This Week in Startups - Hyfe’s Joe Brew on using AI to quantify coughs + Presh on generating content with ChatGPT | E1721

Episode Date: April 14, 2023

Hyfe CEO Joe Brew joins Jason to discuss quantifying cough to prevent the spread of disease. They break down how Hyfe’s technology works (2:48), how his company diagnoses diseases via audio and qualifying health data (14:47), before Presh joins to demonstrate how to easily create business content for your startup with ChatGPT (40:16). (0:00) Jason kicks off the show (2:48) How does Hyfe work?  (9:03) Notion - Sign up for FREE at https://notion.com/jason (10:32) Acoustic AI and its potential   (14:57) What is a cough, and why do they happen?  (17:17) Chronic Cough (21:06) Hyperice - Get $50 off your order of $150 or more with code TWIST50 at https://hyperice.com/ (22:34) Managing a cough (25:18) Diagnosing other diseases through audio   (31:56) Licensing the technology and pandemic prevention  (38:47) Mayfair - Get 4.35% APY on your cash and increase your FDIC insurance coverage at https://getmayfair.com/twist (40:16) Quantifying health data (43:45) Generating ideas from existing content using ChatGPT (49:45) Using Chat-GPT for your product or service from scratch (53:51) Creating micro-content with Chat-GPT FOLLOW Joe: https://twitter.com/joethebrew FOLLOW Presh: https://twitter.com/preshdkumar FOLLOW Jason: https://linktr.ee/calacanis Subscribe to Founder University https://www.founder.university/podcast https://www.youtube.com/channel/UC8bLXgWx7dTGcA9kssYDjdw?sub_confirmation=1 https://open.spotify.com/show/1p39ZzpP0ZcUCcEX293y3j https://podcasts.apple.com/ca/podcast/founder-university/id1648407190 Subscribe to our YouTube to watch all full episodes: https://www.youtube.com/channel/UCkkhmBWfS7pILYIk0izkc3A?sub_confirmation=1

All right, everybody, I got a great show for you today. First up, I interview Dr. Joe Brew. No, it's not about coffee. Joe Brew is the co-founder and CEO of a startup called Hife, like Life with an H. Hife uses your phone's microphone and wearables to track your cough patterns, right? All that stuff. Then they use AI to analyze your coughs and give you preventable measures to take. You can check them out at hYFE.a.i. It's an awesome interview. do AI is changing everything, including how we diagnose your cough. Then a fun segment for you, Prash, who co-runs Founder University with Kelly on my team, is going to demo how founders can use Chat GPT to automate content creation for your business. This is super important. Everybody's got to create content. Presh is an expert at that. He's going to teach you three things.

How to use ChatGPT to come up with engaging topics. Two, how to prompt ChatGPT to create a blog post from. those topics and three, how to take a blog post and create a bunch of micro content for Twitter threads and other social media platforms. And if you get value from this tactical talk, then you're going to love the Founder University podcast. So take a moment, hit pause here, search your podcast player for Founder University, YouTube, Spotify, Overcast, Apple Podcast, whatever it is. And please subscribe to Founder University so you can hear other tactical talks as we do them.

We're going to put some at the end of This Week in Startups so you can learn about it or go into the description here in your podcast player and click the links to go sign up for the newest podcast I am producing.

You heard it all in This Weekend Startups. This is my third Founder University Podcast. It's going to be a great show. So please stick with us. This Week in Startups is brought to you by Notion is now offering eligible startups up to six months free for their Plus plan,

plus unlimited AI for that six months. Start saving your startup some money and apply for Notion for startups today at Notion.com slash Jason. Hyperice. Warm up and recover faster with the Hypervolt 2 massage gun and NormaTech compression therapy boots.

Save $50 off your order of $150 or more with code Twist50 at checkout at hyperice.com. and Mayfair helps venture-back companies protect and grow their cash automatically. Dversify where you hold your cash, increase your FDIC insurance coverage, and earn up to 4.35% yield on deposits. Go to get mayfair.com slash twist to get started today. All right. Next up on the program is an idea that you're going to just hear the idea and go, well, why don't we have that already? What a brilliant idea. I'm glad somebody's working on it. and that person who's working on this idea

that you're about to smack the side of your head and go, oh my God, so obvious, so happy somebody's doing this, is Dr. Joe Brew. He is the co-founder and CEO of Hife, as in,

uh, like rhymes with life. It's spelled H-Y-F-E, and it's a smartphone app. And it tracks the sound and frequency of what, Dr. Brew?

And welcome to the program. Yeah. So thanks so much for having me, Jason. Uh, it's great to be here. So Hife is a call. company. We're obsessed with cough. We think that it's, I mean, we know that it's the world's most

prevalent symptom. Everybody with lungs coughs, you have to cough or you will die. You know, it's a very healthy thing to do irregularly. And if you're doing it regularly, you might be dying for another reason because you're sick, right? And despite being the world's most prevalent symptom and the number one reason that people go to the doctor. By the way, you just coughed and soon detected it. And and cancel the noise. So there's already some audio processing. But despite being so prevalent, it's largely ignored, and it's certainly under service.

So when you go to the doctor with a cough or when you take your daughters to the doctor with a cough, unlike anything else in medicine where they're going to poke you and prod you and stick things in you to measure, blood tests. With cough, there will be this kind of voodoo where the doctor says, how's your cough? and you, the patient, are supposed to answer, you, by the way, are incapable of answering because you don't have a memory. Wet, Fleming. Those are kind of the two, three, maybe four words we have for it.

We don't have the acoustic abilities of a dog to actually hear the different frequencies and cough. And, of course, we have no recollection. Like, I can try to tell you how well I slept last night, but if I try to tell you how many times I coughed yesterday, it's impossible. So, like you said, I don't think we're particularly genius group for coming up with this novel idea. I think it's a somewhat obvious idea that cough matters and it's ignored and somebody needs to solve that problem.

