Jose Luis Ricón Fernández de la Puente is the author of the popular blog Nintil, and Head of Theory at Retro Biosciences. Jose is a prolific blogger, covering a wide breadth of topics across economics, philosophy, progress studies, science funding, and of course longevity. His writing has been published in a16z Future, Works in Progress and by the Adam Smith Institute, and his writing previously won him a fellowship with Emergent Ventures, Tyler Cowen’s competitive program for intellectually ambitious projects.
In this interview, you’ll hear how insightful Jose is about deeply technical topics in biology, and you’ll see why Retro Bio was eager to bring him on as their head of theory (the only role of its kind in the entire biotech industry).
Our conversation is wide ranging, spanning a deep dive on Retro’s work to replace and engineer microglia to rejuvenate the brain and how our cells have the ability to turn back the aging clock but choose not to. We also covered the technological stagnation and why biological engineering is the new frontier of progress, as well as philosophical topics like transhumanism and how a future of total biological control might impact our values and way of life.
Retro Biosciences was seeded with $180M by OpenAI CEO Sam Altman to develop therapies to prevent and reverse age-related disease, and is widely recognized as one of the leading AI for longevity companies. Previously, we hosted Rico Meinl, the head of Applied AI at Retro, so make sure to give that episode a listen as well.
Watch on YouTube. Listen on Spotify or Apple Podcasts.
Chapter Markers
2:49 What is aging & why cells have a tough choice to make
9:02 When cells choose to reverse aging themselves
12:43 Cellular vs Organismal Aging & the magic wand experiment
22:40 How reprogramming plays a role in DNA damage repair
25:42 Do we already know how to cure aging? FOXO3!
28:37 How to cut through the complexity of interconnected biology
32:32 Why transcription factors are so great for intervening
36:49 Does a rejuvenation program exist already in the genome
38:51 Michael Levin: from thinking in terms of genes to morphogenesis
48:14 Tech stagnation and why physics is cooked
55:03 Why doesn’t the world look more futuristic
57:26 Transhumanism & asking ourselves what we want out of life
1:05:34 Do we need war for technological progress
1:09:31 Government role in science funding
1:15:06 How Jose became the Head of Theory at Retro
1:24:53 How AI might put software engineers out of a job, and push them towards biotech
1:27:28 What it takes to get a flywheel in biotech
1:29:04 Rejuvenation vs Prevention
1:34:23 Aging is the coolest hardest problem to work on
1:36:09 What does it take to cure aging
1:42:25 Delivery mechanisms for genetic therapies
1:48:53 Retro’s work to replace microglia and engineer them outside the body
Transcript
0:00 Intro
[Montage]
Daniel 00:00:54
Welcome to the Free Radicals podcast, where we interview the scientists and builders working to dramatically extend human lifespan and bring about a sci-fi future where humanity has full control over biology.
Today’s guest is Jose Luis Ricon Fernandez de la Puente, author of the popular blog Nintil and Head of Theory at Retro Biosciences. Jose is a prolific blogger covering a wide breadth of topics, including economics, philosophy, progress studies, science funding, and longevity.
His writing has been published in A16Z Future, Works in Progress, and by the Adam Smith Institute. His work previously earned him a fellowship with Emergent Ventures, Tyler Cowen’s competitive program for intellectually ambitious projects.
In this interview, you’ll hear how insightful Jose is about deeply technical topics in biology. You’ll also see why Retro is eager to bring him on as their Head of Theory, the only role of its kind in the entire biotech industry.
Our conversation is wide-ranging, spanning a deep dive on Retro’s work to replace and engineer microglia to rejuvenate the brain and how our cells have the ability to turn back the aging clock but choose not to.
We also covered technological stagnation and why biological engineering is the new frontier of progress, as well as philosophical topics like transhumanism and how a future of total biological control might impact our values and way of life.
As context, Retro Biosciences was seeded with $180 million by OpenAI CEO Sam Altman to develop therapies to prevent and reverse age-related diseases. They are widely recognized as one of the leading AI-for-longevity companies.
Last week, we hosted Rico Meinl, the Head of Applied AI at Retro, so make sure to give that episode a listen as well. I am your host, Daniel Shur, and my co-host, though he couldn’t make it for this episode, is Eric Dai.
I hope you enjoy this episode of the Free Radicals podcast. Let’s start with an easy one: What’s aging?
2:49 What is aging & why cells have a tough choice to make
Jose 00:02:51
One definition of aging I like to use is a decline from a state of fitness. If you look at the process that gets you from age 30 to 80, you can see that it begins even before you are born. It is a slow accumulation of damage.
In parallel to that, there is also the process of development. You are building your body and becoming more than just a single cell.
From a biological perspective, you could say aging is the accumulation of cells falling apart due to entropy. In that sense, aging is the same thing that happens to a building or a car. The car ages as it falls apart.
The interesting part is that, unlike those systems, our cells are constantly trying to fight entropy. They are trying to repair themselves all the time. We get damaged and we fix ourselves.
From that process, we eventually get diseases and the phenotypes of aging, such as the loss of muscle mass or hearing, which end up resulting in death.
It doesn’t necessarily have to be this way. Many species don’t seem to age much. Axolotls and naked mole rats still die, but they don’t seem to decline like we do.
Some cells in culture, like induced pluripotent stem cells, are able to stay in culture while keeping a similar state for a very long time—much longer than normal cells. If you can do more repair than you receive in damage, you can stay unaged for longer.
Daniel 00:04:44
One of the things you’ve argued on your blog is that DNA damage is central to aging. Can you talk about what that process is on a cellular level?
Jose 00:04:54
All cells in your body have roughly the same DNA. It is constantly floating inside the cell across different chromosomes. Sometimes the double helix gets a nick on one side, or you get a double-strand break.
This happens from UV radiation hitting your skin or from toxins like cigarette smoke. The cells have to repair that damage.
When that damage to the DNA causes the way the DNA is folded to change slightly, it accumulates over time into aging. In a way, DNA damage is what makes the clock tick.
DNA damage is unavoidable. You can repair it faster or slower, and better or worse. Our cells can actually do more repair than their baseline level.
Many of the things people point to for extending lifespan, like calorie restriction or rapamycin, ultimately upregulate DNA damage repair. Across every species, better damage repair boosts lifespan.
Even in species that live longer than humans, like bowhead whales, they repair damage much better. There is a strong correlation: the longer you live, the better damage repair you do.
Intuitively, it makes sense. If you get this damage and the DNA has to be repaired, it acts a bit like a scar. There’s another theory on top of that put forth by David Sinclair and Lenny Guarente.
They argued that the very same proteins involved in DNA damage repair, like sirtuins or the polycomb repressive complex, are also involved in keeping the chromatin in its proper state.
The cell has to choose between repairing damage or preventing aging. The more damage there is, the faster the clock ticks.
