor is actually working effectively. Sometimes I put it on it works effectively, immediately. Sometimes I'll do the pre soaking where you put it on, and you stick in like an old dead transmitter and then swap the transmitter in. So then it's been in underneath your skin for 12 or 24 hours before it's actually being used. Yeah.
Scott Benner 54:38
You get more accuracy from it right away. What right what pumps do does Android APS support?
Edward Robinson 54:46
So Android APS supports a bunch of pumps. Some like European pumps, and I think even like some from South Korea, as well as Medtronic, and I Omni pod, the original Omni pod as well as Omni pod dash, and I'm using Omni pod dash, which is really fantastic. Because there's nothing to carry around other than your phone which, yeah, yeah, like most people these days, I'm glued to my phone anyway. Right? My, my, it's really nice to not have anything to carry around.
Scott Benner 55:19
It's my expectation that that's the system we're going to put on. When as soon as loop is ready, there's a loop, I think loops, Badaling on the pod dash right now. And it's, I think it's very close to being done. So
Edward Robinson 55:34
yeah, I'm just, it's really great. So then, so then after doing that, for a while I, I did some contributions to it. So I tested that the I was one of the early testers for the dash capability. And then I was one of the early testers for this new algorithm that people have been working on in the Android APS community, that's called dynamic ISF, where it looks at your total daily dose from the last seven days, the last day, the last, like four hours and the last four hours before that, in order to estimate what your insulin sensitivity should be. Which is pretty cool. And there's actually some really cool science behind it. Where, which is why a lot of the pump manufacturers use it. Where if you've given yourself more insulin, over the last few days, or whatever you have, chemically within your cells, like more resistance built up. I don't know exactly how the chem chemistry works. But it's kind of like people say, when you drink more, you're
Scott Benner 56:50
gonna get drunk as fast it takes
Edward Robinson 56:52
Yeah, exactly. I thought that's a kind of similar concept with insulin, right? Which is pretty cool. So I tested out that and over the last four to six months, I've actually taken my experiences from work of working on artificial Mattel, artificial intelligence and machine learning. And built my own algorithm that uses AI and ML to build a model. And it uses a whole bunch of data, so uses those total daily dose, insulin factors, it uses the time of day, it uses activity tracked from my watching from so I wear a smartwatch, that tracks the number of steps that I take. And from the phone, and I'm actually working on incorporating heart rate monitoring. So from the watch, it monitors my heart rate and sends that to my phone. So I can include that in the algorithm as well. Well, that's as easy as
Scott Benner 57:54
well, let's stop for there for a second. So having just a step tracker on, then that allows a sort of like not unknowable variable, variable, but something that it's hard to remember to. I mean, you know, when you're, when you're making, when you're making manual decisions about your insulin, it's hard to think, oh, I walked around more today than I did yesterday, I'll cut this back, or maybe my insulin sensitivity should be a little weak or something like that. But you're saying, Yeah, you're saying that the algorithm would be able to take that into account?
Edward Robinson 58:28
So it does, yeah. So basically, what you do is you feed it, instead of writing out for machine learning, instead of writing out an algorithm, you give it a bunch of data, and you give it the answers. So you say, you know, my blood sugar went up half an hour later, I should have done more insulin half an hour earlier. And you give it all the factors and then it will figure out what the algorithm should be specifically for you. So it's really amazing. And it's just a total change in perception and way of thinking about it.
Scott Benner 59:03
Well, I definitely see the the retail space is moving that way because Omnipod five is is you know, there they won't even tell you what it's doing. So you know, it's like it's learning and I'm like okay, what does that mean it and and if you ask them directly there's there's no hiding it. They'll say this is proprietary information. And in so my expectation is I I've always thought that like I was like, well, if they're working on it, like the DIY community has got to be doing it as well. I just don't have a real in with Android APS to to hear the conversations. But yeah,
Edward Robinson 59:43
so this is something that that I've been working on independently. I've worked with a couple of other people. But it's still kind of early, and then my thought was to then find other people who are interested in experimenting with it. and help roll it out to more people and then eventually make it easy and accessible for everybody in the Android API space. There is actually a French company called Diablo loop. I don't know if I'm pronouncing it correctly. But they actually use AI and ML for their algorithm on your, which is pretty cool. It's
Scott Benner 1:00:23
exciting, actually is what it is. Yeah, yeah. Because but it's I guess, and it's, so you're you, are you using it right now? For yourself? You are? Where's it holding your
Edward Robinson 1:00:37
server right now my blood sugar is 129. It the target is 90. And it can keep it pretty stable there. But the so it's pretty cool. So it uses what else does it use? It uses the insulin on board, the the amount of carbs that I've eaten, I've recently been working on adding in the type of carbohydrates. So recording. Yeah, have I eaten pizza or salad, or, you know, orange juice, and then being able to incorporate that into the algorithm, because the algorithm should be able to learn that, hey, you ate pizza four hours ago, I'm gonna need to increase your insulin in order to account for the fat price that's going to come later on. And things like that. And even like specific types of pizza, have you eaten pizza from Domino's? Or have you eaten homemade pizza, and be able to factor those things in because it can basically see, you know, in the past nine out of 10 times when you've had Domino's Pizza, you've had arise for hours later. But when you've had homemade pizza, nine out of 10 times you've had arrived two hours later, because there's less oil and stuff
Scott Benner 1:01:56
like yeah, it might be less significant as well. Because I mean, right we make we do it here and we that's exactly what we see to the crust is thinner. I'm not dousing it in oil, there's not a ton of salt on it, like you know, it's there's not as it's not covered in cheese, there's cheese here in there. That's um, but that's exciting, isn't it? Well, we gotta get you help them. And we're we got to keep you moving. And your wife needs to be nice, no more kicking you at night, you need to be treated better. What's the sort of is the idea of like, Do you really think you can get it to a place where you'd be comfortable sharing it with other people like that?
Edward Robinson 1:02:35
Yeah, definitely. So then, so then it. So I take all the data. And then I train and build a model on my computer. And then I load that model onto the phone, and then use that model. I built in a bunch of safety requirements. As well, like if my blood sugar's dropping, it's going to be restricted and how much it can give me. You can put in restrictions around how much insulin it can give at any, for each micro Bolus. There's thresholds around how much it can give in total, like how much insulin you can have on boarded anytime. So there's a bunch of like safety around it. But all of those safety requirements are built around what I need for safety, not necessarily around what any person would diabetes needs for safety. You know, I'm pretty, relatively easy. I'm not a young child. I don't go to school, I work from home. I carry glucose tablets with me. Just in my pocket. I've got a little thing and I carry two glucose tablets with me all the time. Do you remember safety isn't really too bad for for me, but it shouldn't be that hard to make it safe for other people as well.