All right. So, and you can check this out at h, y-f-e dot a.a-i. Dot-a-i. There it is. So the, how does it work exactly? I install your app and then like Siri, it is persistently on listening the entire time or I have to put it in record mode. Obviously, there's a bunch of privacy issues there, and I don't know what Apple lets you do with background recording of sound. So how does it work?

Yeah, great question. So, you know, we'd like to call ourselves a capabilities company and not an app company. So one way you can interact with our technology is through a phone app, and that's the, you know, use case you're pointing to. And in that case, you can download free from the app store or from Google Play, and you get this thing, and you say start tracking, right? Okay. it will process sound on device. So, yes, it's listening, but not like in the sense that a human is listening, right? It's registering frequencies.

We have something called peak detection, so it detects these explosive, abrupt departures from baseline decibels. And when it does, it passes that little snippet of audio, again, on device through a neural network, trained on millions of sounds to say, hey, is that a cough? Is that not a cough? When it finds a cough, it timestamps and counts them up, right? Now, if you're running this on your phone, you're competing with a lot of other things on your phone. And your phone is pretty smart about saving battery, about protecting you from things that are accessing your mic, et cetera. So it's not, I mean, it's a great way to interact with the technology. And I highly recommend anybody who's curious about cough tracking to go out there and just download it on their phone.

And we have people who've used our app for years on their phone continuously. Wow. But phones are suboptimal. One, because you're in a war with other. apps. And two, because phones don't always go in the right place. They go in your back pocket. They get left on the kitchen counter. And if you're really concerned, if you mean, if you really have, let's say, chronic cough for an infectious disease, you probably want to be tracking continuously, right? And so we also build a wearable. We have our own wearable that's a kind of

more medicalized use case. And we're looking forward to a future where we think just like step counting is integrated into wearables kind of everywhere. We think there's probably a future where cough counting will be integrated into wearables everywhere and hopefully that future kind of goes through our licensing. I think an app on my

watch would be good too. Does my Apple watch, is that a possibility? Or again, is Apple so concerned with protecting your privacy that, you know, and so concerned with the battery life that it wouldn't do this? And is there a way for me to override? I'm sure on an Android.

watching and do whatever the heck you want. You can jail break it basically. But if I was in serious mode, I might say, you know what, I'm willing to have my battery in the next 12 hours run out or, you know, take half my battery life down because I have a cold and I want my doctor have great information. Yeah. So, you know, right now it's not. We do have an Android watch and it runs.

Like you said, you just have a lot more flexibility with what you can do with Android. There's no way right now within the Apple ecosystem to run the mic continuously. and we need audio in order to do this. We do think there's a future. And most of our first year of existence was about getting really accurate models. And the last year has been much more about getting really fast, light, extremely efficient models. And I think we're coming to that inflection point where an Apple or a garment or a polar or whoever else would say, hey, you know what, the value, the cost of benefit of cough tracking now begins to make sense.

When we're draining your battery 30% faster and telling you something that, that nobody knows what to do with. That's one value prop. But when we were only draining your battery 5% faster and giving you insights into this whole new kind of science of cough tracking, then I think users will want it. And if users want it, then of course the manufacturers will also. All right. You hear me raving about Notion all the time.

We run our company on it. And now Notion lets you leverage really powerful, artificial intelligence. My guy, Prash, works for me. he takes everything we do when we do a tactical talk at foundry diversity or we have an internal meeting and he'll take the transcript right we cut and paste the transcript into notion but then he will highlight it and then he uses the magic of notion to create bullet points and a summary and this is incredible what would have taken him an hour or two of work to summarize a meeting it's done in seconds i mean you're starting on third base right now great news for all of you

Notion is offering eligible startups up to six months free of their Plus plan, plus unlimited AI for six months. Notion is going to help your startup. You're going to save money because you can consolidate many of the different tools you're using. It's going to make you super efficient, all the infos in one place. Notion is becoming more powerful by the day and you need to get in on the action. Notion is now offering eligible startups up to six months free for their plus plan. Plus, again, unlimited AI for six months. through the Notion for Startups program.

You start saving your startup some money and apply for Notion for startups today at notion.com slash Jason. When you use our link, you're supporting the show. Start saving your startup some money. Apply for Notion.com slash Jason. What can you tell from the sound of a cough in terms of the nature of disease?

Or is this something where you're going to record a bunch of people, then they're going to get confirmed that they have certain diseases? okay, this person had COVID, oh, this person had this flavor of COVID. This person has whatever that new respiratory diseases that I think I got that was particularly brutal. It was like 10 times my COVID. So is it something like over time, we're going to learn these, this cough pattern over

this many days is, you know, what results in COVID is what results in whatever that respiratory diseases. So it's like this whole, it's a great question, right? What are we going to learn from cough? We're learning a ton already. This field is just on fire. And it happens to be like acoustic AI is having its moment at the same time as chat AI and natural language processing and image.

And those get all the buzz and attention. And that's actually fine with us because it means we kind of get a head start. But acoustic AI is is incredible. I mean, one, everybody has microphones with them all the time. Like you said, you've got one on your wrist. You've got one in your pocket. And they're attached to computers. So the capabilities are amazing.

