I bought palantir at $7. these 8 stocks are the biology version

0:00 Hello everyone. Today we're going to look at eight stocks, eight names companies which I believe are going to change the world in the next 5 to 10 years time. The broader thesis is that in the next few decades biology is going to deliver most of the returns in the stock market. Welcome back to the world's best investing podcast. I went long Palantir at $7 per share. The evidence of that you can find it in my original deep dive written back in 2022 explaining why Palantir was the most important company in the West and then me actively defending the thesis all the way up with the various 50% plus declines and everyone trying to short the stock on the way up. Today as I said, we're going to look at eight tickers in depth. The very least this presentation is going to do for you is make you smarter. If you relate to the AI trade, this presentation is going to enable you to at least understand the next big huge trade early which is the intersection between AI and biology.

0:46 Nothing I say today is financial advice, nor should it be interpreted as such. This is just for educational and entertainment purposes only. Let's get started with today's presentation. Right. So, life extension per token. Many of you guys have heard me talking about this. This is a new AI scaling law which I've coined myself. Essentially, the idea that every time an AI model runs that's trained on biological data, human lifespan is going to be extended. Longer lives, better lives as an emergent feature of this very simple data set which is your ability to read into someone's biology at present, write back into it, and then be able to tell what happens afterwards. What's the end state in someone's biology. This is the very simple key value pairs we'll cover now that will train superhuman AI doctors. In the future, it's actually happening now and this is going to give way to the most valuable subscription service ever in the history of humanity.

1:42 And that is the bottom line of the thesis. Today we have eight tickers. I know many of you guys like me to cover the tickers early. So, we have Hims 10x AI, Nautilus Biotechnology, Butterfly Network which is the next deep dive. Today we're only going to cover it superficially. Recursion Pharmaceuticals, AbCellera, ImmunityBio, and MiNK Therapeutics, Hims, and Tempus AI at the distribution level. It's very important for the read and write function in biology. I'll explain the importance of that now. So, actually read patients so we can see what's happening with patients before and after a treatment. That is the essence for training an AI. And then there's actually deeper components of this value chain. We have Nautilus Biotechnology, which is very important, increasingly so, for the read function at the proteomic level. Butterfly Network seems to have extraordinary technology that can tell the state of specific organs at present, and this is actually quite unique in the market via ultrasound, essentially chips that run ultrasound in situ inside the chip. Then we have Recursion Pharmaceuticals, which is the drug manufacturing machine which builds itself. It's recursive in nature, apparently. AbCellera, which is a company that I've covered a lot.

2:49 Congratulations to you guys that bought in early. I didn't. It's essentially a factory that just prints antibodies and can therefore target any receptor in the human body. By the way, if you don't know anything about biology, you can go and check out my past videos on that. In the course, I have a module in which I break down the first principles framework so that you guys get up to speed quickly without spending years studying biology like I did. And then we have ImmunityBio, which is basically something that jacks up your immune system by binding onto the interleukin 15 receptor. I do believe that will be fundamental in the next 5 to 10 years, specifically so in the next 5 years. So, that's very important to take into account. And then we have MiNK Therapeutics, which is the only company in the world, at least publicly traded, that I know of, that's capitalizing on a third branch of the immune system, which is not the innate, not the adaptive branch, it's something in between that targets the bad guys via what is known as lipid. So, it doesn't use the peptide complex, therefore does not leverage the major histocompatibility complex, and therefore is extraordinary. You can go and check out my deep dive on that company.

3:54 >> >> Now, let's get started with the worldview that I was promising. Biology will, in my view, deliver most of the stock market returns of the next decade, primarily by extending human health span. This is going to be the most valuable, desirable subscription service on Earth because ultimately, as a human, if you don't care about sticking around and in a good state, what do you really care about? If you don't stick around, you can't possibly care about anything.

