Limitless Podcast - The $2 DNA Test That Could Add 50 Years to Your Life | Google AlphaGenome is Here

Episode Date: July 4, 2025

Imagine spitting into a $2 test tube, dozing off, and waking up to an AI-written health roadmap that flags hidden genetic typos and tells you exactly how to dodge disease. That’s the jaw-dropping future unlocked by Google’s new “Alpha Genome.” 🧬✨ In this episode, Josh and Ejaaz break down how this model reads a million DNA letters at once, pairs with CRISPR to correct single-letter glitches, and slashes genome-sequencing time from 30 days to a single night. They also reveal Microsoft’s new diagnostic AI that outperforms top physicians 4-to-1, why personalized “always-on” medicine could save $225 billion a year, and how bio-engineering is quietly becoming the next trillion-dollar wave. Hit play to see how AI is rewriting the code of life, and why living to 150 might be closer than you think.-----💫 LIMITLESS | SUBSCRIBE & FOLLOWhttps://limitless.bankless.com/https://x.com/LimitlessFT-----TIMESTAMPS00:00 What Is Alpha Genome?04:56 DNA For Dummies10:34 The Next Multi Trillion Dollar Industry11:59 Microsoft's Superhuman Doctor16:30 How Does It Work?22:09 Local Models on Your Phone26:44 The Role of Humans29:39 The Most Important Tech Of This Decade------RESOURCESJosh: https://x.com/Josh_KaleEjaaz:https://x.com/cryptopunk7213------Not financial or tax advice. See our investment disclosures here:https://www.bankless.com/disclosures⁠

Picture this. You climb into bed, pop the cap on a little plastic test tube, and spit into it. While you dream, an AI model rifles through all these three billion letters of your DNA, and it tells you what's wrong with you. So by the time your alarm goes off in the morning, your phone buses with a briefing that says, hey, that unexplained seizure disorder you have, well, I found the exact typo in your DNA that causes it. Also, it turns out your liver makes half of the amount of this particular enzyme needed to digest glasses of wine. Stick to spritzers or you'll spike your cancer risk by 60%. And overnight, a test tube spit becomes the most detailed health roadmap you've ever seen, written by an algorithm that reads the human genome, just like you would read Spotify lyrics.

And this is unlocked this week by a new breakthrough by Google called Alpha Genome. So, Ejazz, you brought this up. You kind of crafted the agenda around this for today. So I'd love for you to introduce what makes this technology so special and what is it going to do to the world now that it's out in the wild. All right. So, Josh, this one hits kind of close to home. So I have a background randomly in biology, which was it was my. university degree. I just spent four years. Yeah, I spent four years studying this stuff. And actually,

my thesis was in genetic engineering. So I'm very familiar with the gene stuff. And if I was to summarize the pursuit of biology as a field, it is simply to understand the human body. And the fact that we have so many unanswered questions in today more than ever just shows how complex our situation is, right? How bodily composition is. And the most common. part of our bodies is the genetic material. It's the genes. As you just pointed out, there are a million different base pairs in just one tiny fraction of a skin cell, right? So if you apply that to the, you know, the hundreds of parts of you, the different organs and stuff, it gets really complex really soon because then you've got to figure out how they work

with each other and all this kind of stuff, right? So it would just blow anyone's mind and it's taken hundreds and hundreds of years to even understand how we work at a basic function. So Google announces this thing this week called Alpha Genome. And it's basically this AI model that has the ability to examine every specific part of your genome. And the genome defines who you are, right? Josh, it tells you the color of your eyes, whether you're going to lose your hair at the age of 40, and arguably what kind of personality traits you're going to have as you grow older, right? And these genes aren't static.

They evolve over time. That's why some people get ailments or illnesses or develop certain types of conditions as they grow older. But the issue with all of this is we've never been able to accurately predict when these kind of conditions and procedures are required, right? In the medicinal field, it always acts as kind of like not preventative, but what is it? Is it proactive or like what happens when it's after the condition? It's like it's treatment, basically. It's like reactionary treatment. Yeah.