If there were no damage, those proteins could stay in that healthy state. But because there is damage, they have to go do their job. This was mostly found in yeast, but in humans, there seems to be evidence of a similar process. Those chromatin support proteins move to different regions when there is damage.
Daniel 00:07:24
You mentioned a really interesting trade-off. When DNA damage occurs, the cell needs to decide how to allocate its energy between repairing the DNA itself—avoiding a mutation—versus correcting the chromatin structure.
Is it right that if there’s a lot of DNA damage, the cell might focus on correcting the DNA damage itself, causing the chromatin structure to degrade?
Jose 00:07:53
You can think of it as short-term gain for long-term pain. It is unclear why we evolved this way. There are theories in yeast regarding nutrient sensing, but it remains ambiguous because you can have both.
Another argument compares this to sleep. Why do we sleep? From an evolutionary perspective, it is risky; you have to shut down the entire system for hours. Perhaps you need to shut everything down to perform certain maintenance.
Maybe the reason we cannot stay young forever is that entering a rejuvenation or chromatin-fixing mode prevents simultaneous damage repair. This might make an organism highly vulnerable.
This could explain why species that live longer are typically those that are not heavily predated; they can afford to spend more time on repair. It is puzzling because our cells have the capacity to roll back their clocks, but they often choose not to.
Daniel 00:09:01
Tell me more about that. When do our cells choose to turn back the clock?
9:02 When cells choose to reverse aging themselves
Jose 00:09:05
The classic example is during embryogenesis. We have the egg and the sperm. Eggs are created at the beginning of a woman’s life and stay relatively stable. Sperm, however, accumulates more mutations and ages over time.
When they combine, that accumulated age is wiped out. This process, which some papers call “ground zero,” is where aging truly begins. Shinya Yamanaka discovered that certain factors involved in embryogenesis can also revert any cell in your body. You can turn on that program and essentially wipe out the age of those cells.
Other cells can do this to some extent as well. If you remove a chunk of the liver, it regenerates. The regenerated area actually behaves and appears younger.
We see this in other organisms like planaria. These are strange worms that, if chopped into pieces, can grow each chunk into a whole new worm. This regeneration seems to bring rejuvenation with it.
There appears to be a connection between regeneration, rejuvenation, and cancer resistance. Beyond that, in vitro experiments show that the niche where you place a cell can rejuvenate it. Fibroblasts, which are skin cells, seem to de-age when placed in a more elastic or youthful niche.
A similar mechanism occurs in the intestine. The intestinal crypts contain stem cells that constantly produce new cells to regenerate the lining, which is damaged whenever we eat.
When damaged, some of these cells can transdifferentiate or dedifferentiate. We see this more clearly in axolotls. If you remove their limb, the cells at the site dedifferentiate into a stem-like state. They then regrow the entire arm, and that new arm appears younger.
Daniel 00:11:34
This connection between regeneration and rejuvenation is very interesting. For something to be able to regenerate, you essentially need cells to change their identity.
Jose 00:11:47
Right.
Daniel 00:11:47
Cells need to sense what they need to turn into, and then they transform. Rejuvenation feels similar because as we get older, many of our cells lose their identity. We need to restore it. Is that the right way to think about it?
Jose 00:12:00
Yes. If you think of DNA damage as making the clock tick, then aging itself is the state of the chromatin. It is a matter of how that chromatin is folded.
When you look at patterns of gene expression in old cells, they are doing less of what they should. Fibroblasts in your skin produce less collagen as you age, and cells in your liver produce less albumin. They are failing to perform their specific jobs.
If you express transcription factors for that specific cell type, they appear rejuvenated. In that sense, the phenotypes of aging are essentially a loss of identity for that cell type.
12:43 Cellular vs Organismal Aging & the magic wand experiment
Daniel 00:12:44
When we think about aging, there seem to be multiple levels. There is cellular aging, which we just discussed, but then there is also the aging of the entire organism. I think of things like atherosclerosis, which is a higher-order natural occurrence. How do you think about these different levels?
Jose 00:13:04
Classical aging theory—focusing on pathways like IGF1, AMPK, rapamycin, and mTOR—was largely studied in unicellular life like yeast. Those pathways translate all the way to humans, but we also face emergent properties that arise when many cells work together.
This leads to issues like cancer, senescence, and atherosclerosis. C. elegans worms, which are commonly used in research, do not have stem cells, senescent cells, or cancer. As a result, they do not suffer from heart disease. Their aging is almost purely cellular.
In humans, we have these emergent forms of aging. In a recent blog post, I proposed a thought experiment called the “epigenetic magic wand.” Suppose you wave a wand and every cell in your body becomes 19 years old. Would you live forever, or would you only live for another few decades?
The answer is unclear. Plaques might still accumulate in your arteries. You could suffer a hydraulic failure where blood flow stops, resulting in a stroke and death, even if your brain and other organs are cellularly young.
We are still trying to understand how much cellular rejuvenation contributes to organismal rejuvenation. As a heuristic, if a certain type of damage already accumulates in children, it probably won’t be reverted simply by rejuvenating cells. You would have to remove those deposits directly.
Daniel 00:14:42
And that is potentially the case with atherosclerosis.
Jose 00:14:45
I’m not an expert in atherosclerosis in teenagers, but I know that people with familial hypercholesterolemia get plaques very early on. I think of these things as a balance: damage is constantly being produced and removed. If the rate of removal or fixing is higher than the rate of generation, you are in good shape.
For some diseases, the rates of degeneration are very high. In children with familial hypercholesterolemia, they will develop issues earlier. For most people, the rate is lower, so they may not develop plaques.
However, for some things, damage is permanent. If you lose your teeth as you age, they don’t grow back. You have to do something specific about it. The hair cells in your cochlea that allow you to hear eventually die. Even if you were to rejuvenate everything else, if those cells are gone, there is nothing left to rejuvenate. You have to put them back or create them.
The same applies to neurons. When they die, you aren’t making new ones all the time, so they won’t come back. Even the lens in your eye doesn’t turn over much. It is a protein structure that is made when you are born and stays there. If it gets damaged, that’s it.
There are many long-lived proteins, like elastin in the skin, that may have to be addressed separately from cells. We don’t know yet if rejuvenating an entire tissue will allow the cells to fix the structures in between. It might happen to some extent, but maybe not 100%.
Daniel 00:16:25
What do you think the priority should be? Should we be focused on curing cellular aging through things like reprogramming, or should we focus on fixing macroscopic damage and replacing missing cells?
Jose 00:16:42
We should focus on everything; there is room to do it all. At Retro, we are pushing a strategy of cell replacement while also trying to fix things in situ.
The field of longevity recently found its “this is it” moment with reprogramming. It is the one thing that seemingly and robustly reverses age. Many things can potentially slow down aging, but very few things actually reverse it in a way that remains after you stop the treatment.
In the past, people talked about antioxidants and telomeres, but this time is different. Those things never did what reprogramming does. Given that we know how to do it, we should focus on it now. Pragmatically, you can test these things in vitro and see if they are working.