Scott Benner 1:03:59
That's what I was wondering is like, Can you can you imagine a world where the thing you're working on could be that what's How do I want to put this where it can be where it can do the job that is doing for you, but be generic enough that it could just be a retail item?
Edward Robinson 1:04:21
Yeah, so I think that this is going to be that the way that it will be for pump manufacturers, I think that they'll probably have another couple of iterations of doing it manually, like writing the algorithms manually, but then eventually, over the next maybe five or 10 years out, it will be using machine learning. And actually, I've you know stocked Omnipod pod and Dexcom online and they actually hire a lot of like data scientists and machine learning people. So they're using a lot of this but then they're translated and sort of what they're learning and what they're seeing in the data into still handwritten algorithms that then use inputs, like, I know that Omni pod uses, like total daily dose of insulin in their algorithm. And they probably use, like, I know, they use total daily dose from last three days, they probably use it from the last day in the last few hours or things like that, to figure out how much insulin to give you, but they don't factor in things like, you know, exercise, hormones, stress, food, types of food, things like that, which you can really get, and a machine learning algorithm can do and use for figuring out how much insulin to give people
Scott Benner 1:05:47
ever do this for me, tell me something that's that exists in the world today that people would be aware of that is employing machine learning.
Edward Robinson 1:05:59
Oh, Tesla cars, that's one of the simplest ones. So like, you know, Tesla's got their autopilot, and self driving car features. So that's all around machine learning. So they take in basically tons and tons of video data. And then the answer is, you've got to stay on the road. And times when the car drove outside the road, from the audit from the AI and machine learning system, but the person steered back into the road. Those are areas where you need to adjust the machine learning algorithm and provide it more data and say you should have turned right, when you tried to turn left to stay in the road. And, and therefore learn from this experience and build that back into the model.
Scott Benner 1:06:44
So in there, in that example, and I don't know why I know this, but I know that Tesla had to actually build their own computers to do that process, because they couldn't buy powerful enough computers to do it. Right. Yeah. And they're getting data from every I guess every car that checks off. Yeah, you can have my you know, you can have my cars data. And right, so So these computers are just running these simulations over and over. And then from what it sees, the computer says, Alright, well, line whatever in the code should be this instead of this, because, or what I know, I'm oversimplifying greatly, but it's learning and making adjustments to the, to the software that's onboard in the cars.
Edward Robinson 1:07:32
Yeah. So like, another way to think about it is. So when you write a traditional program, you say, if your blood sugar is high, if your blood sugar is x, then you need to bring it down. 50 points, right. So for 50 points, if your total daily dose was 100, then your insulin sensitivity is going to be 50. So you need one unit to bring you down 50 points, we're gonna give you one unit, right? So you basically write all that out to say if this than that, and if something else, then do something else. And you write out that logic by hand. But the machine learning basically figures out what that logic should be. So you can actually have a generate out this machine learning algorithm approach called Random Forests, where it will write out those sort of if then statements for you, and it builds kind of like a little tree that you could imagine that at the base, you say, is your blood sugar high or low? Okay? If it's high, then do this or that. And if it's low, then do X or Y. And l build out like hundreds of these trees, and then each tree will make a prediction of what it thinks it should do. And then it will take an average to figure out how much insulin to give you.
Scott Benner 1:08:53
Okay, so it could in that scenario, it could say, Listen, normally, I need we want to move 50 points, that's a unit, but I see that you've walked more today than normal. So we're going to use point 100.8 instead, something like that, or, or over the last 24 hours, we've been using 30% more insulin than we expect to so I'm gonna give you a 1.3 in this situation. Yeah,
Edward Robinson 1:09:19
yeah. So then, one way to do it,
Scott Benner 1:09:23
you understand? That puts, as I said, that puts me out of business. That's, that's for sure.
Edward Robinson 1:09:31
So I've been using an algorithm that it's called neural networks. And so it's kind of similar, but think of it more around probabilities. So you know, when you're high, you probably need one unit nine out of 10 times, but you've exercised so when you've exercised, nine out of 10 times you need half of the amount of insulin. But I can also see that you've eaten Oops, 60 grams of carbs. So when you've eaten 60 grams of carbs usually need, you know, a little bit more. And so then it adds all those pieces up to then figure out how much insulin to give you,
Scott Benner 1:10:10
right? It looks at the carbs, then it looks at the activity. And here's your number. And then this happens. In a blink of an eye, like you don't notice it happening, so yeah,
Edward Robinson 1:10:24
yeah, so it makes predictions every five minutes when it gets the reading from the Dexcom. And then it takes in all those factors, and then calculates what it should give you. So right now, let's see. For example, it says, so this is my, my current blood sugar is 136. And it's going up plus three. And so it says, the AI model predicted a small micro Bolus of point 207 units. And then it goes in and factors in a bunch of safety requirements. And then it rounds it to point 05. So it's gonna give me point two units of insulin. Okay. And so it did that.
Scott Benner 1:11:12
That just happens automatically. It's not done through basil. It's done in an auto Bolus.
Edward Robinson 1:11:17
Yeah. So then every, everything's done through an auto Bolus. Yeah. And and so then it sends the communication with the pump, and then gives me the amount of insulin that it that it should do.
Scott Benner 1:11:29
Yeah, I'll tell you for Arden loop made a big leap for her when we left, the version that was just using basil to try to stop and it moved to the auto Bolus version that was a was a big deal for her. There are still some times very infrequently, but it can happen, where if she has a low, and she puts in some carbs to fix it, but uses too many carbs and then hits her threshold again, it boluses and you're like, No, please don't Bolus here. You don't I mean, and then and then she's low again, 30 minutes later. And then yeah, I've showed it to her, I'm like, you just you can't, you can't over treat it. Because you're also not telling it you took in the carbs. And that's weird. So
Edward Robinson 1:12:15
there's some cool things that you can do. I used to run into that a lot. And so with Android APS, you can set up automations. And so one of them that I had set up was anytime that I went low, I would set a temporary target of 130 for the following hour, in order to avoid having a second low by overcorrecting because you eat fast acting carbs, and it thinks your blood sugar is going to shoot up to 200. But then the carbs are gone. And then there's too much insulin right on and then it crashes down again. And you got sort of a yo yoing effect. So
Scott Benner 1:12:47
you just did a temporary target that was higher.