4:16 Therefore, on a first principles basis, if this does come to life, as I believe it will do, and actually, as I will prove in today's presentation, is coming to life, this is going to be extremely valuable. Semiconductors have been have proven to be extremely valuable, and they're only getting started. As many of you know, I'm an early AMD shareholder since $4.2 per share. At the very least, you have billions of videos of me defending AMD last year when it was when it was trading at $75 per share, saying it's the next AI trillion-dollar AI giant, and so it has been. It's now trading just over 1 trillion. All right, so we have an emerging data value chain that's already extending lifespan per token. This is something that the market is missing. The financials are starting to click, and several key assets were actually severely underpriced. And to me, that's certainly an arbitrage, which which I am exploiting. There are a series of newly started long positions.

5:08 Newly means in the past year, or recently in the past few months, which I will disclose during the presentation. The building block of this new scaling law, which is more lifespan per token every time an AI model runs that's obviously trained on the adequate data, is a single key value pair, as is the case with the rest of AI. The key is someone's biological state before a treatment, then plus the treatment, and the value is the state after. So, if you have a snapshot of someone's biology before a treatment, then you add the exact treatment, and then you see what happens to someone's biology after.

5:40 Enough pairs of that, and AI does become biologically predictive. Now, historically, every major advance in has come from a better read or a better write function in biology. So, as we as we understood biology better and we've gained an ability to then write back into biology, we've cured things. Now, both run on data and compute compute that's recursive in nature. So, you essentially have AI models that are starting to improve themselves. So, each better read actually trains a better write. And we are right now going through revolution in this space, which, you know, is increasingly acknowledged by the wider public, but it's still very niche, right? So, if you tell anyone about this that's not on, you know, in in our sphere of investing and technology, they're going to think you're insane, right? So, you have the key value pair behind predictive biology. As I was saying, it's very important we nail this. You have someone's biological state before captured by the read function, then you have the treatments, which is delivered by the write function. So, we have, you know, something in the write function that's booming right now is peptides, for example. You're just writing back into biology.

6:45 And then you have the state after. Millions of these pairs teach the model the link between key and value. Then, you know, when you have millions of patients on board, for example, and this key value pair is increasingly training an AI model, when a new guy joins the platform, for that guy, the AI is going to be predictive. And that's what's going to change not just medicine. It's not only going to produce the most valuable subscription service on earth, in my opinion, it's also going to change the human condition forever.

7:14 Now, here's where we get into a spotting the asymmetries, you know, the tickets that we were discussing, how good the prediction is. So, how much someone's health span is extended per token, per every time a model runs, depends on how deep the read goes, from how you feel through each organ down to the proteome, epigenome, and genome. By the way, if you've never heard of the proteome, epigenome, and genome, I have lots of videos of that. As I said, we have a module on the course with the first principles framework.

7:41 Let me just illustrate what I mean by a deeper read. How the person feels, how they're sleeping, how much they work out, how their heart is performing. We have that in the wearables today. That is being democratized at scale with Whoop, which I've been a customer and now actually taking a break from it because it was a little bit painstaking. But, you know, Aura stuff like that, this is happening across the board. You have the intelligent mattresses that people are using. That's the very high-level surface biomarkers. Then we have organ performance. This This seems to be some kind of innovation happening inside Butterfly, ticker BFLY. I'll be discussing that in depth in the upcoming deep dive. The proteome is very well done by Nautilus Biotechnology, frankly, in a way which no one else seems to be able to do so at the moment. The thing everyone is using in the industry is mass spectrometry, which we'll be covering now, but it doesn't work to truly capture nuance in the proteome.

8:34 So, this ticker to me is fundamentally interesting and it's to me it seems incredibly undervalued. And then, at the epigenome, RNA, genome level, we have two companies which I think are very interesting, which is Tempus AI, ticker TEM, and then Recursion Pharmaceuticals, ticker RXRX. So, just understand that and I'm not saying these tickers will succeed 100%, but just understand that the deeper the read into biology goes, the better, more predictive, more efficient AI will be at extending human health spans across the board. Now, to test whether this chain is actually viable, it's real and it's not something I'm making up, you have to look at the distribution layer. So, the platforms in direct contact with the customer where read and write actually reach a body. Is this happening across the board or not? Yes, two companies, Hims and Tempus. You guys know I'm a long-time Hims bull. I'm a gigabull as I've been with companies like AMD and Palantir and I continue to be to date.