Other things might be harder. For atherosclerosis, there are already companies working on removing plaques and damage. At Retro, we work on what we do because we saw an opportunity to impact disease by targeting aging biology in a way that no one else was.
We didn’t want to just make another marginal benefit. If we can replace all of your blood cells or the microglia in your brain, that is a unique contribution. Someone should do it.
18:31 What is reprogramming
Daniel 00:18:32
We’ve brought up reprogramming a bunch of times. For people who don’t know, can you explain what reprogramming is and why it is so amazing?
Jose 00:18:40
Reprogramming typically refers to Yamanaka or OSKM-based reprogramming. You take a normal cell, like a skin cell, and turn it into an induced pluripotent stem cell. This is similar to an embryonic stem cell; it is a cell that can become almost any other cell type.
The interesting part is that it also has its age wiped out. You can take a skin cell from an old person and a young person, reprogram them, and the results are very similar. The damage is gone.
The mutations remain, but cells have a remarkable ability to work around DNA mutations to some extent. This de-aging of the cells is a genuine bounce back. It stays that way. It’s real.
Daniel 00:19:40
You can even grow a whole new organism from that cell, and they will be young and then age like a normal organism.
Jose 00:19:46
That goes back to the early experiments by John Gurdon. In reprogramming, you add molecules or transcription factors to cells to get stem cells. Before that, experiments showed you could take a nucleus from the skin of an old mouse or a frog and put it into an egg.
This is essentially cloning. Dolly the sheep, for example, came from one such nucleus. Dolly wasn’t born old, so we knew that something happened to wipe out the age of the nucleus. Yamanaka found a way to make that process scalable and possible for other cell types without having to use eggs.
Daniel 00:20:33
The amazing thing with reprogramming is that we can turn a specialized cell into a stem cell, which can then turn into anything. It turns the age back to zero. This is really what spurred the newfound interest and funding in the longevity space, because it is so impressive and real.
Jose 00:20:54
We have known how to make induced pluripotent stem cells (IPSCs) and understand these processes for a long time. The idea of taking young cells made from IPSCs and putting them back into people is called regenerative medicine. This concept existed long before companies like Altos, Calico, and other aging-focused startups.
The field became much more exciting recently due to partial reprogramming. If you were to fully turn on these factors in a mouse, for example, the mouse would develop tumors and die. You don’t want to create stem cells of age zero; you want to maintain the cell’s identity while making it younger.
Alejandro Ocampo, one of our advisors and a pioneer in the field, found that if you express these factors briefly and then stop, you can achieve a rejuvenation effect without the cell losing its original identity. This suggested we could potentially apply this to all cells in vivo.
Companies like New Limit and ourselves began searching for safer ways to reprogram various tissues and cells once it was shown to be doable. Prior to reprogramming, there was no way to reverse aging. We have discussed slowing aging for decades through methods like autophagy, rapamycin, or genetic knockouts in worms, but reversing it was the “wow” moment that created all this excitement.
22:40 How reprogramming plays a role in DNA damage repair
Daniel 00:22:41
I want to understand more about what is happening with reprogramming and how it relates to DNA damage repair. With reprogramming, you are restructuring the chromatin to express the right genes for a specific cell identity. But then there is this other piece involving damage repair.
Jose 00:23:04
Mechanistically, it is still unclear exactly how reprogramming works. You can think of these factors as keys that enter the DNA to turn on a specific program. Instead of having to turn on 100 genes individually, you can turn a single key.
For example, turning on NF-kappaB activates the inflammation program. Turning on Oct4, Sox2, and Klf4 activates the pluripotency program, which shuts down the somatic program and establishes pluripotency.
During this process, there is a drastic increase in DNA damage repair capacity and an initial burst in mitochondrial function. IPSCs perform at least an order of magnitude more DNA damage repair across every pathway than somatic cells.
Is this DNA damage repair what causes rejuvenation? It is possible that if you induce enough repair, the cell can catch up and return to a youthful state, but this remains unproven. There is likely also something involved with reopening and loosening chromatin that was previously closed. We don’t yet know the exact sequence of events.
Daniel 00:24:33
Is it correct to assume IPSCs have higher damage repair because they are embryonic cells about to divide many times, which is stressful for the genome? Is it increasing its repair ability in preparation for all that mitosis?
Jose 00:24:51
Not quite. It is difficult to say because IPSCs inherently want to divide. If you take normal cells in culture that are also dividing every 20 to 24 hours, they still age. IPSCs do not age in vitro in the same way.
While their chromosomes can eventually become unstable, you can keep passaging them much longer than normal cells. I once tried to find the longest someone has kept an IPSC in culture, and it was several years.
The key lesson is that cells already possess the machinery for this repair. We didn’t have to edit any genes; the capacity is already there. The challenge is simply learning how to switch it on.
25:42 Do we already know how to cure aging? FOXO3!
Daniel 00:25:43
Sometimes when I look through longevity literature, it seems so complicated that we understand nothing. But then I find something that makes me wonder if it’s actually simpler than we realize.
You mentioned that we already know some longevity genes, like Foxo3, which extends the lifespan of various organisms. Is the primary challenge simply figuring out the right way to activate Foxo3 in humans?
Jose 00:26:11
Foxo3, also known as DAF-16 in worms, is like a key that turns on programs for DNA damage repair, autophagy, and antioxidant responses. Instead of targeting many individual genes, you can activate Foxo3 to achieve these benefits.
In humans, genetic studies show that individuals with specific Foxo3 SNPs tend to live longer because they have more active Foxo3. There was a recent paper from China by Guanghui Liu, who was a student of Juan Carlos Izpisua Belmonte—a fellow Spaniard in the longevity field who is now at Altos and was also Alejandro Ocampo’s PI.
They overexpressed Foxo3 in mesenchymal stem cells and injected them into monkeys. The monkeys appeared phenotypically younger, though not necessarily intrinsically de-aged.
If you were to maximally activate Foxo3, it might reverse aging, but there are likely more variables at play. Researchers have found that overexpressing Foxo3 in mice can lead to muscle atrophy because excessive autophagy causes the muscles to shrink. You could have maximal Foxo3 activity and still die.
However, if you had to choose one gene to increase throughout the body to slow or reverse aging, Foxo3 is the candidate that most experts would agree on.
Daniel 00:27:57
There is a process in our cells effectively trying to prevent the damage of aging through DNA repair. These processes maintain the correct genome and chromatin structure. We need to find ways to turn that up and manage any side effects that occur.
Jose 00:28:18
To be clear, there isn’t just DNA damage in the nucleus; there is also damage to the mitochondrial DNA, which causes them to function less effectively. We also see the accumulation of protein aggregates, and autophagy is another process that works less well with age.
DNA damage is a major component, but everything is connected. One of my pet peeves in biology is the idea that if you fix one thing, you fix others, and if you break one thing, you break others. It becomes difficult to determine what is causing what. To me, the right question is where the best point to intervene is, not simply what the cause is.