So like you've already been diagnosed. Thank you. Exactly. Yes, it's reactionary treatment. Thank you. And so what this tool now does is it can sequence millions of base pairs at once. So that's a really important point. It's not something that is like, okay, millions of base pairs and then like you've got to wait like two hours. It can do multiple at a time. So we're talking about like reducing the time it takes to sequence your entire genome, which currently takes roughly around 30 days in some of the best practices and costs like thousands of dollars. to some, to minutes now, right? To your point earlier, Josh, you could sequence your entire genome overnight. And this is just insane for a few different reasons, but kind of like ranking them in priority for me at least.

Number one, I want to know if I'm going to die from some kind of major ailment. I don't want to wait until like, you know, I suddenly get a heart attack or I start losing my hair. I want to know before the fact so that I can proactively prepare for it or prevent it at the first place, right? Number two, I'm sure people who are listening to this podcast have heard this term called CRISPR, CRISPR CAS.

Yeah. Now, this was all the rage when I was doing my degree because it basically describes a methodology where you can use kind of genetically engineered bacteria or viruses. And viruses typically are harmful, but I'm talking about attenuated versions of these viruses, so weakened versions of these viruses, which can help edit certain parts of your genes. And that's really cool, right? because you can take out the thing that which will potentially give you Alzheimer's or whatever that might be. But it's been a highly inaccurate process. That's why it hasn't been scaled to millions of people all over the world. Well, ding ding, now we have a really plausible solution, which will arguably get much better, which will be able to not only identify some of the abnormalities within your genes,

but also directly edit them, straight away. That's amazing. Crazy. Yeah. I would love to kind of step back into a quick primer if biology wasn't people's favorite classes. Because when I was first reading up about this, I was like, okay, I know I've seen these words before, but I'm not quite sure what they mean. So I just kind of want to like quickly define some of the things we're going to be talking about. So we have like DNA, right? And DNA is that double helix fun shape. And it's a molecule made of these four letters. So it's A, T, C, and G. And then these two letters pair up, A goes with the T, C goes with the G, and we call those base pairs. So when EG has mentioned base pairs earlier, that's what it is. It's two of those letters pairing up to create

a base pair. And you stack three billion of those pairs end to end, and that's your genome. So about 2% of those letters are coding DNA. And that is what we previously believed to be the part that we could actually control and edit. And that's how we cured diseases. And the other 98% of it is called non-coding DNA. And what we've discovered through this is that that 90% that we previously perceived as junk is actually super valuable. And not only is it valuable, but now we can kind of read it and we can understand it. And the old technology, it could only glance at like a few hundred letters at once. It was kind of like skimming through this, but it didn't have a lot of resolution, Alpha genome can read, to your point you just earlier about the speed, it could read

one million letters at once, which is a significantly higher resolution product. So there was an example that I read that I loved where there was this five-year-old child who had seizures. And Standard Test looked at her coding DNA, which is the 2% that we thought was valuable, and it came up empty. No one actually thought anything was wrong about it. But because of alpha genome, which can sequence the 98% of the non-coding DNA, they actually found that there was a single letter A that was off, and they flipped it to a T, which was 800,000 letters away from the gene that they thought mattered. And because they flipped it, using the CRISPR technology that you just described, it actually cured the seizures within a few hours. And that was it.

They nailed the root cause by flipping one letter that was 800,000 letters away from the gene. And that's like the type of applications that we're going to see from this, is that you can now really get a fine view of the genetic makeup of our body and see down to the exact letter, where you're wrong, where your body is getting sick, how you are going to die, all these issues with you. That was the thing that was unbelievable to me. I mean, you just simplified hugely complex procedure, Josh, right? Which has only recently been achievable in like the last couple of years.

What Josh, what you just mentioned, you know, 800,000 base pairs away, we're talking about granular microscopic level surgery on your genes. So small. I don't even know how to describe how small on it. It's picture the smallest thing you can imagine than a million times smaller. Yeah, and what you just mentioned was a single letter. So we're talking about like, if you imagine like a letter that you type in a massive essay,

one letter could redefine the entire essay and the meaning of the essay and how it's interpreted. That's basically how I picture what you've just described, right? Which is just insane. Another thing that I think is worth kind of like talking about here is both of us have just mentioned that, you know, this new alpha genome thing can process, one million of these base pair characters at a time.