We're very science first. We're a group of scientists and we work directly with scientists. What we're trying to do is build the capabilities and let the scientists discover what there is to be discovered. Like, you know, the manufacturers of the first x-rays were not trying to diagnose disease. They were laser focused on doing this thing correctly. And then, of course, a ton of radiologists. I mean, a whole field was born out of it, right? Similarly with the thermometer, you have a thermometer in your house, I'm sure, so do I.

not because the thermometer makers were the ones who are figuring out exactly what a thermometer is for, but once you make the tool, medicine and science itself discovers these things. But there's a couple that are super interesting, right? So, you know, recently there was a competition sponsored by the Gates Foundation where they looked at the acoustic signature of tuberculosis in screening cases. And there's a significantly better than random. So AUCs of 0.7, 0.8, just by getting a cough to tell whether or not somebody has TB, right? And TB is infamously difficult to diagnose and it's infamously underdiagnosed and

people remain infected and infectious, you know, they infect others, they die in their communities because they can't get diagnosed. And you think of the fact that there's six billion phones floating around that are all mic-equipped and you could suddenly screen the entire world for tuberculosis, right? Something that public health is trying to do piece by piece, house by house, like you can just so leapfrog current screening technologies. For COVID-19, there was a study that came out where they used HIF, the Emerging Pathogens Institute in the United States, the University of Florida. They put Hife next to the beds of those that were severely infected with COVID-19,

just to look at cough counts, right? Nobody ever looks at this. You're looking at everything with these patients. They're the most monitored people in the world. Blood, oxygen, pulse, respiration, right? Every little level, they're getting poked and prodded. And when you looked at cough counts, there was this really interesting finding, which is that when the coughs start to go down, that is a leading predictor of that person is either going to get intubated or die soon, right? And it's kind of counterintuitive because you think more cough, more sick.

But once you actually start looking closely at this, it doesn't necessarily match with your intuition. In this case, our current hypothesis is that when you're extremely ill with COVID and you start to cough less, it's basically because your body is dying. diverting resources to keeping your organs functioning. And you no longer have the energy for that high caloric activity. Ooh, that's interesting. Yeah. And I mean, we don't know, right? This is new science.

But if that's confirmed, and they did reproduce the study in Montreal, same exact thing. You know, once the coughs go down, people start to die. It might not be that the disease is, you know, receding. It might be your body is receding. You don't have the energy to cough. Exactly. Exactly. And you don't have any other indicator necessarily that's showing you that.

And you can imagine, oh my gosh, we need to get some oxygen over to bed number seven because their coughs are going down, right? And so we're actually very modest in the claims we make about the technology, but we're very confident that putting this technology out in the world is going to lead to a lot of these kinds of insights. Now, a cough at its core is, correct me if I'm wrong, my understanding was you have some irritation in your throat, your lungs, something. and that irritation causes your body to have that reaction and that reaction is in order to expectorate, is that the word for like spitting up something and hacking something up? Am I right in that that's the nature of a cough?

Or are there other reasons coughs happen? Well, yeah, so there are other reasons too. So the correct way to cough is what you just said. You have this kind of feedback loop. You have sensors, if you will, right, in your throat and it senses some sort of obstruction or some sort of fluid build up, and hey, you just kind of feel this urge. You just got to get it out, right?

And that's great. Now, a huge portion of the population, between 10 and 30 percent of American adults, have chronic cough. And with refractory chronic cough, it's not that they're coughing because of a problem. Cough is the problem, right? So you could call it kind of tick cough. So they have an over sensitive cough reflex, essentially, even when there, is no fluid buildup, or even when there is no

irritation, they feel they need to cough. They just have this. They are sick. The coughness is itself the sickness. Oh, got it. So they're not suffering from like a viral infection. They're not choking on something. They didn't burn their throat. They have a tick, a habit, some

kind of reoccurring habit of coughing. And then that causes more coughing because it causes irritation. Exactly. And as is often the case with startup journeys, we went And we're like, oh, we're going to build... Sorry, coffee.

I've got to be self-conscious at my cough. Exactly. I've probably just given you chronic cough now, you know? Well, you know, this is a... You have a nice cool drink that seems to get rid of the need to cough. And when I'm speaking too much... Yeah. You can get a tickle in your throat and then want to clear your throat.

So is that something with performers that's, like, unique to them? People who use their voice a lot. Well, you know, we've actually got a lot of interest from vocal performers and speakers, you know, because they, they're self... Everybody's a citizen scientist. when you're dealing with cough and they're playing with humidifying beforehand and inhalers and hydration and then, you know, counting their coughs. And it's hard to count your cough while you're giving a speech.

But if you have a device, that can do it. But to get real back to chronic cough real quick. Yes, please. You know, the startup journey is so often the case of you have an idea and that idea is so wrong. But because you rapidly go to users, you go to the market, you learn so much. And so our idea, being epidemiologists and public health folks and COVID happening, we're like, we're going to go and build this thing for infectious disease surveillance.

We're going to count the coughs over space and time, and that's going to allow us to see where diseases are spreading. And they're like, oh, we're so smart, right? And then we go to the world, and nobody really wants to monitor their coughs for infectious disease, but because by the time you download the app, you're kind of getting better in most cases. And then you're kind of like, I forget it. Meanwhile, there was this population I knew nothing about, again, 10 to 30 percent of Americans

who the moment we put this app out in the wild, they were like, give me more. And we had these amazing power users early on. I mean, our app sucked, right? In 2020,

we had these stories of like dog barks, a thousand coughs when it's all dogs. One guy went on a hike and it said like he coughed 5,000 times. It was just his like microphone rubbing in his pocket. But they stuck with us because there was such a need from chronic coughers. Again, people I knew nothing about a couple years ago who were like, thank you.

Finally, somebody is paying attention to this thing I have. So you found out about what is a chronic cough. And just so we're clear here in the storytelling of this episode. And so I understand where we're at. You started with, hey, we got COVID. We got people who get the flu. But as we all know, like many things that you could use as a treatment,

including like if you were going to get antibiotics, a lot of times you write it out and you don't wind up going to the doctor or needing those things. It's too late. What you found out was there. as a group of people who have what's called chronic cough. And chronic cough is in some way this like sort of involuntary, involuntary reflex. And it's somehow related to this, you know, nerves in your larynx or something. And it just makes you do this for the rest of your life.