28:37 How to cut through the complexity of interconnected biology
Daniel 00:28:49
A lot of people, when they see how everything is connected, throw their hands up and say it is impossible to isolate anything. But you make the point that a hierarchy still exists and we can give primacy to certain things. What do you prioritize in aging?
Jose 00:29:08
DNA is a primary one. Consider a heuristic: the longer something remains unchanged, the more important it is. My analogy here is gravity. Gravity extends everywhere, so everything in the universe is attracting everything else.
While we can still use Newtonian mechanics to simplify things, consider the solar system. You can think of the Earth orbiting the sun, but that is just a mental model. You could also imagine the Earth as fixed with the sun orbiting it. The math is the same, but it makes more sense to have a hierarchy where the largest structures are most important. Galaxies function in a similar way.
This applies to turnover rates. If you have a mutation in your DNA, it is essentially permanent unless the cell itself is removed. In contrast, proteins turn over constantly. They are produced and cleared easily.
For example, liver fibrosis or a scar is constantly being created and destroyed. It is not a static mass of collagen; it is being added and removed continuously. This gives you a clue that if you alter the balance of cells involved in that process, the fibrosis would go away because the cells are capable of removing it.
The same is true for the chromatin state. For a cell to function healthily, it must have specific regions of chromatin open and others closed. This is determined by epigenetic marks, such as methylation groups attached to the DNA and histones.
Nucleosomes are the structures around which DNA is wrapped. Depending on the state of the histone tails, the DNA is more or less folded. These markers are constantly being added and removed stochastically.
Imagine a stretch of DNA for a fibroblast containing the collagen gene, COL1A1. The cell makes collagen because that area of DNA is open, allowing RNA polymerase to transcribe it. It remains open because of specific histone marks and a lack of methylation.
Transcription factors bind to keep it open, but they only stay for a few seconds at a time. They are not covalently bound; they move back and forth via hydrogen bonds. Histone marks are added and removed by histone deacetylases, and DNA methylation is managed by DNMT enzymes.
The epigenome is turning over, but it does so very slowly. If you alter the epigenome, everything else downstream conforms to it, rather than the other way around.
Conversely, if you put one broken protein into a cell and wait, it will eventually be degraded and replaced. It doesn’t matter as much in the long run. DNA and chromatin are the primary drivers of everything downstream, with the exception of proteins that do not turn over, like those in the crystalline of the eye.
32:32 Why transcription factors are so great for intervening
Daniel 00:32:33
Everything is connected. We need the right chromatin structure for the cell to have the correct identity. That structure leads to the translation of the right proteins, but you still need the right transcription factors to exist in the cell to bind to those chromatin sites.
I think about the path dependency of aging. Eventually, a cell could get so unhealthy that even if we restore the right chromatin structure, you might not be able to recover the correct cell identity.
Jose 00:33:09
If you fix the chromatin structure or these transcription factors, you have to consider where they come from. They are produced from open DNA being read. If you were to fix the chromatin itself, you would likely address everything else as well.
Cells exist in an equilibrium state. If you compare a young cell to an old cell, they differ in many ways. The metabolites, protein abundance, cell shape, stiffness, and chromatin are all different.
This is all part of an equilibrium that slowly shifts toward an aged state over time. The question then becomes how to shift it back to youth. In principle, there are many ways to accomplish this. In fact, there are many ways to reprogram a cell without using Yamanaka factors.
If aging is characterized by a loss of identity for a specific cell type, you can try to reinforce that identity. For example, hepatocytes in the liver are defined by a transcription factor called HNF4alpha.
A study in mice showed that expressing HNF4alpha in the liver makes the organ younger. Fibrosis disappears and it becomes more functional. You can achieve these effects without Yamanaka factors simply by helping a liver cell be more of a liver cell.
In that case, transcription factors have great causal power to rejuvenate the cell. In contrast, a random protein like albumin does nothing if you introduce it to a cell; it just makes more albumin.
Transcription factors are the programming code for cells. While there are feedback loops, TFs are the tools that are truly able to shift a cell’s state to other places.
Daniel 00:35:36
Is that because TFs act at the epigenetic level?
Jose 00:35:40
It is because they act in a coordinated way to shift entire programs across the cell. When a cell undergoes a process like inflammation and needs to produce cytokines like IL6, IL1, and TNF alpha, it would be inefficient to build separate regulatory elements for every single component.
Instead, you have programs where you turn one thing on. Transcription factors bind to specific motifs in the DNA—almost like a CRISPR guide—to turn on those genes. It is a very clean process.
This means you can identify which genes are controlled by a specific TF by finding those motifs in the DNA. You are switching on an entire genomic program.
If you wanted to activate 100 genes with CRISPR, you would naively need 100 CRISPR guides, which is too much work. With one TF, you can turn everything on at once.
36:49 Does a rejuvenation program exist already in the genome
Daniel 00:36:51
Do you think a rejuvenation program exists naturally—a set of transcription factors that will turn on the program we care about—or will we need a synthetic transcription factor to drive a new kind of program?
Jose 00:37:10
It depends. We seem to have a regeneration program that carries some de-aging effects. Normal cells don’t perform Yamanaka reprogramming on their own; it typically only happens in the uterus in a very controlled environment for the egg and sperm. We are forcing it to happen elsewhere.
In principle, we could force a program from the liver to happen somewhere else because the DNA is the same everywhere. The transcription factors we have could work, but some are short-lived or ineffective when introduced to a cell.
This is where modifying TFs becomes interesting. Instead of using four Yamanaka factors, you could modify them to be more potent so that you only need two or three.
You can either combine existing TFs or use AI to make the ones we have more potent. These proteins have specific domains; we know what makes them enter or exit the nucleus.
We can rationally design them, and with AI, we can go even further. You can use existing factors for a proof of concept and then tweak and improve them when creating a final medicine for humans.
38:51 Michael Levin: from thinking in terms of genes to morphogenesis
Daniel 00:38:53
You brought up planaria earlier, and I’ve noticed several references to Michael Levin on your blog. Can you talk about how Michael Levin influenced your thinking about biology?
Jose 00:39:02
I think I discovered Michael Levin before everyone else. Perhaps people got into his work because of me.
Daniel 00:39:08
Now even normies are into Michael Levin.
Jose 00:39:10
I discovered Michael Levin on YouTube a while back and was very impressed. I suggested his work to Allison Duettmann at the Foresight Institute, and it spread from there.
Michael Levin thinks in terms of higher-level structures rather than just genes. Initially, he focused on bioelectricity, which led to the ideas of morphogenesis and morphostasis. He views cells as existing in a state where they react to and adjust their environment in a sort of dance. When that dance is broken, you get aging.
Levin points out several experiments that reframe how we think about biological systems. One way of thinking is to look for a broken gene to fix or determine which genes to overexpress. However, Levin highlights many phenomena that involve no genetic changes at all.