Now, we're not talking about simple letters. We're talking about like a genetic code. So what really matters is not only what the letters are, but the sequence of these letters. Totally. But not only the sequence of these letters, the different sequences of groupings of these letters. So if you imagine just a straight line, right, of A's, C's, T's and Gs, they usually get bracketed up

into groups of four and sometimes groups of eight. And the reason why they're grouped like this is these different groups end up creating different kinds of proteins. And they also end up regulating how much of that protein gets expressed at one particular condition. So if that sounds familiar to you, it sounds something like hormones, which govern, you know, our mood, our emotions, how tired we are when we're at the gym, or whether we're feeling really hyped up about something, right?

like I am right now. Right. So my hormones are going crazy. My genes are expressing literally different things, right? For an AI model to be able to not only analyze the combination of these letters, but also identify these different groupings and then extrapolate as to how these different groupings are going to interact with each other and then identify the exact grouping that is potentially wrong or abnormal.

And then the exact letter within that grouping is just insane. It needs to run like thousands of simulations to even unassimulation to even unassailable. on what it's doing in the first place. This is a remarkable breakthrough. And it's something that can only happen in this new world of AI, where it's trained on so much data of so many people. And what you start to realize is that the human body, the human genome is built no different than a computer is. It's just a series of lines of code. And if you could understand and interpret those lines of code, you could actually manipulate it and you could change it and you could evaluate it. And this is this totally new world that's unlocked. And I think probably one of the

biggest trends of the next decade is, like, part of this, a lot of this conversation that we have is around the idea eventually reaching transhumanism where we kind of merge with these robots and we kind of become these enhanced versions of humans. But this is, this is deeply human. This is actually allowing us to change and alter the makeup of our bodies inside of ourselves, to make ourselves healthier, to remove diseases, to prevent the process of aging. And I think this, it's just, it's incredible to see. And actually, I wish I knew more. I'd love for us to get one these guests on the show who actually can speak at a length about this because it's fascinating stuff and it's actually working. It's actually like curing people's issues. Yeah, I'm super curious to

see how this applies to personalize medicine, right? Like maybe a decade from now where they just know exactly what Josh is eating. His heart rate is regulated and monitored over time and it knows when to tweak your diet, when to kind of tweak your exercise routine. I think that's going to be a multi-trillion dollar industry. And what's interesting is. is not many people are talking about it because right now we're obsessed with chat boxes and chat dbt acting as our therapist but what happens when it can act as our doctor and not just any doctor like a doctor that knows anything and everything about you just crazy would that be right it's like we talked about the new open-AI hardware device that's always on it's always monitoring you right

it kind of knows what you're doing 24 hours a day well what if you pair that up with a biometric device what if you were whoop with that that was constantly tracking your your health data and you just combine it all into this one large language context version of you. It has your blood data. It has your genomic data. You have a full view of your life and with the ability to alter it using these CRISPR gene modification tools. It's like, it's unbelievable the potential that this has in affecting our lives. It would almost make my insurance bills every month totally worth it. Like I would gladly play 500 to a couple thousand bucks a month if it meant that I for sure can be guaranteed to live an extra 50 years, right? And I live to like 150 or whatever that might be.

Okay, so if what we just described was all at the genetic level, right, it's at the code level, Josh, I'm sure some of the listeners on the show is wondering, okay, well, I'm not going to be getting my genome sequence anytime soon, but is there anything like, you know, at the consumer level that maybe I can, you know, go to my doctor about and AI can help accelerate that? Well, have I got a bit of news for you. Microsoft this week revealed a new, I'm not going to call it an AI model, but it's more of an AI framework, but it's a very cleverly designed AI framework that basically acts as your doctor.