9:35 These two companies, as I've covered in the past, perform the same function but are coming at this through opposite doors. Both are actually growing, as I will show now, for one simple reason. And that's that data improves outcomes. Whether you believe in the emerging data value chain that I'm going to cover or not, you can't deny the fact that actually the top line of these two companies are growing really fast. Whether AI is truly driving patient outcomes or not, we can debate back and forth. But the fact that these are two data-driven and the top line is growing exponentially for now, just because the intel that the data is generating enables these companies to deliver more value per dollar spent is actually increasingly harder to debate. Right?

10:19 So, this is a hypothesis. There is no guarantee of success in investing. But you have Hims doing this revenue in the last 12 months, 11.3x revenue growth with respect to the start of this timeline. Tempus 1.4 billion in the same period, 5x revenue growth. You will see there is a tentative arbitrage in the market cap. So, Hims 6.7 billion dollars market cap, Tempus AI 14.9 billion dollars. Although there is a disparity in the revenue. This is because the market at present interprets Hims's revenue as being lower quality of that of Tempus. However, as we know in technology, things that come to be very meaningful eventually actually look like toys at the beginning. Hims has been labeled as an erectile dysfunction company, GLP-1 company, whatever pill company. It's actually building an infrastructure that's D2C. It's just a different way of nailing this key value pair. Right? So, Hims Labs picks up data on people before treatment, biomarkers and increasing volume of them, then the treatment that they receive, and they do so via this closed-loop infrastructure, and then the state after. So, on a first principles basis, it's very much training in AI, just like Tempus is.

11:31 Now, there's some interesting qualitative remarks from the co-founder and CEO of Tempus AI in the Q2 2026 earnings report. He said, "They, the customers, don't just want our data. They want access to Lens. They're uploading data. They're building models that remain in Lens, and the business just feels super healthy, super sticky." Algorithm attach rate 45%. That's up from 40% in the latest report. Data licensing revenue growth 36% year over year.

12:02 Now, the company is actually for the first time ever positive GAAP net income. This is because Tempus is becoming a machine that not only delivers value just by selling raw data, it's actually a place where customers are beginning to generatively build their own AI models on top of Tempus's data. So, Tempus is entering this new chapter in which at a marginal cost, it prints AI models that then drive additional incremental value at a marginal cost to their customers inside their infrastructure. So, rather than just a data platform, it's becoming a place where new AI models become bootstrapped by customers via prompts, which is extraordinary.

12:46 Now, that we've covered the top line and we we can sort of broadly agree that something is happening for these two companies to grow so fast, it's interesting to go deeper into the value chain. Because here is where I believe also the case at the top of the value chain is where we will uncover many multi-baggers in the decades to come. This thesis, life extension, health span extension per token, I believe is extremely fertile ground for extraordinary investments as has been the sort of vanilla standalone AI value chain with companies like AMD and Palantir over the past few decades.

13:22 Recursion is a machine that essentially explains why a treatment should work. And Tempus owns the outcomes, right? So, they can read a patient's state before and after the treatment. So, they own a higher level of the same key value per abstraction. Now, they recently closed the deal in which Recursion pays Tempus $42 million over 3 years for data regarding their patients. And Tempus pays $12 million to Recursion Pharmaceuticals over 2 years for data for their TX FM RNA model. So, for a deep comprehension of what's happening inside cells at the RNA level.