Edward Robinson 1:12:51
Right, right to say, you know, let's avoid doing a second low, which then I'm going to treat and over treat. Let's just keep it steady for a little bit, and then bring it down slower.
Scott Benner 1:13:01
You know, the woman I spoke to yesterday, who I told you a 72 She's looping, which I'm like, just I thought was terrific. And she said when she goes and does things like meets new people, she gets a little adrenaline. And she's like, I don't like the Bolus for instead, I set my target lower. So the Yeah, so then the algorithms a little more aggressive, not a ton aggressive. And then the bump she gets wasn't as you know, isn't as as significant. And she doesn't end up low later.
Edward Robinson 1:13:31
It's really Yeah, the other thing that Android APS can do, which is really cool, is that it can see based on how many carbs you put in and how much it thinks it's seen of those carbs. It can say, you know, you're likely to go low in the next 45 minutes, and you should eat, you know, 15 carbs, or 10 carbs are eight carbs, which then really helps with overcorrecting, and it can see it far enough out that you can have it to then totally avoid the low. That's which is really nice.
Scott Benner 1:14:06
It pops up and tells you like you should have eight carbs. Yeah.
Edward Robinson 1:14:10
Right. Right, exactly. And then the other thing that I really like is so I use another app called X strip, which works with Android APS, but it allows you a lot finer control over the the alerts and alarms. And so I always find it funny when I hear you ask on the podcast, what people have their alarm set to because what you can do is you can set different alarms. So at like 130 I just get like a nudge that's like, Hey, your blood sugar is going up a little bit. At 160 It's a little bit more of a poke. And then at like 190 My phone starts vibrating like crazy. It's like hey, your blood sugar is really going up. You should do something about this and and then also when the alert goes it starts with vibrate. But if you ignore the vibrate for a few minutes, then it starts with like a small chime. And then it gets louder and louder and louder for the alert to make sure that you actually acknowledge it. Which is really cool.
Scott Benner 1:15:15
Is this a job interview? Or did you think I'm gonna come on this podcast, I can get a job with one of these. Somebody over here, because they're going to hear it. At this point, I think every diabetes company listens to the podcast. So I feel like I feel like you're on an extended job interview right now you don't even know. Like, why? So I'm fascinated by this part of it. And maybe you have insight in this. And maybe I'll just ramble and you'll say, I don't know, Scott. But when all this is possible, right now. I don't understand why. You know what I mean? Like, I was in a meeting with somebody one time, and I said, like, I don't know who Ivan is, but find him and hire him. You don't mean and? And, you know, I don't know who Edward is. But is there a reason somebody hasn't run in and gotten him and put them in a room and said here, do the best you can? Let's see what you come up with. You know, like, I don't, I don't know why. I mean, I guess big companies have more to consider than I'm, then I'm absorbing right through the FDA and all this other stuff. But it's just I mean, what you're talking about sounds like the future, but you're sitting in front of me using it. So
Edward Robinson 1:16:22
yeah, yeah. And a lot of it is just like common sense. Like, it gives you an alert, if it thinks you're gonna go low within the next 15 minutes. So then, when that comes up, I just take, you know, two glucose tablets. And then I avoid, though entirely, rather than getting the alert when you go low. And from what I've listened to of the interviews that you and other people have done with, like Kevin Sayer from Dexcom. When he answers the questions around the alerts in the apps, it does sound like Dexcom is going to build some of those things into the next version of the CGM. And I really do hope that they build in a lot of these types of alerts because it just makes it so much easier. Rather than getting the Dexcom alert, that's like, incredibly loud, when your blood sugar's 130. Just getting a little nudge is so much more convenient and effective.
Scott Benner 1:17:21
Yeah, it also doesn't give you that like, feel anxiety. Yeah, she's this again. And then you look down, you're like, Oh, I'm only 130. And you know, it's not right, I take your point, I just I'm not even pointing out anybody specific. I mean, pumps, pump manufacturers, CGM, or whoever. It just, it's all right here. I don't know what you don't I mean, so
Edward Robinson 1:17:44
we don't Yeah, and the funny part is that the code is all written, where it's, it's open source, I mean, you can just take the source code and just have that be the Dexcom. App, you don't have to do anything, you can just take it and use it.
Scott Benner 1:17:58
Right. Even that part of the world, I don't understand I for the life of me don't understand, like open source. Like I think it's wonderful. But like, my brain doesn't understand why people do it. Because it doesn't, it doesn't lead to I guess it does. It just leads to it. So slowly, getting to mean like, like, Wouldn't it be better if you just sat down and wrote out your thing and incorporated it and made it? And then? Or does that like the as soon as you get into regulatory stuff, it just it stops that I
Edward Robinson 1:18:29
get? I think yeah, because I think that the regulatory stuff, and then also making it safe and effective for everybody. So like, I can build this algorithm and test it on myself. But if I'm doing it in a company, and you know, I get low blood sugar, I just treat it whatever, I don't, I don't care. But if a company gives somebody low blood sugar, then it's like, well, what are you guys doing? This is terrible, right? And if it falls apart, and somebody dies, you know, then the company is dead. It's over for them. I completely understand they have to be extra safe and cautious when working on all these things.
Scott Benner 1:19:09
I'm being to people here. I'm being like, 10 year old man who's like, Just do it. And you know, the rest of me is like, No, I know why this isn't happening. But it's just it's, it's frustrating. You know, it's and it's, and yet, in the diabetes space over the last eight years, things have moved forward. Incredibly dramatically. Yeah, you know, just really it's
Edward Robinson 1:19:31
amazing. Having lived through like multiple daily injections and eating a certain amount of carbs at a certain time to now having this like incredible algorithm. Yeah. Which so the other part that's really cool about the algorithm is that there are no settings in it. There's there's no insulin sensitivity factor. There's no Basal rate. There's no Are count, there's no car sensitivity factor, it just figures it all out for you. And all I do is I put in the amount of carbs that I'm eating 20 minutes before, I mean, you still have to like, put it in before so it can Pre-Bolus. But 20 minutes or half an hour before I eat, I put in what I'm eating, and the amount of carbs and then it slowly builds up like a wave of insulin as a Pre-Bolus, which is really nice. Yeah. And then you also get fewer like crashes because it's slowly building up the insulin through multiple doses, rather than just like one dose that then so I use fiasco and I would find that if I Pre-Bolus. With the right amount of insulin, it would all hit within like a 15 minute period, half an hour, 45 minutes later. But the food absorbs over a longer period of time. So being able to have it, slowly build up the wave of insulin makes it a lot more effective.