And the takeaway headline is it diagnoses you four times better than the best doctors in the country. I want to let that sink in for a second. So you can go to this. You can go to this AI and you can say, yo. I've got a sore throat. I have no idea what's wrong with me. And within an hour, that AI will know exactly what your issue is

and would have already ordered all your blood work, all your labs and tests, and you'll see him probably the next day, and you'll have the next step being recommended. Before I get into the detail of how this works, Josh, take away, please. What's your reaction? This is incredible. And part of the study, it ran, what, 300 of these, like, tricky hospital cases. And humans were able to solve one of the first of these, like, tricky hospital cases. five of these, this new software stack solved four out of five and spent basically no money

curing these patients. That's a huge deal. I mean, I don't love going to the doctor because I just feel like they're generally stupid and not like, not against doctors. It's just they don't have the information. Inefficient. Exactly. They don't have the context required to properly diagnose things. So half of the time spent there is trying to gather context to properly diagnose what's wrong with you. And this, so with AlphaFold or with the Google's Alpha Project, we kind of had the proactive approach. And now this is the time of illness approach where when you actually are sick, it is able to properly diagnose you. And I think that's unbelievable. So many people hate going to doctors. Not only that, so many people just don't live in proximity to good doctors.

I mean, we're lucky we live in a city that has access to high quality people, but there's a lot of, most of the country doesn't have access to top-tier medical support. And if you could offer this at a fraction of the cost. That changes the health of everybody. Exactly. I'm going to pull up this tweet from Satya Nadala, the CEO of Microsoft, who basically kind of like gets into detail about this new tool. And he mentions this thing here, which I've highlighted, which says it achieves 85.5% diagnostic accuracy, four times that of an experienced doctor. I just want to point out that that means that an experienced doctor had a 20% diagnostic accuracy

before this AI tool. That's so depressing, man. Do you know how insane that is? Like, people spend thousands of dollars, I'm not joking, in the US, per month to get access to some of the best doctors in the world or just even general level doctors, and they're hitting you with a 20% accuracy of your diagnostics,

which would basically theoretically mean that you would need to go to multiple doctors before you can get a certainty as to what the hell is wrong with you, which is just insane and a very inefficient process, as you've just pointed out. But I'm sure people are wondering, like, how the hell does this thing work and what is it based on?

So I'm just going to give like a very loose breakdown. So in the USA, if you want to practice medicine, you need to pass a rigorous test, which is called the United States Medical Licensing Exam, right? And this exam naturally became a benchmark for any AI models that were trying to, advance the medical field as is. So you're probably wondering, you know, how did these AI models perform over time? And the short answer is over three years, these AI models now score a pretty much

perfect score or perfect marks on these tests, right? So you might be thinking, oh, well, great, like, I'll just get treated by an AI model and we should be good to go. The one major flaw of this test, which they're benchmarked against, is that it's all multiple choice. So you don't actually know if the AI model's smart or if it's just kind of working on memory recall, right? So we kind of can't be too entrusting of these AI models until Microsoft revealed this new method, which they train these AI models on, which is known as something called sequential diagnoses. So I'm actually going to pull up this tweet, which describes it really well. He goes, how did they test the AI doctors?

Well, they took 304 real cases from this test and turned them into an interactive game. Here's the setup. Step 1, you, a human doctor or AI, gets a tiny intro like 52-year-old man with fever and breathing problems. That's it. No test results, no detailed history, just like a patient walking into the ER. Step 2. There's a gatekeeper AI that has the full case file but won't tell you anything until you

specifically ask. Step 3. You can do three things from here. You can ask questions, for example, any recent trace. travel, do you have any chest pain? Number two, you could order tests. So like a chest x-ray, a CT scan. And number three, you can make your final diagnosis. For example, I think you have pneumonia, right? And then step four, the gatekeeper then answers the question, but it only reveals what you ask for. So if you don't think to ask about travel history, you won't find out that the patient

just returned from a cave expedition, which is a real case study that they tested for. So the point of this technique is it emulates or it simulates a real-life doctor-patient interaction, where you go in, you have no idea what might be wrong with the patient, but you need to deduce by your own analytical preference what might be wrong. So, you know, you ask a question, you eliminate certain options, and you go down that pecking order. And to be honest, Josh, this sounds like the perfect use case for an AI. If you can just digest the entire medical degree that I've ever spent five to seven years,

of God knows how long, depending on the certain niche medical craft that you're studied in, in a matter of hours, I would trust the AI because it has constant memory recall of all of this. And so much lower of an error rate, it's just when you think about the medical process today, we're so dumb and we seem so stupid. Like we don't do a good job. And our buddy is the most important thing, but it's not even our fault. It's just a lack of knowledge. It's a lack of context.