You're clearing your throat and you're coughing. And it is a little bit my understanding because I know somebody who had this was constantly coughing or hacking. And we thought behavioral tick. And then there were all these kind of theories that came up. Like maybe there's chemicals around that are irritating them or maybe they have asthma. So did you, is that actually your beachhead market is people with chronic coughs? It is our beachhead, but it's certainly not the only market because the examples you just gave are also very real and relevant.

Some people cough because they have asthma. Some people cough because they have allergies. Some people cough because they have infectious disease. Even chronic cough is not necessarily tick cough. There are other forms of chronic cough, which could be caused by other chronic disease even. Wow. Right. What's really cool is once you start paying attention to this, there are patterns in the temporal

nature of cough that indicate what it might be. For example, if you just like, if you, if I notice looking at your chart, right, that you're coughing more after meals, you likely have GERD, right? Some sort of reflux related cough. And if I notice that you don't cough at night, you likely have kind of a tick cough because basically go into some form of paralysis. And when your brain stops telling you, hey, cough, Jason, cough, you stop. Whereas if I do notice that you're coughing at night, it's very likely that you have an

infectious disease, right? That is continuing to produce mucus and swelling and keeping. So nobody's ever done anything in this space because nobody is actually counting coughs at night because your doctor's too busy to sit by your bed and say, oops, 343, another cough, right? But once you have a tool, it opens up this entire thing. just like before there was, let's say, a glucose meter. Think of diabetic hair. Yeah, you're guessing, basically.

You got a bunch of areas, you're guessing. You can't measure it. You can't manage it. Okay, everybody, are you sore and tired? It's hard to keep this peak performance up, but I have an answer for you. Right as I get off the slope, Hyperice is waiting for me. Here it is. If you've never seen it, these devices change everything.

This is the HyperVolt 2 Pro that I have. It is extraordinary. Here's what it looks like. I love this thing because when I use it, I can take the knots in my back and ah, that feels so good. Takes out all those knots, loosens my muscles. And generally, it's going to make you feel good throughout the day. Hypervolt comes with five head attachments plus three different levels of percussion massage. You can customize it.

However you like, it's lightweight. It's super quiet. Hyper ice also makes heat and massage wearables for your back. That's what I'm getting next. And we all know a bunch of people. They're suffering from that back pain, right? They also have these normatac air compression boots. These things are insane.

Here's a really important call to action. Get 50 bucks off your order of 150 or more with the code Twist 50 at checkout. Go ahead and go to Hyper Ice. That's H-Y-P-E-R-C-E.com and use the code Twist-V-O-T-V-O for 50 bucks off your purchase of 150. or more. C sincerely, my wife got me into this, and then all of a sudden they showed up to be partners on this program.

I was like, I used that product. What? So it's amazing. You're going to love it. So what does management then look like? You're measuring it. Let's assume you've been doing this for a couple of years, I understand now you've raised a bit of money.

So congratulations on that. And you've got a product in market. So have you moved over to treatment? Or is it just, hey, we're going to diagnose this, talk to your treatment provider? Yeah, I think it's a super excited. space. So to answer the last part of your question, hey, we're going to tell you about your cough and you're going to bring it to your doctor and they're going to help you manage it.

That's very much where we are right now. And in fact, we are currently running a clinical trial to validate so as to submit to the FDA so that this can be a medical device that we can actually say this is for your doctor to use to tell you about. But treatment is a really interesting space, right? Because chronic cough is massively undertreated. The current generation of molecular or pharmaceutical interventions, they're not extremely effective and they are, as pharma goes, you know, somewhat toxic. So they cause side effects. Meanwhile, there's a lot of therapeutic content out there, speech therapy, for example, which has been shown in research to be very effective. But speech therapy doesn't scale well, right? So there's little methods. For example,

you stick your tongue between your teeth and try to swallow. And when you have, if you're someone of those people with an hypersensitive cough reflex, you do. suddenly it goes away. And now you're sitting at a public restaurant where people were staring at you for a few minutes because you were coughing so much. Or you're at the opera and you don't want to cough? Exactly.

You can do these techniques. Fascinating. So these therapeutic techniques have a lot of promise. And in fact, I mean, the market is so large for chronic cough that if you can reduce people's chronic cough through pharma or through therapy, there's a lot of value there. Behavioral therapy, basically.

Exactly. But everybody knows that therapy. that therapy is effective is most effective when there's a feedback loop. Right. So you want to lose weight. Good luck doing that without a scale. Right. Or you want to get faster.

Good luck doing it without a stopwatch. You want to cough less. If perhaps you're a refractory chronic cougher, you need a cough counter. Right. And so we do think that there's kind of a future version of Hife, which is integrating the output, which is the coughs with the input, which is some sort of therapeutic that could compete, even with molecular or pharmaceutical interventions. So I'm coughing, I have some wearable device,

and then it gives me, you know, some sort of vibration on my app, hey, you should try doing this for five seconds or whatever it happens to be. Absolutely. That kind of stuff in real time. But then also you could be with your dedicated device, which I looked for the dedicated device online. I couldn't find it. So is that something you've built already or something you're working on building?

Yeah. So the dedicated device is built already. It's called the cough monitor. it's going through these regulatory validation trials. We basically need to show that it works, that it does what it says it does. And then we can market it and sell it as a medical device. Honestly, I don't see the future of cough tracking being dedicated devices. Like how many people really want another device, right?

I really see the future being integrations. But the dedicated device has a real use case in kind of medical settings. And we see it as a stepping stone. And those exist in medical settings. There are devices to track coughing. Of course, yeah, yeah. Well, no, the gold standard in cough tracking is human annotation. So in clinical trials, usually when they do it, they actually record you.

And then they have trained listeners go. And it's extremely inefficient. It scales terribly. You have to carry around like a microphone attached to you. It's not a good system. So we're pretty confident that Hife can kind of win in the medical space. but we're much more ambitious than that. And we think cough tracking is for everybody with lungs.