Scott Benner 1:21:05
Arden had a five guys the other day, and we went and picked it up. And we left there. I was like, I think I think I said something like Bolus 20 carbs now. And then we drove home. And then we did 40 More like I'm telling you this meal is right. It's got to be 100 at least carbs. And then we did a little more and did exactly what you just said, like we created. You know, in my you it sounds like you do listen to the podcasts like in my mind, like I'm creating like different timelines of Bolus is different. You know, they live on their own plane of existence for me, this one's going to peak here, this is going to peak here, this is going to peak here. And that should lay over top of this fat and protein that she's going to see from this food pretty well. That I'll tell you that Bolus went so well that when it was over, I was stunned by it. Like Like, like, like four hours later, I looked at her graph, and I was like, huh, wow. Like, even I was like, okay, yeah, that's crazy that that works. And then the next day. She has a waffle in the morning. And we did a little bit. And then I said I was cooking and I said, Hey, Bolus 30 more. And then I don't know, 15 minutes later, she sat down, and I said, Okay, we're gonna do the rest of the insolence she goes, we haven't done any of it yet. Just the first bit. And I was like, No, I said, Do 30 more. And she goes, I didn't do that. And just not hearing me say that. kept her blood sugar 180 for like, four hours. I appreciate the waffle. Just just missing it in that spot. You know what I mean? Was it's really something. But yeah, obviously, I, I want this because that I mean, what you're describing does what I do. And I think there's other things in the world, I would do, Edward if I didn't have to think about this. So sounds pretty cool.
Edward Robinson 1:22:50
Yeah, yeah. And it's amazing. So it really, machine learning works well, when you have a lot of really precise data. And there's, there's just so much precise data for diabetes, like the amount of insulin that a pump gives you. It's just mind blowing how precise it is, I think like, point 05 units that the Omni pod can give you. I think you if you take like a drop of water, and slice it up into like 1000 pieces that that's the size of one of those bonuses. So it's like incredibly precise, it's just mind blowing. And then, you know, people are pretty good at counting carbs. So that's pretty accurate. Be able to track your exercise and activity and heart rate and be able to incorporate that one of the things that I'm really excited for potentially doing this with other people is to be able to factor in things like location. So you've got the GPS on your phone, it knows if you're at school, if you're at work if you're at home. I mean, I work from home, so not really that important for me. But you know, being able to factor in like hormones and menstrual cycle, you can easily just put into the app, like hey, today's the first day of the menstrual menstrual cycle, and then from that it can figure out you know, throughout the month, how much more or less sensitive it needs to be based on that.
Scott Benner 1:24:19
Yeah, I'll tell you that's what we just went through that last week where through the event, Arden didn't need very much insulin. And then when the event ended for like four days her needs were significant. And Rachel and I, we track it in a in a period tracker, but that's just enough for you to go Okay, looks like the four days after need more edits. You know, for people who think like all this stuff is just magic like Arden's insulin sensitivity and needed to be changed for Bezos needs to be changed, like a lot needed to be changed in those four days and then when it ends Edward, it's fascinating. Like, I watched it end, at four o'clock in the morning on the fourth day, it was just it. It was like someone walked in a room and was like, Oh, who left this switch on? Click, she does not need these settings any longer. And boom, we had put them all back. Stop. Hello. And then Yep, good, was fascinating.
Edward Robinson 1:25:21
And that was that was one of my frustrations with with the other Android APS is that algorithms is that on, on some days, I skipped breakfast and do intermittent fasting, some days, I'll even skip skip lunch as well. And then you have to go into it. And you have to say, okay, reduce the profile settings, and then potentially reduce them further, if you skip lunch as well. And needing to do all that I just found to be annoying, and it should just be able to figure it out for you.
Scott Benner 1:25:52
Right. And the override for loop is I know we're jumping back and forth between systems. But the overriding loop works for her for lower days. But for higher days, it's just not enough. Like you can't like you can go 150% On the override, and it still doesn't do it. And then there's a moment when you get too high, and you're just going to cause a low. And then those are the moments when you're just like, Well, I'm just gonna, I'm just going to open the loop and run it like a regular pump, then. And then you inevitably forget that the loop is open. And you're and you're reminded three o'clock in the morning and you're like, Why is her blood sugar low? Oh, we didn't put the algorithm back on. You know, like that kind of stuff. Those things being for Android
Edward Robinson 1:26:32
APS, you can disable it for a time period, which is really awesome. You can just say, pause it for an hour or pause it for four hours. Yeah. So then you don't drop it into that?
Scott Benner 1:26:41
No, it is really well listen, it's amazing. And here's the here's the scary part to me. That it's it's all based on whether people like you find out about it, and are interested enough to put effort into it. Because this is this could literally be like a softball league, where one day we all just don't show up for practice. And then the team has gone. You know what I mean? Like like that could happen. Like I it's so far hasn't, but it could it like what happens if there could you know what I mean? Like, but
Edward Robinson 1:27:17
I mean, I think it's, I could see that happening if there is you know, but I think because people still have diabetes, it's still going to happen.
Scott Benner 1:27:26
Yeah, I hope so. I mean, I'm I'm like what you're saying is, is exciting to me that you're interested in it. Yeah, that you're doing it is fantastic that you think that other that the companies are looking at that stuff as well, is very cool. And I'll tell you why even more important. And not just because I spoke to her yesterday, but 72 year old woman using a loop. And one of the questions I had to ask her was, Are you worried about the moment when this stops making sense to you? Or that you can't keep track of how this technology works? And if it was the machine learning technology would help you later into life? Because you wouldn't need to know much of anything, you know? Yeah, like, I
Edward Robinson 1:28:10
don't know. I don't know at all how it works. I mean, I know like, you know, how it all works. But the idea is that the algorithm, it it's multiplying so many numbers that you just can't possibly read through it and understand how it works. And but the magic is that you can see that it's effective, because it's saying at the end of the algorithm, it says how accurate it is. So like right now, it says that it's off by like point 05, on average point 05 units on average. Initially, when I started the algorithm is off by like, point eight or things like that. But I've seen as I've developed further that it's gotten more and more accurate, even though I can't sort of crack it open and tell you how the algorithm really works. You can build similar ones that kind of show you that like tree structure that are readable and understandable. Which is pretty cool.