14:01 In case you haven't seen my Recursion Pharmaceuticals deep dive, they own roughly 50 petabytes of lab data. They have a factory which processes cells and generates deep low-level biological data of what's happening inside cells. This is hard to do because you have to build out all the physical infrastructure to get the data of what's happening inside cells, and then you have to build the AI models on top of them. Of course, as AI scaling laws continue accelerating, the intelligence built per petabyte of data generated in Recursion's factories continues going up. So, every day that goes by, Recursion, in my view, is actually exponentially harder to replicate. Bottom line is Recursion has a deeper level understanding, lower level understanding, of what's happening inside cells and how the chemistry and drugs interacts with that biology.

14:52 And this deal just shows you a delineation in the data value chain. These two companies are paying for each other's data because they can't replicate that data. So, this is a very strong signal from Q2 2026. Now, what really caught my eye in Q2 2026 for Recursion Pharmaceuticals is that the model seems to be discovering novel biology. Not a hypothesis that someone prompted the model and to gain clarity on. The model came up with something that no specialist actually even thought of apparently.

15:26 This is the result of a novel atlas that Recursion has built by processing over 1 trillion lab-grown neurons and microglial cells. Neurons, as you may know, are essentially the cells that power the central nervous system including the brain, and microglia are the resident immune cells inside the brain. This is mapped together with 17,000 genes that Recursion has essentially looked at. So, this atlas has produced, apparently, a novel neurological target, so something in the brain that can be interacted with to produce to improve patient outcomes that Genentech is now taking to the clinic.

16:04 What's interesting is that this atlas is obviously reusable. This was an emergent property of the atlas, and this is sort of increasingly generalized intelligence such that it would seem that Recursion's model, there is a combination of them, but I think we can speak of these models as one emergent model, came up with something which really no one thought about, according to management's qualitative remarks in the Q2 2026 earnings call. Now, I'm going by what they say because this seems like an extremely capable management team, and I think I increasingly trust them. What's interesting here is the relationship.

16:39 So, we talked about how the growing top line of the top part of the value chain, which is the distribution layer, seems to be powering demand for lower levels of the emerging data value chain. What's interesting, however, is the recursive nature of the relationships there. Meaning, if this thing is coming up with novel biology, does that supercharge the top line itself? Does that open a field? Does that open a field of possibilities such that Tempus and Hims can radically improve patient outcomes in a way which was previously considered impossible.

17:14 This is absolutely the case, and this is actually happening now. Again, we just go back to the idea of more treated patients, more key value pairs, and then that sort of feeds back into a richer Atlas. So, this emerging value chain is extremely real. And as I will explain now, across these series of places, it's actually being tremendously undervalued. >> >> Recursion by itself is an extremely interesting company to me because the engine seems to run at a level of efficiency which the broader pharmaceutical industry cannot match. And as I said, it seems to be recursive, meaning that the infrastructure improves itself. They do have a closed loop where they have a physical factory that generates the data, then the models, and then the models sort of improve the factory itself, and that's recursive in nature.

18:01 But, if you look, for example, at REC 4881, which is essentially a molecule, like I covered this in depth in the deep dive, but it's a molecule that reduces polyps in the gastrointestinal tract. It did so by roughly 43% actually, concretely 43% in phase two. And what's interesting is that this is done at a level of efficiency which the market, the broader market, cannot match. Right? So, the industry norm is roughly thousands of compounds synthesized in order to produce one drug that does something useful in the clinic.

18:36 So, we can see, however, that REC 7735 only required 242 compounds. And just the the recursion average is 330. So, it's a lot lower than the thousands required by the industry across the board. What's incredibly interesting, even more, is that Recursion Pharmaceuticals is doing this while the guidance in terms of OPEX, operating expenditures for 2026, is 40% below the OPEX in 2024. So, the machine is not only outperforming the broader industry, it's not only allegedly discovering novel biology, it's actually doing so in an increasingly agile, lean manner. To me, this is extraordinary and it is very much worth keeping an eye on. However, this is part of this map, obviously, is the whole multi-omics thing. So, here we we we said they processed 17,000 genes.