For instance, if you take a cancer cell and put it into an egg, you can grow a healthy mouse. The mutations are still there, but the cancer disappears. While the mouse may eventually develop cancer because the mutation “primes the pump,” the outcome is not deterministic. Destiny is not written in the genes.
He also points to the strange nature of limb regeneration in lizards and axolotls. When they lose a limb, it grows back. How does the organism know to grow exactly five fingers or reach a specific length? Where is that information encoded?
While the genome created the embryo, using it to explain regeneration is like trying to use quantum mechanics to do mechanical engineering. It is technically true, but not useful. We need to look at a higher level of structure, specifically the signaling between cells.
This takes us back to development—how one cell becomes an entire being. Cells signal to each other to determine their position, whether they are in the head or the legs. From this perspective, aging is a disorder where cells lose coordination.
Levin recently published a paper describing aging as a “lost astronaut.” Initially, every cell is working together on the project of building and developing the being. Once development is finished, that shared directionality is lost.
Challenging the primacy of genes allows us to think beyond genetics when fixing biological systems. This is important because gene therapy, while powerful, cannot yet be delivered everywhere. Working at the cellular level offers other paths.
Take fibrosis, for example. If you ask a group of scientists how to cure it, they will offer different ideas based on their approach. One might suggest a GWAS to see who is prone to fibrosis, or use single-cell RNA sequencing to find a gene to inhibit.
Another way, inspired by systems biology experts like Uri Alon, is to look at cell turnover. This shifts the focus from genes to cells. You could potentially cure the condition by depleting myofibroblasts, which produce collagen, or by boosting the activity of macrophages.
These approaches don’t touch the genes; they target the cells. We see similar systemic perturbations in experiments where young blood is exchanged for old blood, or in bone marrow transplants. These methods change the system without having to edit genes one by one. It opens up new frontiers of thinking.
Daniel 00:43:47
The way I initially learned biology was through mechanisms and complicated flowcharts of protein activations. While those are extremely useful, they don’t provide an abstraction layer that helps you theorize more broadly about biology.
Jose 00:44:08
The genetic view is powerful and easy to explain. In many diseases, if everyone with a specific mutation gets the disease, you can trace that chain step by step. That is very clean and satisfying to the human brain, but aging isn’t like that.
There is no single gene you can turn on to live forever. Biology is a “hot mess” compared to computer science, math, or physics, where entities and laws are crisply defined. In biology, things are rarely that clean.
You can do two things: you can try to find small areas where you can impose a model, or you can give up and decide we need AI to solve everything. Alternatively, you can decide that we don’t need to understand every small detail.
We can build another level of abstraction on top and observe what is happening there. We don’t always need full mechanistic evidence for everything.
It is similar to how we don’t need to know how every individual molecule of water is bouncing around to see that a pot of water is boiling. We can see the bubbles and understand the state of the system without knowing the exact path of every molecule.
Daniel 00:45:51
Levin emphasizes that one of the unique aspects of biology is the nature of repair. When you’re repairing a car, the vehicle is inert; you’re simply trying to fix it.
But when you’re trying to repair an organism, you’re dealing with individual agents at every level. Each cell and each tissue is trying to accomplish its own goals and will respond to your actions.
Levin frames it as thinking of the body as an alien we’re trying to communicate with. How do we give it the right goal?
Jose 00:46:24
In the car analogy, a vehicle takes damage passively until it simply breaks. Cells experience random damage as well, but they don’t act randomly.
If a human being is punched, they react in a predictable way to protect themselves regardless of where the damage came from. Cells do the same, using identical mechanisms. Even though damage is entropic, cells react predictably.
They can also enter states that resemble cellular trauma. Much like people with trauma who are locked into self-harming behaviors, aged cells suppress the mechanisms that would fix damage.
Chronic activation of inflammatory pathways, for instance, downregulates repair mechanisms. These cells become trapped in a state where they are damaged and can’t fix it because they are actively stopping themselves from doing so. They are trapped in a local minimum instead of jumping to a better equilibrium.
Rejuvenation involves improving coordination within and across cells to give them a high-level goal: to play better together. Aged cells are almost trying too hard to maintain a faulty identity. They become locked in and lose their flexibility, adaptability, and resilience.
48:14 Tech stagnation and why physics is cooked
Daniel 00:48:17
Stepping away from molecular biology, you’ve shown a lot of interest in the roots of progress in science. Can you talk about your views on progress? Are we currently in a state of technological stagnation?
Jose 00:48:35
I was thinking about this even before the progress studies movement. My interest came from two conflicting narratives: Ray Kurzweil’s singularity, where everything is accelerating, and Peter Thiel’s stagnation.
I wanted to find the individual truth between those sweeping statements. I’ve spent time looking at various areas of technology to see if Moore’s Law is slowing down or if the cost of electricity is decreasing.
Biology is making progress, but in areas like fundamental physics, not much has changed in a long time. We might be nearing the end of discoveries in physics. While some things remain to be resolved, they likely won’t lead to technologies like faster-than-light travel or time travel.
Daniel 00:49:41
Wait, really? Even given infinite time, you don’t think there are deep truths about physics left to discover that will enable incomprehensibly advanced technology?
Jose 00:49:56
I don’t think so. Sean Carroll has a paper on this topic attempting to prove this formally. Physics is more likely to be completed than other fields.
The unification of quantum mechanics and general relativity has not been done, so there is something left to do there. There are phenomena involving very small things moving very fast or massive objects that require that explanation.
However, for the phenomena we observe in daily life that could be used for useful purposes, what we see now is likely all there is. There are still things to unfold in fields like material science—we have better battery chemistry now, for example.
There was no fundamental physics discovery involved in that, but we still achieved improvements. At some point, we discovered electromagnetism and the nuclear force, giving us nuclear energy and magnets. I don’t believe there is a new force waiting to be found that would allow us to do something radical.
Daniel 00:50:58
That’s very disappointing.
Jose 00:51:00
It is sad.
Daniel 00:51:00
I don’t believe you.
Jose 00:51:01
It’s like the Metallica song, “Sad But True.”
Daniel 00:51:06
It’s funny because my initial interest in longevity stemmed from a childhood fascination with physics and science fiction. I wanted to visit other planets, meet aliens, and discover the secrets of the universe.
Since we don’t have faster-than-light travel and space exploration takes a long time, I realized we would all have to live much longer. Many people in the longevity field share this motivation.
They want to see everything in the universe and understand deeper truths. It’s ironic to think that the fabric of reality might not actually have that much more to discover.
Jose 00:51:48
There are certainly strange things about the ground truth of reality yet to be figured out, such as how consciousness works. Perhaps meditation will eventually provide that answer.
But the physics portion seems very stable. I’m open to changing my mind, but I occasionally see claims about things like reactionless engines, such as the EM Drive.