Yeah, and there's also, like, people would get the HALTA monitor when they were trying to do, like, check their heart rate for a while. You'd just scrap it onto them and they'd wear it for a couple of days. I guess in an acute situation, you could have something like that. But if you're doing it when you're sleeping in your bed, that's a time when your phone isn't being used. And it could be plugged in, and it's stationary. So that actually works really well.

I guess you get a lot of data at night. Yeah. And, you know, I mean, just like the example you just gave, There's acute situations where you know you're sick and now you're going to start monitoring the illness. That's all good and well. But where we think there's this kind of paradigm shift with the fact that we're literally wearing microphones on our bodies right now is you don't have to wait until you get sick before establishing a baseline. And we've noticed in our studies that there are lots of people who are who they profess to be asymptomatic, but they're in fact symptoms ignorant. And I'll give you one anecdote. A study we did in northern Spain, a big population study, COVID-19. A young woman was diagnosed with COVID on February 8. She got a, you know, kind of a drip in her nose, a little bit of a tickly throat. And she went and got a test. She had COVID. She said that her symptoms started that day. But when we looked at her cough counts because she was tracking using Hife, her symptoms started 48 hours earlier. She went from coughing zero to two times per night to eight to 16 times per night. So there was a huge increase in her symptoms. She didn't.

notice because nobody really notices these these kind of fluctuations and symptoms. But the device did. And so I actually think there's a lot more promise, not so much in the treatment of or in the diagnosis, that moment where you already know you feel unwell, but prior to that. Imagine, for example, lung cancer, notoriously hard to diagnose and notoriously late and time to diagnosis really matters for survival. It's foreseeable that somebody with lung cancer would begin to have symptoms. perhaps unbeknownst to them long before they would go seek care.

And imagine if you're, I mean, imagine if you're watched right now buzzed and said, hey, the longitudinal characteristics of your cough and the acoustic characteristics of your cough are consistent with lung cancer. That's why don't you go get trained? That would be like, that's kind of the end game, right? Not just waiting for somebody to get sick and then keep an eye on them. Also, I wonder people's breathing in general, is there a point at time and just acoustically, I don't know what it takes, but I'm sure if you've been working with acoustics to figure out coughs,

there are people who are working on breathing as well, is on your product roadmap maybe to build just breathing technology as well, because what happens in between the coughs, I'm certain is relevant in some way. And boy, this data with AI over time is going to get so valuable. You talking about the cough, you know, with lung cancer, if people just had some way of recording their breathing for some amount of time. Just like you do now with your Apple Watch,

you can do the EKG, I guess, and some people do glucose monitoring for a period of time. All this stuff is coming together. So what about just breathing in general?

Yeah. So, you know, this is, we're having the classic startup struggle between narrow and why, right?

And our beach head, our narrow focus is on cough and it takes a lot of discipline to stay focused on cough because there's so much opportunity and sound. and sound and AI. So we do have ambitions for the entire health soundscape, right? But we're trying to do it in a logical sequence, starting first with the low-hanging fruit.

So cough is loud, percussive, explosive, unambiguous. You can hear a cough. And so we start with counting cough. Once we fill our database, which by the way, you know, we have hundreds of millions of cough-like sounds in our database. Once you fill those, you can start to say, okay, what are the acoustic characteristics of different diseases. Once we've shown value there and people are like, oh, wow, I want to actually, I want somebody to actually track my sound, then we can explain it outwards. And we think of doing it, you know, to sneeze and sniffle because those are very indicative of the onset of respiratory illness, exactly.

But voice is also very information rich. How is my voice change, right? If I'm talking raspy, what does that mean? And there's a few different levels. There's Raspi and horse and the kind of upper rest. respiratory infection, but there's also a lot to be said about cognitive states with voice. Oh, so I'm coming across as very clear and articulate, or I'm coming across as foggy and fatigued or aggressive or at its most extreme early dementia, Alzheimer's, or, you know,

once you start to quantify something, you can really pay attention to things. How many times have you filled a conversation with an unnecessary utterance, like, like like or uh or um how how abnormal or certain speech patterns you have. I mean, that's interesting, but it's not necessarily relevant to health. Meanwhile, researchers are showing that cognitive disease can be diagnosed through very short voice samples. Unbelievable. Right. And then respiration, like you said, respiration is hugely important. You know, I've had a sick infant and you're sitting there seeing, you know, they got a fever, they're breathing fast. They're very clear clinical guidelines. And it's not a hard problem to solve. Right. You just need to get the mic close.

enough and filter out the background noise. And you can count up a respiration rate. And if you do that well, you can deploy this places where there aren't doctors and nurses who are going to do that for you. And so you're going to, I guess, with all these other wearable and phone companies out there, it seems like a company like Apple licenses. Is the idea here is that you can have a lot of licensing partners once you have this data set and they don't have to rebuild it. You're like an API. Yeah. We feel that we have a really strong mode. And it's not just quantity of data, because if Apple or Amazon or Google wants to collect data at scale, they'll pass us in two days, right?

It's really quality of data because it's one thing to have 100 million coughs or a billion coughs. It's another thing to attach each of those coughs to a disease state to know that this is a cough from tuberculosis, this is influenza, this is pertussis, this is COVID-19. And in order to reliably collect that data, Apple or Amazon or Google has to go through the exact same hoops we've spent the last few years going through, which is ethical committees, institutional review boards, consent forms, recruitment of patients. If you really want reliable medical data, you have to play the medical game, which is complicated legally and regulatoryly. So I imagine that.