19:30 That's at the genomic level. They did say in the Q2 2026 earnings call, they do that at the other levels of abstraction that we covered, like epigenetics, proteomics, and so forth. However, Recursion Pharmaceuticals is doing this with what is known as mass spectrometry. And this, until Nautilus Biotechnology came along and actually recently did something great with their new technology, this has been, until recently, state of the art. The problem with mass spectrometry is it can't tell the difference between proteoforms. So, a protein is essentially just a combination of amino acids per electromagnetic attractions.

20:06 The amino acids take on a specific shape. However, mass spectrometry cannot tell between, for example, an OH group, oxygen-hydrogen group attached to a specific location of a protein or another location. And that actually changes the function of that protein variation a lot. Mass spectrometry cannot do that. Nautilus, with iterative mapping, and I've actually made a quite a few videos about this company already, can do it. So, here what we have is, again, an emerging data value chain and we can go deeper. Nautilus can read intact proteins one molecule at a time. It understands perfectly the difference between one proteoform and another. And therefore, it gains a level of nuance that traditionally has been impossible.

20:50 It's unblocking, unlocking a data layer which is foundational to everything in this emerging value chain that we are talking about. If you can get an exhaustive read at the proteomic level, there's nothing that you can't do in the body to mediate any physiological process. Here's what's very interesting about the company. The first assay, the first vertical read, the first read of a small part of the proteome took them 5 years. This was the tau protein. Now, the last one, oncology, which they're moving into now, took 1 year. And now they're looking into months and projecting 20 assays by mid-2028.

21:28 This data layer, I think, is going to be among the most valuable parts of the value chain. Because if you can't read the proteome, you can't really do all that much, and AI actually doesn't fully extend. I mean, you can't play the LEGO puzzle. And therefore, I don't believe you can unblock the whole health span thing. Anyways, there's there's essentially a pathway called AK-1, and it's the same pathway that Recursion 7735 is working with. 7735, remember, is essentially this one that just received the FDA green light. And it was designed in 10 months and with just 242 compounds. However, Recursion doesn't really have an in-depth read of of what's happening in the proteome there. So, how much better could the molecule get if they do gain a read on that on the AK-1 pathway, which, by the way, is the one that I mentioned previously on oncology? The answer is Recursion's machine could get a lot smarter, a lot more predictive, and actually just turn this condition, AK-1, it's essentially a pathway involved in a good number of tumors. It could just turn this thing into background noise. If you as a human have these biomarkers being generated in the background and AI models working for you, and then the right function is seamless and delivered via the distribution layer, be it Hims, be it Tempus, be it Oura, be it be it Whoop, essentially this becomes a nothing. You know, it just disappears into the background, and AI truly becomes predictive for you, and thus extends your health span. So, therefore, Recursion is doing very well. It's already outperforming the broader market, as we're talking about in this slide, but it's doing so with a very thin read of proteomics. So, my bet is, you know, these guys are exchanging money now for data. My prediction is Nautilus essentially and Recursion form another link in the chain very soon.

23:17 Now, we can trace the emerging data value chain along separate things that these companies are doing, which and you know, maybe that data value chain hasn't been formalized yet, but if you look at the biological advances, then it can be. There's a thing called PI3K, which essentially changes stuff in the cell membrane, and that change when it goes, you know, when the when the switch goes uncontrolled, it basically makes the cell proliferate and divide and grow uncontrollably. It does so via the AKT pathway that we talked about. It switches on the mechanism of grow and survive, and this absolutely seems to make cells grow uncontrollably.