Scott Benner 1:29:09
But but that's not what most people need. What most people this is where I actually I firmly believe that Omnipod five is on to something, they're just like, look, we don't think you want to know this. You mean like we want this to work, right? You and without you having to understand what all these things are? And that's what I'm saying is they absolutely either you're a person who gets it. I mean, that's one of the reasons why looping and all this hasn't blown up more than it has like, it's exciting that, I don't know, 10,000 people have downloaded the loop app or whatever the number is, but there's like 1.8 million people with type one diabetes. So basically, none of them have loop. You know, statistically speaking, what keeps it from blowing up is that you have to understand all this stuff and that most people either aren't going to have the ability or the time or the inclination. And so if you need it if you want it to really grow, it's going to need to, like, you know, steal the line from Steve Jobs, right? It's going to need to just work. And, you know, that's it.
Edward Robinson 1:30:09
Yeah. So that's actually what I want to do, eventually, when I make it so that other people can do it. And I've actually done a lot of this for myself as well, just to make it easier. But have it just like automatically it uploads your data to the cloud, the cloud trains a new algorithm. It says, Hey, this algorithm, is this much better? Do you want to accept it? Yes or no? And then you're on your way? Or even just have it automatically upload the algorithm if it is more effective,
Scott Benner 1:30:40
right? Is there? Is there a business way, maybe you haven't looked into it, but is there a business way where, where that could be set up, and someone could earn a living from it, without it being a legal entanglement.
Edward Robinson 1:31:02
Um, it's not easy. Because as soon as you start having people give you money for it, it's kind of like a service. And then if they run into any problems, you know, you've been providing this to them, and therefore you are, you're responsible, liable. Now, it's, but in the open source, it's just I'm providing I'm setting this information out into the world. And take it or leave it up to
Scott Benner 1:31:32
you, because someone needs to, I mean, this is what I figured out for making the podcast is somehow is completely different and exactly the same. If the podcast is helpful to you. It's because I can treat it like a like a business. Like, I can make sure that there's an episode there for you when you turn on your phone, and you expect it to be there. And the only way I can do that is because I can pay my bills and live my life while making it. And the only way and the way I found around, and I didn't want to charge people, because I not only do I not want to take money from people, I very badly don't want to take money from people. But I also know that the amount of people who would be willing to give money to it probably wouldn't be enough so that my wife wouldn't make me go get a job. Right? So like so. So I thought, all right, well, I'll take advertising, but you can't take advertising until you have numbers that are big enough to support advertising. So as much as this might just seem like a podcast to you. I've had to put a number of different structures in place, and will them into existence and make them popular enough to support the next step to get you to the content that would help you. And this is the same problem. Like you don't even Yeah,
Edward Robinson 1:32:44
I think I think with the open source community, it's like, I could have never done what I've done with the machine learning stuff, and building my own algorithm. If Android apps didn't exist on its own, previously, I wouldn't have had the time and the effort to build out all the foundational pieces to then build on top of it. So it's all like just building on top of what other people have done. It my point. And that's I think the magic too.
Scott Benner 1:33:14
Yeah. And that's it is but the way to make it faster is if all those people could have drawn some sort of a salary from it and put their entire eight hours a day, like I'm sure I'm assuming this is something your hands go to bed, you know what I mean? Right,
Edward Robinson 1:33:27
so yeah, when when they go to bed, and they're not awake? Yeah, well wake up screaming, then
Scott Benner 1:33:32
think about it. Everyone's got to put his kids to bed, then he's got to look at his wife for the nod, the nods, not there. And then he's like, Alright, I guess I'll work on the injury. I want I want I wanted to get up in the morning at 8am with a cup of coffee in his hand working for eight hours, 10 hours on it. I want 510 20 people doing that. And then all the sudden, here we go. And what are we really talking? Yeah, you don't I mean? Like, how much money would you really have to generate every year to pay a small band of people to do that? It wouldn't be overwhelming. Somebody just needs. It's the situation where I don't I'm not calling anybody out. But I'm using him as an example. Like, I don't know how Helmsley doesn't hear this conversation and wrangle up 10 people and be like, here, look, how much do you make a year? Well, you work for us now. Go make this, go do this thing. I mean, they support companies when they do it. But then the minute you support the company, you get caught up in the you get caught up in the red tape. And then it's Yeah,
Edward Robinson 1:34:32
I think that there are potentially ways that they could get around it. Like if Omni pod just built the pods, and they just built the pump, but the pump and the pods, they're open to any algorithm, and then they just provide and then they work with people who build the open source algorithm. And they publish the algorithm for free and then it's easy for people to connect the dots. Then they they make their money Selling the pods, yeah. And then people use the pods because of the open source that they release. Right,
Scott Benner 1:35:04
right. And then that started to happen with tide pool as an example. But then tide pool seems like it got sucked right into the machine. And now, like, I it feels like it was two years ago where they're like, look, what we're gonna do when I'm like, it's two years later, my daughter's using three different versions a loop since then, you know, like, so it's the time that bothers me. And I knew, you know, there's gotta be a, there's gotta be a way to get around. That is all I'm saying. At least
Edward Robinson 1:35:32
Yeah. And that was kind of how, how I got here was that I thought, the 630 GE was going to come out, and it was going to be amazing. And then a few years later, they're going to have, you know, an algorithm that is just night and day, so much better. And so by the time I build something, you know, there's going to be something incredible by the time I do that, but 630 D was good. And it was really effective for me, but it's just a rate of innovation and releases just too slow. And that was how I then ended up building my own algorithm and stuff like that.
Scott Benner 1:36:11
And that, and even the idea that something might be coming could be holding back other people like you who are like, I could probably make an you know, an improvement to this. Anyway, I I'm not, listen, I'm grateful for all this stuff. It's amazing. You don't I mean, like it's without these companies. I remember what it was like, I want a lot of fun. So yeah.
Edward Robinson 1:36:34
And I've seen people's graphs on on Facebook and things from Omnipod. Five, and it, it really is fantastic. And yeah, for most people, that's maybe just good enough. And you know, it provides excellent control. And, you know, maybe that's good enough control that, you know, you're gonna run into the same long term health issues that other people do. Like, on average, people maybe get their knee replaced when there. Yeah, 75 or whatever. And if you have control through Omnipod, five, you're just as likely to get your knee replaced at 75 or things like that,
Scott Benner 1:37:16
which is good news, because it means you're alive at 75. And you need your knee replaced. Yeah, yeah. I listen, it's amazing. I love having these conversations is great. Is there anything? I mean, I'm out of my element here. So did I not bring up anything I should have?