A medical device is a certain process you have to go through. Yeah, that takes a couple of years. Yeah, and we're deep in that process right now. And you have to show that, hey, this thing works. It counts golf. But you also have to show, hey, it's usable. And it doesn't electrocute people. And it's not going to do certain things.

or it's not going to tell a doctor that somebody's coughs are getting better when they're really getting worse and therefore the person gets off treatment. You know, there's a very high bar. And I think that any large kind of integratable platform, right, whether wearables or just the large tech companies will rationally look at that challenge and say, you know what? Cough counting is valuable. We don't want to reinvent the wheel. Let's work with these guys.

It's just like temperature when I was in Japan, everywhere I went, when you walk in, there's a camera and a monitor. and it's a video of you walking towards the camera. This is in the lobby of the Park Hyatt in Tokyo. And as you walk towards it, it zooms it on your face, it knows it's a face, and then it tells you the temperature.

And it tells you your temperature, green, red, whatever. And I don't know if it alarm goes off or whatever, but it's just kind of like letting you know if you have a fever or not. And that is just incredible for society. You could be an early warning system for the next COVID. If we start seeing coughing, spike up in different regions

in some place like China where they have control over people's phones and devices in a different way than we do in the West because just turn this on and they would know oh my God, this province has too much coughing. And I'm so glad you brought that up because I love to talk about the potential for pandemic prevention,

disease mitigation, and we haven't talked about it yet. One of the amazing things about using cough is that it is relevant for so many different respiratory diseases, right? It's not just like one disease. And the problem with classical epidemiologic surveillance is you wait until you know what pathogen you're looking for, and then you go out and start giving tests and you don't have enough tests. I mean, all of us went through COVID, right? We know the drill. It's if you do surveillance in a pathogen agnostic way,

so rather than counting the cases, you count the symptom, you count the coughs, you would be able to identify abnormal departures from baseline well before you know that there was even something to look for in the first place. I mean, imagine if you had people in the Wuhan market in late 2019 using cough counting. You would have seen this bright red dot on the map and you'd say something is going on there. We don't have a name for it. We don't have a molecular test for it. But we need to go there and see. And if we had done that, imagine how many cases, lives, lockdowns we were saying. They could to just stop the flights out of China immediately. And then any other place that had coughing fits going on or some coughing pattern that's acute.

Listen, it's just a great, Dr. Brew, that you are doing this instead of starting a coffee company or a microbrewery with a name like Dr. Joe Brew, I thought for sure you were going to be headed towards a different profession or startup. It should be coffee for sure, because you can brew a cup of Joe, you know, and I actually, I worked in college with a girl at a coffee shop. Her name was Laurel Coffee. And so it was brew and coffee opening the coffee shop on Sunday mornings. I love it.

Yeah. That is literally a Laurel and Hardy sketch waiting to happen. Who's opening? Brew? Yes. When we open, we brew. Yes, that's right. Joe?

Yes, we're going to brew Joe. Wait, Joe's opening or brew's opening? You know, speaking of health sciences, I'm so relieved every time these new studies come out that say coffee is actually good for you. Because I drink so much coffee and I always feel a little bad about it. And then it's like, you live longer. You feel okay. Don't tell me about the bad ones because I don't want to even hear about them. I just want to live in my fantasy that there are only good things to say about coffee, you know?

I think one of the great things about this moment in time is that we're down to such a finite moment in figuring out what's good or bad for you. And it's like, you know what's really bad for you? Smoking. Obesity. Drinking a lot. And like as you take those things out, we've just seen lifespan. go crazy and to the point of which we haven't been able to control obesity. And then we saw a lifespan in countries where obesity kind of got off the charts directly

correlates with lifespan. And so it's almost like we know how to live to 80 or 90 or 100. It's low stress. It's, you know, a controlled diet. It's being a certain BMI and not smoking and not drinking too much. You know, well, one of the, so I come from a public health background. We can have an average lifespan of 90 if we just did that. Yeah, and you know, the life expectancy, the real drivers of it, the reason why there are these sub-Saharan African countries with a life expectancy of 55 and other countries with 85 is not that they're dying at age 60, it's infants and children.

So it's zero to five's. And in fact, one of the things that makes me so excited about cough tracking is that one of the drivers of childhood mortality is pneumonia. Right. And so people live in a place where getting care is very difficult and getting diagnosed. diagnosed is very difficult. And it doesn't matter how much of, let's say, Bill Gates's money you have. You're just not going to build fast enough to save that kid's life next year, right? And so I'm really excited about tech for leapfrogging certain interventions and access to care in public health.

Tuberculosis, pertussis, pneumonia, influenza, novel pathogens, the ability for people to finally have access to top-notch diagnosis, screening, remote patient monitoring, without necessarily having the good fortune of having been born in a place where you have a doctor and a nurse and a practitioner nearby. When you're a founder, you need access to cash no matter what goes down. This means having more than just one FDIC in short account, of course. And Mayfair is a cash vault powered by Stripe that's designed to keep your money S-A-F-E safe. Mayfair is built for startups looking to diversify and grow their funds automatically. It takes 10 minutoes to set up. That's it, 10 minutes, and it works along your existing bank accounts.

And it's FDIC insured. Here's how it works. You sign up for a new account. You keep your operating cash at your existing bank, no problem. Then you deposit the rest at Mayfair for safekeeping. Then move your money back when you need it, or that Mayfair manage your balances automatically. Mayfair also helps you earn 4.35% APY with no investment risk for our lock. your cash stays in cash.

And it's never invested. Withdraw your funds at any time, just like a checking account. Your funds are held in the custody of Evolve Bank and Trust with powers companies like Mercury and WISE. Sign up today and earn up to 4.35 yield on daily balances. That's more than 10 times the national average savings account. Go to getmayfair.com slash twist. Just start managing your cash like a Fortune 500 company today.

go to getmayfair.com slash twist. Start managing your cash like a Fortune 500 company today. It is going to be amazing when this technology makes its way to all ends of the earth. And we're lucky to have some great founder like you working on it. So carry on, get back to work. And when you get to the next level, come back on the pod. And when you get this new product in market, you get your FDA approval, I guess, for a medical device. Is that how it works?