These claims arrive like comets, then they pass away. The likelihood of them being true is small because our current laws of physics fit our observations incredibly well.
Where could new phenomena be hiding? Before we understood magnets, there were observations of strange magnetic stones and people wondered why they moved. That led to new technologies and the discovery of a fundamental force. Today, is there any mystery of that scale left?
Daniel 00:52:57
What about dark matter? It’s a huge portion of the universe.
Jose 00:53:01
I’m open to exploring what dark matter actually is. Some claim it doesn’t exist as a substance but is actually a force. Either way, can we use it for anything? There might even be some dark matter in the room with us right now.
Daniel 00:53:19
Is the dark matter in the room with you right now?
Jose 00:53:22
Dark matter interacts very weakly with normal matter. What are you going to do with it? I guess you could use it for large-scale interstellar galaxy engineering, but short of that, its utility seems limited.
Daniel 00:53:39
We often think that we’ve gotten most of the good stuff out of physics. There is obviously a lot left to do in engineering, but what about other fields? When it comes to stagnation, how do you see the rest of science?
Jose 00:54:00
Biology is currently the endless frontier. There is so much happening there that we don’t yet understand because it is so intricate. It isn’t simple, and you cannot easily break it into broad categories.
At the same time, it isn’t just “stamp collecting.” If you take stars, for example, you can essentially catalog them and say, “Yes, there they are.” That isn’t particularly exciting; it’s just one more star.
In biology, the interactions between all the small components are incredibly intriguing. There is still plenty of ground to cover because we know how much we don’t yet know. We see this with longevity research. Some animals can regrow limbs.
Daniel 00:54:47
Limbs.
Jose 00:54:47
We cannot grow limbs yet, but there is no physical reason in principle why we couldn’t eventually figure it out—how to grow an arm, or even a third arm. That remains to be done.
Aside from that, we have AI, which is also advancing rapidly these days.
55:03 Why doesn’t the world look more futuristic
Daniel 00:55:05
Another claim Peter Thiel makes is that we’ve basically only seen progress in the world of bits versus the physical world. In the physical world, everything looks the same.
We didn’t get flying cars or major breakthroughs in manufacturing; we just didn’t get much. Perhaps now we are finally seeing cool things happen because of AI.
Jose 00:55:26
To some extent, yes, but I’m not entirely sure about that. There is an idea that the future should look like the future. If you look at cities like Chongqing, Shenzhen, or Singapore, they actually look like the future. That is because they chose to pursue it. The US could look like that if people really wanted it to.
If you look at this room, for example, the camera looks like it could be from fifty years ago, but it has much better optics now. We often arrive at convergent shapes that are simply optimal for what they do.
I touched on this in my metascience article two years ago. If you take the laws of physics and the constraints of being human, some designs follow semi-deterministically. Take a restaurant, for example. Human beings like food and gathering together. We are a certain height, so we need tables and surfaces that are easy to clean.
Restaurants haven’t fundamentally changed in a long time, and as long as human beings remain as they are, they likely never will. You can add bells and whistles like conveyor belt sushi, but the basic idea remains the same.
Stagnation doesn’t mean no one is trying; it might mean we have reached an equilibrium or an optimum for many things. Take flying cars. People have been trying to build them for a long time. It is only recently that the power-to-weight ratio of motors has reached the point where flying cars can be affordable or practical.
The same applies to self-driving cars. I arrived here in a Waymo, which we didn’t have ten years ago. It is still a car, but it is new. Why does a car look like a car? Wheels are a great invention, and you need space for people. People like comfortable seats. When you add those constraints together, you end up with the designs we have.
57:26 Transhumanism & asking ourselves what we want out of life
Daniel 00:57:29
We have a lot of constraints because of our biology. What do you think of transhumanist ideas? Do you think fundamental aspects of human biology will change in our lifetime?
Jose 00:57:42
Aging is the part I’ve thought about the most. If we could live radically longer lives, how would that change people? Right now, everyone follows a specific life arc: you grow up, find a career, settle into an identity, and eventually die.
If you could live much longer, you could have multiple life arcs. You could be a Buddhist monk for ten years, then a movie director, then go back to software engineering.
Some people say that death gives life meaning, but I don’t believe that. If we lived longer, people would just have more fun and take things more lightly. Instead of hyper-optimizing everything because you only have so many years, you could just chill and live life.
You could finish one life arc and then start another, changing over time. For women, for example, it would remove the pressure of the biological clock.
Life would be less driven by external constraints imposed by biology. It would feel more in the moment because you wouldn’t have the existential fear that time is constantly ticking away.
Daniel 00:59:09
I’ve always disagreed with the argument that death gives life meaning as well. Life itself gives life meaning. We find things we enjoy and we create meaning.
However, it does raise an interesting question: the constraints of our biology give rise to many of our values. We like food because we need food, and then we develop fancy cuisine on top of that. Everything stems from the identity of our organism.
As we adjust those biological boundaries, what do we turn into? What do we end up caring about in the end?
Jose 00:59:54
I’ve thought about this occasionally, though I haven’t written about it at length. We have preferences—things we like and things we don’t like. Some of those are informed by our biology.
Take modern art, for example. I used to think it was just nonsensical lines. Then I tried to understand it and eventually saw what the artist was trying to achieve. Suppose you generalize this.
Suppose you had a magic wand that could change any preference you have. You could decide to dislike peppers, choose to like all foods, or stop liking food entirely. Some people wish they liked food less so they could be thinner; they wish they had that preference.
Imagine someone wishing they were bisexual so they could be attracted to everyone, or someone wishing they were more introverted or extroverted to suit their work. You could change these things instantly.
What does that do? We often take ourselves for granted and make choices based on who we are. But if who we are is up for grabs, what happens? On what grounds do you choose?
The classic conservative argument for many things is that human nature is incompatible with certain structures. People choose what to do, or even set policy, based on those constraints. If you don’t have those constraints, the question of what you should want becomes quite intriguing.
I don’t have a good answer for that yet. There are things I would not want to change. I very rarely lie. I have strong values around being honest and having high integrity. I would not want to change that. Even if I could press a button to lie 50% more, I would not press it.
You have core things that define how you want to live your life. Everything else—even having four arms—wouldn’t change the fact that you’re the same person. Those core things I would not change.
Maybe the answer is that you keep your core values and then stay playful with your biases. You could decide to become the kind of person who is into wine and develop an extreme taste for fine vintages.
Or you could choose to become a monk, throw away all earthly interests, and meditate all day. You press a button and you don’t like food anymore because you want to do a “monk arc” for ten years.
I don’t think people will all convert to being the same. They will have arcs, much like in songs or dances. There are sequences and moves, with micro-moves within them, but they don’t all result in the same thing.
Daniel 01:02:48
When I think about the end goal of biomedical technology, I see it as giving us increasing freedom to make these choices. With an indefinite lifespan, time is no longer as limited in terms of all the different things you want to experience.