Edward Robinson 1:37:34
Um
No, I don't think so.
Scott Benner 1:37:42
Nice. You have notes in front of you.
Edward Robinson 1:37:45
I do have some notes. I wrote down a few things. I even built an Excel chart of my agencies over the years.
Scott Benner 1:37:52
Well, what's not wasted? What have they been?
Edward Robinson 1:37:56
So I downloaded the data from from LabCorp. I, I wish I had it when I was diagnosed, but it only goes back to 2012. So it was in like, the sevens, the high sevens. And then in 2016, when I started on the 630, G, it dropped down below seven for the first time. And then it was still in the high sixes but then eventually got down to like the lower sixes. And then there's just a dramatic drop at the end. When I started using the Android APS,
Scott Benner 1:38:36
we're on to 5.6 5.6 is about where you sit.
Edward Robinson 1:38:40
Yeah. So that's where I am right now. I think that would. So that was before the last day once you that I had was before using the machine learning algorithm that I built. So I'm curious to see what it'll be with the machine learning algorithm.
Scott Benner 1:38:59
Do you have any insight through clarity or something like that where it might be?
Edward Robinson 1:39:04
So I think it's probably going to be very similar. I want it to be lower. And I think that I can kind of make it lower. Right now the big benefit of it is not having any settings and having it be more dynamic and less, less manual overrides and input.
Scott Benner 1:39:25
Well yeah, that sounds much better. Do you eat a certain kind of diet?
Edward Robinson 1:39:30
No. So with with young kids, sometimes, you know I have a pop tart or half a pop tart for breakfast, which I love. There's a great quote and one of your episodes where you like pop tarts are poison, which I think is very correct. Especially when it comes to diabetes. It just like shoots your blood sugar through the roof.
And then but in general, we eat pretty healthy. Lots of fat adjustables but eat pasta, I don't restrict myself in any way.
Some days, I do intermittent fasting for like the health benefits, there's some really amazing science out of like Harvard, and some other places around extending lifespan. And one of the effective ways to do it is to do intermittent fasting, which is really cool.
Scott Benner 1:40:28
Yeah, I haven't been like, precision careful about it recently, like this time to this time. But in general, I have a pretty short eating window in the course of every day. And I think it's a major, I can see a major benefit from it just on on how my body weight stays, you know, not necessarily eating, like Jenny. And if we're keeping that window, shorter stops me from gaining weight. If I expand the window, my body starts to hold on to weight. It's really interesting. And I do want to pay more attention to it. I, it's tough, because we're at a really rude time right now, Edward, where a lot of people have opinions about things. And you don't know if their opinions are rooted in any kind of, like, the only me like, there's, it's so funny, because I'm about to like, I guess what I'm gonna say it's gonna make it sound like I could be in that group, too. And I understandably, could be, but like, I have a podcast. So I know how to make my voice on the internet. And I know how how people can find it. A podcast has grown the way it has, and it's existed the way it has, I believe, because what people learned from it ended up being valuable, and they told somebody else. But I don't know if I look for intermittent fasting podcasts, I'll find 25 of them. I don't know what those people know, or what they don't know. And it's the same even with simple things like I want to do more content about just like vitamin supplementation. But how do you find somebody who knows what they're talking about? Who's not a crackpot? Or just trying to pimp their
Edward Robinson 1:42:10
How do you know that they're not correct? Exactly. And that's the challenge. Yeah. And
Scott Benner 1:42:14
when you talk about intermittent fasting, the same thing comes into my head that most that some people are going to hear that and go, Oh, yeah, sure. You know what I mean, or I'm sure that I'm sure that's the feeling people get when I had somebody tell me the other day, they stood up at an event, and they're like, who, who listens to the Juicebox? Podcast, there's a bunch of people there with type one diabetes, and I thought this was gonna be a story about how they all did. And it was not many of them, you know? And then, and so the, so the process starts, like expanding on why they think they should listen to the podcast. And he said that at the end, a guy came up to him and said, Amen, I'm not gonna listen to that, man. Just sounds like you're doing a sales pitch. And the guy's like, no, he's like, I believe in it. Like, you should try it. And so how do you you don't I mean, how do you make that leap? Anyway, there's no answer to that statement. It's just it's, we're in a time where anybody can express their opinions. And it's easy for other people to get to them. And how do you pick? Who's who, because I've had people come up to me, they're like, I know about supplements, you start talking to them? Like, I think you think that there are spaceships coming to get us? Like, so? I don't know. I don't know how to take you on vitamin D right now, you know. So, anyway,
Edward Robinson 1:43:23
yeah. So personally, I like finding stuff that's more like research based with and like, read, or even sometimes reading the research. But it takes a lot of time to do that. And then it's amazing how often, like research is just done terribly, terribly bad. There's this fantastic book about, about pregnancy, and it's called expecting better. And it goes into a lot of like, the deep dive deep dives into a lot of the research around pregnancy. And one of the fantastic things was they recommend, you know, not drinking when you're pregnant. And I'm not a medic, don't take this as medical advice. But the the book then talks about how in some of the studies, the people that they studied, were drinking, but the people who like admitted to drinking in the study also had a number of other really bad complications, like doing heroin or crack or things like that, that you probably would conflate the results. And so it's amazing how even like, research with the best intentions can still be done poorly. And just because you publish some paper doesn't mean that it's valuable. And so it's really hard to sift through that and find the right answer.
Scott Benner 1:44:44
Basically, think about during the height of COVID. When you couldn't, you couldn't make a turn on the internet without somebody telling you like just zinc and vitamin D. And I'm like, listen, there's no doubt I should probably have a certain level of zinc in my system and a certain level of vitamin D in my system, but they would say Like, that's all you need. And yeah, that COVID It just, you know, vitamin D heap your vitamin D up, you'll be fine. I'm like, how, what? And, yeah, can you prove that? You know, I heard it. And then that ends up being the problem is that someone says it, and then people start repeating it. And then before you know, it's a rule, and you're like, wait, what, like, did anybody look into this? And you know, it's, it's just very interesting. I'm not saying even that, you know, maybe thinking vitamin D did stop COVID? I have absolutely no idea. I just don't know how to believe any of it is my bigger problem. So, yeah, I want to ask you one last question before I let you go. And I have you wait too long, I apologize. It's still with the system you're using right now. You still have to Pre-Bolus your meals, right.