And applying this knowledge and context through AI is incredible. And it's not to discount doctors, because doctors, they're not going extinct. They're still going to be important. Their role will just kind of change where AI kind of crunches the data, but humans still, maybe they'll ask the awkward questions that the algorithms can't. I mean, they'll be there as the interpersonal relationship in the case that you are diagnosed with horrible diseases like cancer or anything else. They'll hit the breaks on maybe when an AI recommends something that isn't quite right,

because they're still not 100% perfect, right? They still have some form of hallucination. So there still will be the use case for doctors, but man, these doctors now become super, leveraged doctors. And some of the examples that I was reading about this were unbelievable. There's AI at the University of Florida will watch your smile frame by frame and will detect Parkinson's with 88% accuracy, which is crazy just from watching a video of you. Like there's the idea where in a UK lab, it'll sequence newborns for six hours, $200. And then it tells you exactly which baby needs which care at what time. And there's so many amazing applications for this.

Another one that I found interesting because I actually wear an Apple Watch frequently is it'll take Apple Watch data, 8Sleep data, and then ping, your heart rate variability has dropped. And it'll kind of analyze this and say, like, hey, you're most likely to be getting sick in the next 48 hours. And we've kind of seen this a little bit, but this all gets hypercharged. And you get this heightened awareness of your health structure and how to diagnose it. And that, to me, is super powerful. A lot of people hate going to the doctor. It's long, you wait on lines. You never really get the best information. This flips all that on its head. And this is like, man, this is So cool. Dude, I want personalized medicine. That's what I want. Same. And I would pay an in excess amount of money to do that. So let's kind of like highlight a few things that we're getting out of this, right? Let's just pretend that these medical AI models from Google, from Microsoft, get much better over time, which as AI is trending is definitely going to be the case, right? What world do we end up in? It sounds like number one, we're going to get constant, we're going to get a constant stream of

medical feedback day in, day out. You mentioned a few tools there, right? You wear an eye watch. I have an eight sleep in my bedroom, which tracks my heart rate over time. If you wear a Fitbit or a whoop or whatever kind of medical device, basically you can feed that data directly into your own personal AI model that might live on your phone or live on your computer and it's tracking you. It's in constant touch with your real-life human doctor or, dare I say, an AI doctor, which suggests or orders blood work for you or lab tests so it arrives on your doorstep, right? It gets delivered the next day

or for the cost of a Netflix subscription. Now, I've heard a lot of pushback on this being like, ah, you know, the insurance providers are still going to charge you an arm and a leg. And my counter to that is probably not if it means that, number one, humans live for longer

and they can, like, conduct micro-transactions on a number of different things. Can you imagine that? Like, humans live longer, and you and have fewer fatal elements, but more regular normal elements. And if they own the AI kind of like tool or service that is like kind of like diagnosing you,

they could just charge a kind of regular monthly fee. And I bet you they'll end up making more money than an individual cost of someone coming in saying, yo, I've got a headache, right? Maybe twice a year or whatever that might be, right? Number two, I'm really obsessed with what this means for AI models living on your first. own, Josh. I saw that Google announced something where they launched something called, I think it was

Google Gemini 3N. I'm not going to try and pretend to know what the N stands for. And I would love for you to go into a little more detail, Josh, but the TLDR is it's a small enough model that can run on your phone. I think it's between a four and an eight billion parameter model, which compared to the trillion dollar, a trillion parameter models that are being created today at the frontier, that is not comparable, but where it really excels is it's multimodal and it's private to yourself. Think about it. You don't really want to be sharing personalized private medical health with the entire world. Certainly not with Sam Altman, right, where he could probably use that data to sell to pharmaceutical companies and it becomes this whole rigmarole of getting advertised weird things at weird moments.