23:57 So, we have Nautilus working on the AKT1 proteoform assay, getting an exhaustive read of what's happening there with the proteome. Then we have Recursion Pharmaceuticals with REC 7735, which seems to switch off the jam switch. So, the thing that goes wrong that makes PI3K trigger this uncontrolled growth in the cell. And by the way, it does so with 100x plus more selective efficiency than does anything in the market, which is the reason the molecule is working so well, and it seems to be far less toxic than alternatives in the market. This is all covered in depth. Then, we have Tempus sequencing the H1047R tumors, which seem to be an ultimate expression of this pathway being uncontrolled. And you know, they they do acknowledge that it does emerge from the PIK3CA pathway, which I believe is more specific to solid tumors, but it's essentially a specific version of this pathway that I'm explaining here. Bottom line is these companies are doing things along the same pathway and they're just not connected yet. They're naturally going to be connected because they have modes that to me are clearly defined. Nautilus is very hard to replicate at this stage even though this is pre-commercial and obviously the financial risk in these companies I think is meaningful. Less so as you move up to the distribution layer which is where I think we're seeing meaningful traction at the moment. But this is clearly just an emerging data value chain and it'll goes back to the key value pair. For now the company that I'm most interested in out of these and it doesn't exclude me going long any of these companies that I'm mentioning is Nautilus.

25:37 The iterative mapping technology that they use that I cover in depth in a deep dive no one's close to. Recursion efficiency is unmatched and then Tempus has a strong hold on US hospitals meaning I have clear visibility into who is going to be using the iterative mapping technology that Nautilus is bringing to the world. And they are now starting to commercialize the platform. I covered this in the Q2 2026 earnings reports digest. I believe I didn't make a video. Essentially what's happening with Nautilus biotechnology is they've pivoted away from releasing into the market as generic platform that allows you to read any parts of the proteome you want and they're now focusing on verticals in which they have seen they allegedly claim to have seen real demand from the marketplace and the science is at a place where it could actually unlock meaningful resolution at the proteomic level. So this thing is working and my view is that every condition at some point runs through proteins. I believe that on a first principles basis this is a reality. Of course proteins emerge from the epigenome which tell cells what parts of the genome to read.

26:44 So you could actually very much change the proteome just by getting deep into the genomic layer. But certainly the proteomic layer is how you read where this function is manifesting and perpetuating. And so, this is incredibly important. I believe that every condition runs through proteins. And so, you have Nautilus Biotechnology working on these various conditions that emerge from dysfunction in these proteins. So, you have tau, alpha-synuclein, AKT1, P53, and EGFR, which are the ones in the oncology data value chain, which are very important. Insulin for diabetes, cardiac proteins for heart disease, cytokines for for inflammation. This is where we have the thinnest read at the moment. I don't believe that reading into the genome is a problem. I think RNA we're doing pretty well. Protein amounts with mass spectrometry is actually quite well solved. What we don't have is proteoform resolution. So, we have Nautilus, which at the time of me recording this video, I'm actually I started doing this before market open. Yesterday, it was trading at $170 million. That's the market cap.

27:50 They have a thin balance sheet, but I I have seen management exhibiting a clear ability to push the science forward. Commer- They They've started to commercialize. the first revenue actually came in in Q2 from a foundation that they're working in. I believe it's the tau proteoforms that they're leveraging. And they're running lean, too. So, OPEX is down year over year. There is meaningful financial risk here. So, this company is burning $14 million a quarter and actually expects to raise capital before mid-2027.

28:21 So, there is clear dilution risk. This is very early stage, but to me, this is a very appealing company. Right? So, if management continues executing in a disciplined manner and just continues doing what they've been doing, this valuation for a key components of the value chain to me seems ridiculous. If this works, what are we looking at here? A company valued at least in the tens of billions of dollars, because everyone is going to be pulling from this data layer that no one else really seems to be able to unlock for now.

28:51 >> >> However, you know, I'm long Nautilus, but my allocation there is micro. I do continue to plan to build on it, but my allocation is Hims I have a giga position, Immunity Bio I have a smaller one. It's reasonably sized, but I would say very much pursuing a level of asymmetry in which, you know, the whole biology thing could go up 100,000, 200,000 X, whatever over the next 5 to 10 decades, but it's something that I'd be prepared to have it go to zero. So, it's an additional layer of asymmetry in my portfolio. As I said, there are companies that are also building appealing components of this emerging value chain. We have AbCellera, which has essentially built the discovery platform once and can really just bind onto any target onto the body. I covered that in the video, I believe it was last week. Immunity Bio, I'm long this company as I said, has the ability to reboot the immune system and program it in any specific direction. And then Mik Therapeutics, which I'm not long, but I'm quite fond of that third branch of the immune system, which has also been widely abandoned by the rest of the industry.