Edward Robinson 1:45:48
So I don't actually Bolus you but you still have to put in the car at least 20 minutes beforehand. And then it will it will Bolus on its own Yeah. Sometimes if it's like there's safety restrictions, if my blood sugar is dropping or below the target and things like that, it will be conservative when Bolus thing. And so I still will Pre-Bolus Then, because I know that even though it's dropping, the food is going to make it go up. And so then I still Pre-Bolus
Scott Benner 1:46:16
Yeah, but I just want people to understand that even when we're talking about something that sounds as futuristic as what you've talked about today, they still have a responsibility. There's still
Edward Robinson 1:46:27
Yeah, yeah, exactly. So the, the French company that I talked about, the CEO of that company has given a few interviews, and they have said that, you know, if you don't Pre-Bolus And if you don't even Bolus your meals, you can still get reasonable results. Like, it's just gonna take a while for the insulin to catch up with the food is the reality. But I mean, there's still things that any algorithm just can't predict,
Scott Benner 1:46:58
right? There was a company, I mean, I assume they're still working on it, but it feels like it's 10 years ago, now. I'm sure it wasn't that long. And they said they were gonna have an algorithm where you could just tell it, I'm having a large meal, a small meal, a medium sized meal. And yeah, I don't know, whatever came up with that.
Edward Robinson 1:47:16
So that's, that's interesting. That's one thing that I want to try and do is see if I just use purely the description of the meal, to see if that can be sufficient. Because usually, when I eat pizza, I eat about the same amount of pizza. And when I eat a bowl of cereal, it's roughly the same bowl of cereal. Yeah. So it should be sufficient. And then sort of the next mind blowing idea would be to take a picture of it. You just take pictures of your food. And the there already already are algorithms that can tag food and pictures. So you just take the tags out of the picture. And Google can already tell you if you're eating, if a picture contains a hamburger or a chair, or, you know, a salad. And so you could just use that to then predict how much insulin to give.
Scott Benner 1:48:12
I mean, pretty amazing. If one day you just held your phone up over your plate and got insulin for that'd be that'd be insane.
Edward Robinson 1:48:18
I think that there is an app that tries to estimate the number of carbs based on doing
Scott Benner 1:48:23
is because they tried to buy ads on the show one time and I was like, let me try the app. And then I tried it. And I was like, Thank You didn't exactly how I was hoping. But yeah, I
Edward Robinson 1:48:35
think if you did it like on a if you trained it on like a per person basis, because most people eat pretty similar meals most of the time, right? And like humans are creatures of habit. So you can learn those habits.
Scott Benner 1:48:49
Yeah. Well, that I believe and trust me. I don't know anything about science or machine learning. But I knew the say that years ago, and I think the first time I said it, the Jenny. She was like what now? And I'm like, people eat the same things. Right? And she's like, Yeah, I guess so. And I was like, yeah, it's not like you're like, you don't I mean, not like you haven't frog legs one day in a McDonald's cheeseburger the next day, you probably live in a certain existence, by the way, it was Ed Dimino that I was thinking of. And this is back in 2014. I think it's at Boston labs. I forget, but he was working on. I forget what he calls it even some sort of bionic pancreas was what the phrasing they were using back then. But it's one of those things. I have no idea where that is in the process. It's eight years later, and I haven't really heard much about it. But I think that he was the one who was saying, or maybe it was the other company. That's the other problem is that sometimes these people have ideas that they incorporate. And then suddenly, and I'm not saying this about anybody specific but then suddenly raising money to keep the company going He comes the company's job. You don't I mean, yeah, you get in that situation and
Edward Robinson 1:50:04
trying to roll it out just becomes So, so difficult. Yeah, yeah. i The other thing that I found, though, is that, so people are creatures of habit. But for me at least, we, we eat the same meals, but just not often enough where I've forgotten by the time we come back around to the same meal. So you know, we'll have pizza from this great pizza place near us, but we only eat it. I don't know, every two months or something like that. And I don't remember exactly what I gave myself two months ago. Sure. Yeah, but an algorithm can remember that and learn that for you.
Scott Benner 1:50:42
Right. And you can say, you're even saying that at some point in the world, it might be able to say, I know you're a Domino's right now. And yeah,
Edward Robinson 1:50:53
I mean, you can track your location on your phone, right within Domino's versus Pizza Hut versus school or whatever. And, and you can use that to, to predict how much insulin to give,
Scott Benner 1:51:06
hey, I want to give whoever on the pod hire to do this search engine optimization. I want to give them a big shout out because I Googled Ed Dimino insulin pump and I got a return back for Omni pod five at the top. So somebody's doing their job extra well over there. I could be
Edward Robinson 1:51:24
that or it could just be Google monitoring everything and put recommending ads for you pushing you to ads. Yeah. Because you've gone to Omnipod site.
Scott Benner 1:51:32
And here it is. He was calling his pump. I let from beta bionics i l e. T. Yeah. Interesting. delayed by the FDA. In mid to late 2021. This could be delayed by the FDA. It says, I don't know. Yeah. Anyway, listen, I am. I was drugged into the world of looping by a lovely woman named Gina who just wanted to hear more about it on the podcast, and I can't thank her enough. I think it's made a significant improvement in my daughter's life. And as to be perfectly honest, I don't see a world of we've used on the pod five, and it was absolutely terrific. But, you know, we went back to loop after after a few months. And we really understood I understood, I understand on the pod five forwards now. But yeah, you know,
Edward Robinson 1:52:24
do you still use an orange link?
Scott Benner 1:52:26
Yeah, but I heard there's a better one now. So
Edward Robinson 1:52:29
yeah, so that's what we're using dash though using on the dash, you don't have to carry that round anymore. So it's just carrying around your phone. And it's it's fantastic. So
Scott Benner 1:52:38
Edward, if you would do me a favor of keeping in mind that you and I are recording something that people won't hear for six months, so that I have to keep the timeline of my podcast intact and keep Yeah, keep this to yourself for the next six months. So your episode comes out. So aren't used on the pod five for a couple of months. It was exactly what I expected it to be. I learned how to use it. I understand how to use it. And she said to me, can I go back to loop? Because she didn't want to carry the controller for the only part five she loved having loop on her phone. Right? And yeah, it's a good, it's interesting, because it wasn't about the care to her. It was about the functionality, right? So what I did was because you have to carry the controller, yeah, she wanted to carry the controller, she didn't wanna have to carry an extra thing. And so we have listened. Everybody should be getting enough. I have enough arrows pods to last us for a little while. Okay. So when she went from arrows to on the pod five. And then we left on the pod five, I did not go back to Eris pods, I went back to dash pods. And I'm going to try so we're going to use up the arrows pods we have and then by that I think the dash version of loop will be ready and out of beta. And then I'm going to translate Yeah, I'm thinking around the holidays, that should probably happen. So she's gonna go off to college on loop. And we're actually sending her with the on the pod five stuff as her backup. Yeah, in case her algorithm crashes and somebody we can't do something about it from a distance because she's going away pretty far to school. But yeah, that that that's the next step is I think looping with Dash for us. I would imagine by the time someone hears this we're doing it is my expectation.