And I'm like, I don't want to buy that, but maybe I will, right? I want it to be personal. I want it to be private. I don't want anyone to basically, abuse that information, right? I wanted to be private. Running privately on your phone is really important. And to date, we haven't really seen some kind of combination where you can have a model run privately on your phone. And now you have it. What are your thoughts on this, Josh? Because I know from our conversations about Apple running local models on their hardware, you like this idea. I do like this idea. I always think, so I think in the end, cloud will handle most of the compute. And we will end up offloading a lot of that compute to the cloud. But there is.

an interesting use case that Andre kind of describes in his tweet, which is, hey, maybe we could just create local models that do these specific sets of things very well. So he uses a specific example that it doesn't know that William the Conqueror's reign ended September 9th in 1087, but it vaguely recognizes the name and can look up the date. So the idea is that it doesn't know all of the facts about the world, but it has the awareness and the tool set that it can successfully get to answers in a timely manner. And I think this is the big difference between these older small models and these newer small models where they're really focusing on only the data that matters and outsourcing the data that doesn't matter to the cloud. So like the way

Apple designs their phones, they have the ship inside that has the secure enclave. And that's where it stores your password. That's where it stores your biometric data. And it's actually a hardware protected device similar to private keys on a hardware wallet in like if you're storing crypto. And I think applying that to your health data is a really interesting take. And you kind of use these proofs to give that data to certain systems that you want. But a lot of the rudimentary computing, the processing power, the understanding of the data can actually happen on device. So in that sense, it's really cool. And I'm excited for people to lean more into these smaller models that are much more resourceful and much more powerful than they used to be. So I just, I had a question for

you, which is the medical industry is hundreds of billions of dollars, if not trillions of dollars a year. And that's kind of, there's a lot of people making a lot of money from this. And I'm curious your take on what this, what, how is this perceived by those people who are now, like, we're coming in and we're saying, wait, we could do your job four times better than you for a fraction of the cost. Like, what does that look like? Are we actually able to roll this out? Or is the limitation going to be regulation who tries to stop this from happening? Okay, so I don't think regulation will stop it from happening, but I think that the professionals themselves are not going to go down quietly. I have an anecdote to this. So I have a friend who lives in the UK. He's in the tech world,

but his parents are both doctors and they work for the NHS in the UK. They came to him about this news before he found out about this. So before he found out about Microsoft's AI and Google's genome thing, they came to him and said, hey, there's this new AI thing, which is now diagnosing way better than us. And our supervisors have been trialing it at our hospital. And it's been triaging much better than most professionals will be doing. And I asked him how that discussion ended up. And they came to the conclusion that they don't think it's going to replace their jobs, but they do think that they need to become more well acquainted with AI tools, because they think that this medical professional AI interaction is going to have to become more of like a relationship versus a tool that they

just use, right? I was kind of like thinking about how this kind of maps out, and I think it's going to be a re-org. I don't think either of us can kind of predict what the new medical field will look like, but I think it's safe to say it's going to look drastically different than what it looks like right now, 10 years from now, because if I were to guess, I'm going to say robotics are going to be highly precise and really specific in a decade's time, so you could argue that you could go from diagnosis to medical operational surgery, all handled by AI technically. and AI tools and machine operated by AI, right? So the question then becomes, what do these humans do?

I think at least within the next decade, you're going to need still a human supervisor, but I think they're going to be able to, one single professional is going to be able to do a lot more than the average doctor, right? I don't know if you agree with me here, right? No, I do. A second thing that I think this will enable is

a lot of the national health providers of different nations are very beholden to their demand for doctors. But if doctors now see that they could just take an AI tool and go and kind of like diagnose people themselves and scale themselves up just by using these different tools, they might go out into the private field. So my random prediction, but I think the private medical sectors

for each individual nation is going to skyrocket. I think we're going to see more individual investment from different funds into, I basically think bioengineering is actually going to skyrocket much higher than it already has. And the thing that's been holding bioengineering back quite a bit over the last five years is that the ideas are good, but they haven't been quite as good to get to the point of human trials and then like personalized biohacking, right?