You’re no longer limited by disease or your health. There might even be another layer involving psychological interventions. People make choices, but they often have trouble implementing them.
What if there was a drug, an electrical stimulation, or some other method that allowed you to choose? If you wanted to be bisexual, you could receive electrical stimulation and now you are. We can find ways to give ourselves more leverage over our choices.
Jose 01:03:34
My ultimate take is similar to what I said about cells earlier. You exist in an environment, and ideally, you are well-adjusted. You fit in where you are, and as the world pushes on you with incentives, you can push back.
For example, if you’re in venture capital, you will probably start a podcast. If you have a podcast, you’ll probably do VC. Those things go together. You could fight it and wonder why you want to do it, or you could just go with it.
Hopefully, the result is that people become better adjusted to their environment. If someone wishes they were more of a certain trait to fit in, they will be able to achieve it.
Ozempic is the first button to press: “I wish I was thinner.” Granted. Next might be, “I wish I was smarter” or “I wish I was more reliable.” People will be able to try these changes and see how they react and how they fit in their environment. Eventually, they will just be happier. Ultimately, that is what everyone wants.
Daniel 01:04:44
I agree with your point that we wouldn’t all converge to the same thing. People simply want different things. Giving people more ability to choose the things that make them unique would be amazing.
Jose 01:04:54
I like using dance metaphors. If you’re doing a partner dance and you make a move, maybe you repeat it because it’s fun, but eventually it gets boring. Something else arises and you do something new.
Even with the same person, you could go for hours and new things come up. If everything were the same, someone would decide they don’t like being the same and they would become different. You end up with a constant flux between sameness and difference.
Some people might end up being the same. Cults form when people want to be like the Borg and become identical. That would be interesting to see.
1:05:34 Do we need war for technological progress
Daniel 01:05:37
Regarding progress, I saw you wrote some counterarguments to the claim that government pushes technological progress forward, specifically the idea that we need war for technological advancement. What do you think of that?
Jose 01:05:55
That argument is made in various contexts. For example, some say that because so much money went into the Second World War, we got many innovations out of it. Therefore, it must be true.
But we should consider the state of those innovations prior to the war. Many were already in development. Additionally, we have to ask: what would those scientists have been doing in the absence of war?
There are ways to look at this using econometrics. A researcher named Alexander Field looked at productivity growth and found that war does not necessarily lead to an increase in productivity, despite what you might imagine.
I prefer looking at both aggregate statistics and the history of individual technologies like radar or lasers. Statistics come from models of the economy, not the real economy, so you have to look at both simultaneously.
There are cases where the argument might be true. Nuclear energy, for example, happened when it did because of the government. However, you could argue that because it was driven by the government, we actually have less nuclear power now than we would have otherwise.
Daniel 01:07:13
And we had a ton of nuclear weapons.
Jose 01:07:15
Exactly. Nuclear power and nuclear weapons became forever tainted. They are viewed as scary, explosive, unsafe, and military-driven because they were pushed into production so quickly.
In a world without war or nuclear weapons, the development might have gone slower, resulting in safer reactors. People might not have panicked about nuclear energy, and we could have more of it now than we do.
This is a hypothetical, but given that much of the opposition to nuclear power stems from the mental image of a powerful bomb, it is plausible that people would be more accepting without that association.
The same applies to disasters like Chernobyl, which involved a bad reactor design. These catastrophes happened because the technology was rushed. I don’t know for certain if things would have gone more safely without the war effort, but you can imagine it.
Another example is the iPhone. Economist Mariana Mazzucato wrote a book making the claim that the government essentially invented the iPhone, or at least funded many of the technologies leading up to it.
When you examine those claims, they aren’t quite true. Because of its size, the government is involved in many things. For instance, the U.S. government was involved in Apple’s early days because they gave a guaranteed loan to a bank that then gave money to Apple. But Apple already had money; would they have failed to exist without that specific loan? Probably not.
Siri did come from a government grant to the Stanford Research Institute (SRI), but is Siri really the core part of what makes the iPhone successful?
Daniel 01:09:08
It’s the worst part of the iPhone.
Jose 01:09:10
You can trace back many things, like LCD screens, in a similar way. Perhaps a charitable reading of Mazzucato’s book is that it highlights the many ways the government was involved, but she overstates the case by claiming that role was central. It is very hard to determine the counterfactual of what would have happened in its absence.
1:09:31 Government role in science funding
Daniel 01:09:35
This reminds me of when I was young and read Frédéric Bastiat, the French economist. He has a quote about the costs that are seen and the costs that are unseen. This comes up constantly with government intervention.
You don’t know the counterfactual. When the government is this large, they are involved in everything. Take the NIH, for example. Virtually every drug developed is a result of government funding, but that’s because we live in a paradigm where the government is the primary driver of fundamental research in biology.
What do you think of the current state of biology research?
Jose 01:10:16
Because we lack easy natural experiments, it is difficult to measure productivity growth in many areas. One impactful example is stem cell research. Japan leads in this area partly because the US banned embryonic stem cell research at one point. Even with private funding, that blunt prohibition significantly hindered research.
Basic research is often exploratory and random. You never know what will come out of it, and much of it doesn’t seem to lead anywhere immediately. It can be very niche and specialized, which leads to questions about its immediate value.
There is also the ongoing discussion about the reproducibility crisis in science. If a study doesn’t replicate, it isn’t necessarily because the data was false. It could be that the replication attempt was poorly executed or that the original experimenters omitted a crucial detail in their documentation.
If we slashed the NIH budget by half, we would eventually see the impact in patents and trends. It might be similar to Twitter under Elon Musk—the company was cut in half, but it still functions. We might find we don’t need the same level of overhead.
For problems like cancer or Alzheimer’s, we could fund research more directly. Instead of just throwing money at various projects, we could use Focused Research Organizations (FROs) with specific goals. The government has done this successfully in the past with moonshots.
SpaceX is a perfect modern example. They set out to make reusable rockets and solved the technical problems to get there. In contrast, the “War on Cancer” hasn’t felt like a moonshot. It hasn’t been an iterative process of building toward a specific goal; it’s often just a series of isolated research papers.
We should also experiment with how we structure research institutions. Currently, the average age at which an academic receives their first grant is around 40. Historically, scientists like Louis Pasteur became professors in their early 20s.
We should try giving fresh PhD graduates their own labs. Youth might bring more creativity and a lack of fossilized ideas. While they might lack the wisdom of older scientists, we won’t know the impact unless we give them the opportunity. Success in such an experiment wouldn’t fix everything, but it would provide valuable data for specific fields.
1:15:06 How Jose became the Head of Theory at Retro
Daniel 01:15:11
How do you think about your role as Head of Theory at Retro? It is an incredibly cool title. You mentioned earlier that there was a backstory to how you ended up with it.
Jose 01:15:28
Before joining Retro, I was considering my next move after my previous project, Rejuvenome, ended. I was on my second O1 visa and needed to find a new role to stay in the country.