Edward Robinson 1:54:23
It's it's really great. Being able to use the dash without needing to carry around anything else.
Scott Benner 1:54:30
When we bring that when I bring this up to Jenny, she keeps bringing it up that she thinks that you can't change the impact time in the dash version yet of loop. I don't know you probably don't pay attention to loop. But in the you know what I mean? The ers version, you can say I think this meal is going to hit over three to five, six hours. And at the moment okay, I guess you can't do that and dash but I hope I'm hoping somebody adds that to it.
Edward Robinson 1:54:55
Interesting, I wouldn't have expected that to have any impact because you Like the pump, at least in Android APS, because it's poured so many pumps, that pump like integration is its own separate piece from all the rest of the algorithm stuff. So the algorithm should continue to work, no matter which pump you plug in to use.
Scott Benner 1:55:15
Yeah, I think specifically, what Jenny heard was that the dash version of the loop did not have that ability yet. And she was hoping that somebody would put it in. But it does, it does. It's a big deal, because you can, you can make a Bolus at three hours. And this thing's like alright, well, you're good. But you're looking at a blood sugar that's 170. If you just go in, that's one of the amazing things about loop that I missed with Omni pod five is that I could go back in and change the parameters of that Bolus from two hours ago and say, hey, you know what, this is actually hitting over four hours now. And boom, here's how much insulin I think you should put in for that. And you can't you can't edit your boluses and Omnipod. Five. So once you put them in there in that, that's one of the things I love about loop the most is being able to go back and edit the previous Bolus.
Edward Robinson 1:56:08
Yeah, I do think that for Omnipod, five, they, they will roll it out to more phones pretty quickly. I mean, so pretty quickly would be like, over the next six months to a year or something like that, that they're going to add more and more phones.
Scott Benner 1:56:26
There's also the last part of this is I want her to be on a retail system. I want it to get to the point where I'm like, Yeah, put that on it does these three things we needed to do for me. Yeah, for me, the truth is that Android APS loop, tandem control IQ, and on the pod five, these are not things that are competing with each other. They're not they're not apples to apples situations, like I think I think control IQ and on the pod five are similar. And I think loop and Android APs are similar. I don't like somebody asked me like, are you going to compare the two? And I was like, that's, that's just there's just no reason to do that. You know, right. They're not the same thing.
Edward Robinson 1:57:13
Yeah, they each have like pros and cons of like, tighter control versus more, more effort and work to set it up. And right. Yeah, no, there's a complete trade off so many different things. Yeah,
Scott Benner 1:57:25
if we've only had five if it if you could target a little lower. And if it would be more aggressive on a rising blood sugar, I think, you know, we would have held out until the phone was available. And the problem is that art is a young girl, she's, you know, she's hormonal. And we were having to leave the algorithm to stop some of these big rises. But having said that, I do that, too. It's not like loops was magical with the blood sugar's I would say that the biggest difference is that loop holds Arden steady at a lower number. Like that was it especially like overnight and stuff like that? Overnight? Other than that, 85 Yeah. Other than that, they were very similar for us. You know, yeah. So, you know, but needing to
Edward Robinson 1:58:15
carry around a phone and an extra controller. I would imagine it would be a pain
Scott Benner 1:58:20
when I opened the Omni pod five box, and the controller came out art and said, and this is a quote, what is that?
Unknown Speaker 1:58:28
And I said,
Scott Benner 1:58:31
that's the controller. And she goes, Oh, no, no. I said, No, we're gonna we're gonna use this and we're gonna try it, you know? And, but yeah, it was just, she was like, I don't want to carry another thing. And I was like, okay, yeah, like, and that was a big deal to her like, so I'm like, Alright, so
Edward Robinson 1:58:46
yeah, totally. Yeah, totally get that. So I just carry my phone, my wallet, and two glucose tablets in a little like, pill container that I found on Amazon. And that's all I carry with me.
Scott Benner 1:58:59
That's amazing. Yeah, that's, that's what you're looking for. You're looking for less impact. They're all gonna get there. I mean, I can't tell you. I have no trouble suggesting on the pot five to people. I think it was really, it was terrific. It just was slightly less customizable than I was accustomed to. And she was carrying another thing. So who knows if that app was already on her, like ready for her phone? She might not have ever brought it up. You know? Yep. So because she doesn't use the functionality and on the in loop. Like, you know, it's not like it's not like artists like oh, I should probably change my insulin sensitivity because it's after my period. I tell her that's what's happening. And she goes Oh, okay. That's why I loved what you were talking about so much. Because as interactive as Arden is she's still not she's not that person. You know what I mean? That if she becomes an adult, I get less involved. I imagine that means her blood sugar's are gonna go up if she doesn't have something that's going to think for like that.
Edward Robinson 1:59:56
Yeah. Anyway, and it's, it makes so much sense for the algorithm to just do cared for you. It does the me it has all the data. Why can't it do it? Yeah,
Scott Benner 2:00:04
especially after talking to you it does to me so Alright, let me say thank you very much I appreciate you doing this
of course, we're gonna thank Edward for spending so much time explaining all of this to us. And I actually spoke to him last night for a second his a one C last one 5.5 And this episode was recorded six months ago. So it seems like things are working pretty well for Edward and his choice of insulin delivery. Or the thank us Med and remind you that we get our supplies from us Med and you could to us med.com forward slash juice box or call 888721151 for links in the show notes links at Juicebox Podcast a US med touched by type one and all of the sponsors. Since this one went long, I'll just tell you check out the private Facebook group where you can meet people like Edward Juicebox Podcast type one diabetes on Facebook private group. You'll love it. And goodbye. Thanks for listening. I'll be back very soon with another episode of The Juicebox Podcast. It won't be as long as this one
Please support the sponsors
The Juicebox Podcast is a free show, but if you'd like to support the podcast directly, you can make a gift here. Recent donations were used to pay for podcast hosting fees. Thank you to all who have sent 5, 10 and 20 dollars!