29:56 So, this is going to be a long journey. I'm going to continue studying companies, I'm going to continue building positions, and I'll let you guys know. But the bottom line is that every ticker adds key value pairs or makes them sharper. We are gaining a read on biology, which is going to be probably at the atom level resolution, and we're just going to have something which is going to be far more valuable than anything we've seen in the world today. So, you have companies like Amazon and and Netflix and stuff like that being pretty valuable just by letting people buy whatever or letting people watch whatever. This is the equivalent, but on the biology side of things. The risk actually goes up as you go down the value chain. So, in my opinion, the distribution layer makes a lot more sense. If you look at the unit economics of Hims or Tempus AI, they are far better. Hims is actually has been producing producing cash from operations for a long time. Even free cash flow per share went up exponentially before they entered this recent period of investment. So, the risk goes up as you look at smaller components of the value chain, but also does the upside. So, just to recap today's video.

31:03 This superhuman AI that we're building that's going to extend health spans is a function of reading into someone's biology, understanding the treatments at the atom level increasingly and then reading the state after. We have all these tickers that we covered. Butterfly for organs, Nautilus for proteoforms, Tempus for tumor DNA and RNA, Hims is at the distribution level, but also very much getting a read into an increasing level volume of biomarkers via Hims Labs, for example. Grail's Galleri cancer test, which recently received some interesting encouraging data, and the stock went up a lot. Then we have the treatments, you know, we have Recursion Pharmaceuticals discovering novel biology and understanding how to tune drugs to address that biology. XmAb antibodies and T-cell engagers as verified during the Q2 2026 earnings report.

31:55 Immunity Bio for the IL-15 super agonist to bind onto that receptor and reboot natural killer T-cells. And then we have MiNK Therapeutics for off-the-shelf invariant natural killer T-cells, which again I did cover in depth and the science is fascinating. For the state after again, it's very much the same. We have Butterfly, Nautilus, Tempus AI and Hims. Every every one of these tickers adds pairs or makes them sharper. So, the resolution is going up. But again, the lower you go into the value chain, I think risk goes up. Eventually, if I had to bet, the most extreme rewards will come from the lowest level participants in the value chain that also produce unit economics that no one else can replicate. A candidate to that tentatively in my opinion is Recursion Pharmaceuticals. Very much and this is also the case with Nautilus, although as I said, risk goes up. But everything is going to emerge from the proteomic layer. So if Nautilus is the single dominant player in that, the rewards can be extreme, although again, it can go to zero. So I am certainly allocating accordingly. But if you look at Recursion and its ability to uncover novel biology, that's extremely interesting as well. So I wouldn't discard a company because it's lower down in the value chain, but certainly I would bear in mind that the risk does go up.

33:18 So that's it for today's video, quite a dense one. lots of information to put together. You have the written form available in the blog. You can see the link there on the screen. My positions are Hims, ImmunityBio, and Nautilus Biotechnology to date. And as I said, I'm long these companies in essentially logarithmically declining allocation. I'm very comfortable making a massive bet on Hims. I've been building that for a long time. And then I'm comfortable making bets on ImmunityBio and Nautilus, but in decreasing levels of allocation to the point that Nautilus is essentially a micro allocation in my portfolio that nonetheless, I do plan to make bigger.

34:06 All right, guys, giga video today, lots of information. I hope that at the very least this video has transformed your worldview and opened your eyes to an emerging value chain which I as I said, I think it's going to give birth to the most valuable subscription service on earth. A big candidate to that by the way in my opinion is Hims. So thank you very much for joining me today. As always, if you enjoyed the deep dive, could you please like and subscribe and share this with someone else whom you think will enjoy it. These deep dives updates videos are for free. And so, the only way this grows is with your help.

34:35 So, thank you very much in advance. Take care, and until next time.