Yeah. So into this year, we're hoping for our FDA clearance for a medical device, exactly, meaning that you could buy this for medical purposes and talk to your doctor about your cough. But yeah, we'll keep you updated. Yeah, I mean, I got to tell Tim Ferriss about this. I mean, you know, these quantified self people are obsessed with this. And so you need to get on some of these quantified self podcasts because they got early on to glucose monitors, temperature monitoring, sleep monitoring, heart rate monitoring, step monitoring.

and my hope is that all this stuff comes together and we can just basically give what would be amazing is if all this data could come together and they would give you a custom insurance policy based on that and then every day, week, whatever you choose month, adjust it based on your behavior. Just like there's now mileage-based, you know,

and behavior-based insurance for cars, right? So everything comes down to incentives. And in the West, like, what a great incentive. of you can make your health care half the price just by being a certain BMI. I know this is triggering for folks, but, you know, or just based on some healthy habits, the number of steps, et cetera, all wrapped up together. It could be incredible. Yeah, absolutely.

I agree. And please talk, talk to the quantified self. We've put very little attention just because we have assumed it to small market. But it actually, it's super. I'm a kind of quantified self person. And I love to look at this stuff, right? Especially when it's a novel data stream, something that nobody's looked at. Because everybody does their footsteps, everybody does their heart rate, you know.

But when you have something new and fresh, you just get to learn so much. And then to your last point, you know, like your insurance, right? Imagine if you had, for example, in the United States, a lot of the time it's employers that are actually insuring. It's not just about your own health. When it comes to cough, it's about infectious disease. And if, you know, if you were coughing last night, what if you just got an alert? Congratulations, Jason. Time to stay home today.

And I keep you out of the office and you don't infect four other people. and my workplace actually benefits from that. So there's a lot of use cases. I mean, people may not even know the frequency of their coughing. So just telling them, hey, by the way, you coughed 16 times last night. People definitely do not know the frequency. Nobody knows the frequency. Even if you try to count for the next 24 hours, you will definitely forget

and then not keep an accurate count. It's impossible, you know. All right. Listen, great work. And if people want to learn more, where can they go? And are you hiring right now? Hife. AI, we are

kind of the slow trickle of hiring always, and we're especially looking for people that are extremely talented with AI and data, and extremely ambitious in terms of global health impact. So that's maybe a small thing. If you have a passion

for global health impact, and you're good at AI machine learning, all that kind of good stuff, well, this could be your life's work here. You can do something really great for the planet. All right, we'll see you all next time. Bye-bye. Okay, thanks so much to Dr. Joe Brew for an amazing episode. Next up, my guy, Prasch, is going to show you how to use chat GPT to automate content creation for your business. All right, get to it, Prash.

All right. So firstly, we're going to start with generating content ideas when you already have existing content. Let's say you have an existing blog. So for this example, we've got the Founder University blog, which has a bunch of topics here. But now I want to find more topics to talk about. So what I'm going to do is just compile a list here, which I've already got of titles that we've already written about and created content for. And so we're going to take these titles, and this is going to be the existing content. Let's say you have an existing blog. You would take those titles. We're going to head over to ChatGBT, GBT, and I'm going to simply just talk to it like I would talk to a human. So I'm going to say, can you study the following blog?

post headlines and come up with five new, I'm saying five because I want to be specific, five new ideas that are related. Okay, and then I'm going to paste my titles that I just grabbed for my existing blog posts, and let's see what we get here. All right. So right here, we've got five, because I asked it for five specifically. I can ask it for 10 or more. And I've got topics that are based off of what I fed it already.

So it's studied these topics, and now it's given us new titles and given us a little description of what we can talk about in those. So this is great. I'm going to take it a step further and ask it, please give me five new ideas of potential blog posts. Okay, so now hopefully it'll give me a new idea. All right. So here we've got five new ideas that are still similar to the previous information,

but these are new topics that I can write about. And so from here, what I would do is, you know, get these topics and go over to my Notion page and say new blog post ideas. And I will just paste them in here. And then I'll take the other suggestions here. And I'll capture that in my Notion page as well. So in, you know, under a minute, we fed it information. of existing topics that we've written about.

We've then asked chat GPT to create different variations of those blog posts. And then we further asked it to create new blog post ideas. It's given us a description so we can then go deeper into this. So, for example, here we've got the ultimate guide to growth hacking for startups. Uncover the top growth hacking strategies and techniques to rapidly scale your startup, learn how to analyze and optimize your marketing, product development, and user experience to achieve exponential growth. So this is a huge topic.

Let's say I want to expand on this and now go to the content creation part of building out a blog post. So it's got the context here. Now I'm going to take it a step further. Okay, so I'm writing a basic line. It's still pretty broad, but there are some specific things I'm asking. So I'm saying, I like the ultimate guide to growth hacking for startups as a blog post idea. Can you help me create this topic into a short and concise blog post with a clear

your hook, body, and conclusion. So let's see our results. All right. So look what we've got here. We've got a title, actually, the ultimate guide to growth hacking for startups. We've got the introduction. Are you ready to take your startup to new heights? Growth hacking is the answer. So a little bit salesy, so I probably wouldn't, you know, include this or I might ask it to refine it. But the cool part is here in the body. I've not got these seven points of key growth hacking frameworks to use. Right? And so one is understanding the growth hacking mindset.

One is identify your North Star metric. One is leverage the AARRRR framework. One is implement viral loops. So these individual topics can serve as separate blog posts itself. But the cool thing is here, I've got like a 101 to growth hacking blog. And so I can take these, and I'll just save this

information, I'm going to copy this over to here and paste it just so I have this. And I'm just saving it a notion. You can use Google Docs or Coda or whatever tool to save your information. But the cool part is here. I've got essentially a high-level growth hacking blog post that I can then repurpose. I can edit some stuff here. And then take it a step further, create an individual blog post for each one of these topics.