But now we're reaching a point where the gaps are going to be filled, the bridges are going to be built, and the world's basically your oyster. And I think now is the time to kind of like reinvest and kind of like orient yourself around all these different tools. Josh, I have a question for you, right? more of a guesstimate question, right? Okay. I just looked up the cost of diagnostics,

medical diagnostics in the US per year. Do you have a guess about what that figure is? Nine figures. It's got to be hundreds of billions of dollars, right? Yep. I maybe, I don't know. Maybe let's go with $150 billion. Okay, not bad.

Times that by four, you're at $450 billion. Okay? So $450 billion just in cost, to diagnose people, and that's at a 20% hit rate, right? That's insane. Now, let's assume these AI models get really good, and they take over 50% of that. That's $225 billion of savings per year. But also, so if you spend $450 billion for a 20% hit rate, that means you're essentially,

if you want to get 100%, you're spending $2 trillion, which is an absurd amount of money, where as now that number is split in half, and you're reaching 80% or higher, That, like, those unit economics are, that shift is insane. That like materially changes the way that healthcare works in this world. Crazy. Absolutely crazy. That is crazy. I mean, it's going to take a while for me to process this. Yeah. And to your point about this being, bioengineering being an important field, I was strolling Twitter as I always do the other day and came upon a tweet that said, I forget the exact phrasing of it, but it was basically like, hey, what is the technology that's kind of like Bitcoin was 15 years ago? Where what is the technology that a few nerds are really

excited about, but the rest of the world doesn't really understand how big it's going to be. And the answer that I concluded after that was bioengineering. I think it's such a slept-on place because we've just kind of come to consensus that, like, healthcare is what it is. And we can't change it because we don't understand it. And it's this very complicated thing. Yeah, and to be involved in it, you must go through a decade of school and millions of dollars of costs for education. And what AI is doing, it's democratizing. Not only that, but it's improving it by orders of magnitude now on a monthly basis. I mean, we're seeing these improvements happen so quickly. And what we're seeing from it is that all of these genomic data, it's really just computer code.

And if we can understand and parse this code and manipulate this code and begin to read and write from our bodies, we have full control over the human experience and the human body. And that, to me, like, over the next 10 years, that is incredibly exciting because that's something that directly affects all of us. And I am like so unbelievably excited to watch this field grow. Let's live to 200. Let's get that, dude. Yeah, Brian Johnson, he's ahead of the curve, man. Yeah, he doesn't want to die.

And I think a lot of other people will very quickly be joining on board as becomes easier. Well, hopefully we won't have to spend $2 million like he did to get that genetic blueprint. It's going to drive the cost. It's going to completely ruin his product, actually. Yeah, well, think about it. I mean, he spends $2 million as version number one. I mean, that's a pretty reasonable starting point because one order of magnitude down from that, you're at $200,000 and suddenly people can afford that.

And then the third generation of that, well, you're at $20,000. And then you're at $2,000. And you could kind of see these like incremental steps down that we've seen with technology and how quickly it could become accessible to so many people. So I'm excited. Brian, thank you for the research. Google, Microsoft, thank you for the research. This is great.

We are stoked. And I'm sure this is going to be a topic we're going to talk a lot more about on the show as we kind of evolved through this process. Cool. Are you going to round us out? Yeah, guys, thank you for listening. You might have noticed David wasn't around this week. He is in Cannes. He is in France at a crypto conference, which is kind of like crypto stuff, whatever.

Not quite as cool as genomics, but whatever. But thank you guys for sticking with us and listening. We really appreciate all the comments, all the feedback. If you enjoyed the show, please remember to like, subscribe, share it with a friend. It really goes a long way when we're getting started to share with people who you also think will be interested. And we'll be back again next week with a lot of more fun AI content. So thank you so much for joining us. Thank you for watching. And we will see you guys soon. See you folks.