And so now I've got one piece of content, which is a blog post. I can take it a step further and create an individual blog post for each one of these topics, which then you could still ask GPT to, you know, dive deeper into that topic. So I'll run a quick example here just to show you the continuation loop that you can use to further your content creation. So I'm going to say, leverage the AARRR framework. So I'm going to ask it, can you expand on the, AAR framework.

Please provide some examples of this. Okay, so here are the results. It gives me a full explanation of the AARR framework, and it breaks down into four pillars, which, or five pillars, I should say, which is acquisition, activation, retention, referral, revenue. It gives me the definition or the explanation of what they're talking about. And then it gives me an example of a startup because I asked it for, please provide an example of this.

Each framework has now an example to go along with it, which is incredible. So this is the first example of using ChatGBTGPT with existing content, but now using that existing content to create new ideas, and from those new ideas, creating new content. All right, now let's move on to using ChatGBTGBTBT to create content for your product or service that, let's say you have no content published, and you've just got a product out, and now you want to start building content for it. So the way you can do this is just talk to GPT as a person again and explain what your product does.

So in this example, we're going to do a real life example here of a founder university company, cohort three company, called Clear Space. All right. So here's the website of the company. Their tagline is eliminate digital distraction, get control over your screen time. Okay, so the product helps you eliminate distraction, obviously, and it's an iOS and Android app, and it helps you track your screen time. Okay, so let's go back to chat GPT and explain this. All right, so I wrote this prompt in GPT.

So I says, my product is an app that helps you eliminate digital distraction. Our mission is to help people get control over their screen time and be more mindful of their phone usage. Okay, so this is me just prompting chat GPT here. And now I'm going to ask it, with this context, I'm going to ask GPT to help me come up with some blog post ideas. So I want to create a list of 10 ideas for potential blog posts. All right, and here we have it. So I've got now a list of 10, because I asked it for 10, of potential blog post ideas.

So let's go through some of these. Clear space. your key to a mindful digital life. And that gives me an explainer of what I can talk about. Introduce your app and explain it, and explain how it helps achieve balance in their digital lives. Let's go to another one here. The impact of screen time on mental health and well-being.

Discuss the negative effects of excessive screen time and how clear space can help mitigate them. Let's go one more here because these are great. Creating a healthy work-life balance with clear space, tips and tricks. Provide practical advice on how users can leverage clear space. space to establish a healthy balance between their personal and professional lives. Okay, I just read off three, but this list is great. We've got 10 solid ideas that I can then expand on to create a blockpost.

So, let's take it a step further. Let's say, parenting in the digital age, using clear space to foster healthy screen time habits for your children. Okay, so I'm going to say, I like number six. So my prompt was, I like number six, which is here. And I said, write me a blog post outline. And I want it to be short and concise with a clear hook, body, and conclusion. So it gives me the title here, which it already had created.

Then for the introduction, it gives me options, right? Because I said outline, it's not going to write it for me. I can get it to write it for me. But I said outline, so it's going to give me the prompt back. And so I can then, you know, go back and edit. and figure out what I want to add. So it says, present the challenge of managing children's screen time in today's tech-driven world.

And then another one is says, introduce clear space as a solution for parents seeking to cultivate mindful phone usage habits in their children. Then I move on to the body. It's given me just pointers and prompts that I can then expand on. So like the importance of managing children's screen time or impact on mental health or effects on social media and family time. So it continues to give me more that I can include in the body and I can take these points and then expand. Again, I can just ask GPT here to expand on the impact on health or expand on the impact or effects of social skills and family time.

So I can ask GPT essentially endlessly to expand on the content that it's already created. So again, this is just another example. I explained the product right at the top. I explained what my product is and then I asked it to create a list of ideas. and then from these ideas, I expanded it and asked it to give me an outline for a blog post. So you can do this for any product, try it out with your own and see how it goes. All right, now let's move to the last part. So the first two parts were about content creation, idea generation.

Now let's say we've got the blog post created. Now we want to create microcontent. And microcontent really is just a Twitter thread, a LinkedIn post, any sort of social media post that we can create based on the meat, which is the blog post. So, back to GPT. I'm going to switch to GPD4. So what I'm going to do is I've got a blog post here. It's a great blog post on how to give a great presentation written by Jason Calcanus.

And I'm, I've copied this blog post, right? I've just taken the text, copied it. And now it's in my clipboard. I'm going to ask it, I'm going to paste the blog post here. And I'm going to ask ChatGPT to create. a Twitter thread for me because I want to create content on my Twitter account. So what I'm going to do is, again, talk to it like a person. I want to create a 10-part Twitter thread that summarizes the following blog post.

Okay, I'm going to paste it and enter. All right. And these are the results. this is pretty incredible. So first tweet says, want to create a successful product demo, follow these tips. First, make sure to show your product

within the first 15 seconds. And I can just go back to the source of the blog post, and we see that as the number one tip here. But the way it's structured is now it's in a tweet thread format. So I can then refine it by asking ChatGPT to maybe, let's say, take out the emojis or let's say provide more examples of this. But as a base of, level, we've now got a 10-part Twitter thread that I can then publish on Twitter based on

the existing blog post. All right, that's a wrap for today's video. We went over three things. We used ChatGPT to create blog post topic ideas based on original content that we had already created. We then use ChatGPT to create brand new idea by just explaining a product and service. And then we use ChatGPT to outline a blog post and then from there create microcontent that we can then share as a Twitter thread or LinkedIn post.

So next, I want you to try this out for your own product or service and see what you create. Don't just use the content raw. Take the content, maybe structure it a little bit more, fine tune it by asking additional prompts, and there you have it.