I have a diverse background in data science, electric cars, AI, and biology. I even worked at Twitter for a period after it was acquired. I eventually decided that I should be a founder and build a biotech company based on what I had learned.
I was specifically interested in Alzheimer’s disease. I wanted to understand why antibody drugs kept failing. After developing an idea for an Alzheimer’s company, I sought advice from Joe Betts-LaCroix, the CEO of Retro. We had met previously at a conference.
When I visited the Retro space, I was blown away. The labs were built inside shipping containers with custom HVAC systems designed by Joe himself. It felt like the biotech version of seeing the Tesla tent assembly line or the SpaceX facilities.
It was unconventional and inspiring. I asked if I could join the team, and that was that. I just wanted advice initially, but I ended up finding a home there.
Daniel 01:17:42
The experience of showing up and seeing where the rubber meets the road with building stuff is remarkable. It’s funny how much work goes into actually doing biology.
Jose 01:17:59
When we started Rejuvenome, Adam Adelstone and I were thinking about how to start a lab. We considered outsourcing to a CRO, partnering with an academic lab, or building our own, which seemed complicated.
The idea that you could take a warehouse and turn shipping containers into labs never crossed my mind. That’s what agency is: thinking of the unthinkable. I realized I needed more of this in my life.
Daniel 01:18:43
I’ve met people who have performed genetic engineering on themselves. They use electroporation to put plasmids into their muscles, specifically for follistatin.
There is a company doing that in Prospera now. That is true agency. People in Silicon Valley are electrocuting their muscles to insert plasmids.
Jose 01:19:10
That was the original story. When I considered what I could contribute to Retro, I thought about my ability to synthesize literature and my background as a software engineer. The CEO suggested I become the Head of Theory.
We have since semi-abolished job titles, using them only externally to explain what we do. When I joined, I helped modernize our computational biology infrastructure and established better tools and libraries.
Recently, I’ve been building internal tools like Retro OS. It’s a tool that handles data visualization, label making, sample management, and more.
Primarily, I help people think through their options for experiments. We discuss whether a program is worth starting, the market for osteoarthritis, or new opportunities from other companies. I provide the information they need to do their jobs, either by building visualization tools or doing the research myself.
Daniel 01:20:58
Once superintelligence arrives, will you be the most replaceable person or the least replaceable person?
Jose 01:21:06
I’m not sure. So far, I haven’t found ChatGPT or similar tools very helpful for my work. I mostly use them as a “second Google.”
My work involves thinking about things that aren’t currently on our radar. Once you know which question to ask, an LLM can give a decent answer.
However, how do you know which questions to ask? If you already knew you had a problem, it would be easy to solve. I spend my time thinking about the things we, as a company, are overlooking.
Daniel 01:21:49
We talked about agency before. It seems the most powerful aspect of thinking is having the agency to choose the right questions and decide what to focus on.
Jose 01:21:59
I recently wrote a blog post about how to be more agentic. Sometimes it involves a bit of serendipity.
Regarding the Retro OS tool I mentioned, there was a week where I felt like things were going fine and I wasn’t sure what to do next. While walking around the block, I realized I could just build this tool.
It started as a small solution to a minor problem and eventually grew into something much larger. You have to be in the right environment and pay attention to what is happening. Sometimes, you have to stop trying in order to get what you want.
Daniel 01:22:42
You mentioned Retro OS, which makes me wonder if we will one day have a human OS. Will we have a software layer for our biology?
I wonder if we will reach a level of engineering control where we can alter our biology as easily as writing code—curing cancer or even giving ourselves wings. What do you think that would look like?
Jose 01:23:16
It’s unclear. A “human OS” implies personalization, but deep down, human biology is quite universal. One version of that question is how much we could eventually change ourselves.
Aging might be the easiest target because there is a natural path from young to old that we can try to reverse. Something like growing a third arm is much harder because it doesn’t happen by default.
If we were to attempt that in the future, you would likely need to attach a bioreactor with the right growth factors to trick the environment into growing a new limb.
Biology is somewhat plug-and-play, so nerves might eventually connect and make the limb usable. Michael Levin published a paper on trying to regrow limbs in animals using bioreactors and growth factors. We are in the very early stages of that research.
1:24:53 How AI might put software engineers out of a job, and push them towards biotech
Daniel 01:24:54
We’ve covered a lot of ground. Is there anything else you’d like to talk about?
Jose 01:24:58
Before working in biotech, I was working in AI. I left just as everyone else was joining because it felt too crowded. There are so many smart people in AI, and it feels like the money, attention, and talent are being absorbed by that field. These days, it feels a bit lonely in biotech. I want everyone to come join us.
Since so much AI development is focused on automating software engineering, and there are so many talented people in that sector, I predict that software salaries will eventually decline as the market crunches. Those people will then have to find work elsewhere. I hope that “elsewhere” is biotech, because we truly need more people working on these problems.
Daniel 01:25:47
That is one of the primary goals of this podcast: to reach a wider audience and showcase the exciting research happening in longevity biotech. We hope to inspire software engineers to enter this space.
It would be amazing if economic factors pushed people into biotech. It would be even better if it wasn’t due to declining software salaries, but because bioengineering salaries were increasing.
Jose 01:26:19
Biotech has different dynamics than software. Software often has power law dynamics where you can hack something together and have an unbounded upside. Even successful companies like Eli Lilly, despite the GLP-1 boom, are only a fraction of the size of Nvidia.
If the software market becomes smaller, biotech becomes relatively more attractive. Salaries are determined by productivity, but it is difficult to identify a “10x” or “100x” biologist because the feedback loops are so long.
In software, you can identify a talented engineer within a few months, allowing them to command a higher salary. In biology, even the smartest person on earth has to wait years or even a decade for a drug to be approved to prove their value. While biotech doesn’t always offer software-level money, it offers deep meaning.
1:27:28 What it takes to get a flywheel in biotech
Daniel 01:27:33
What needs to happen to create the same flywheels in biology that we see in tech? One obvious factor is the need for faster feedback loops.
Jose 01:27:44
We need to work on the right problems. For a long time, biotech and pharma focused on specific drugs for small patient populations that couldn’t be scaled. GLP-1s represent a revolution because they work for almost everything; obesity affects so many other conditions, from arthritis to muscle loss.
Since you can sell those drugs to everyone, you have much higher margins. To me, the most obvious target is aging. Most health complaints are age-related, so if we had an aging drug, the market would be everyone on earth. You would no longer be limited to a specific cancer population or a small genetic subset.
At Retro, our strategy is to build something that works for a specific condition, but can then be applied everywhere else without changing the fundamental approach. The only way to make a biotech company that looks like a tech company is to target the biggest possible market. Aging is that market.
1:29:04 Rejuvenation vs Prevention
Daniel 01:29:09
You brought up an interesting point: there is currently no clinical pathway within the FDA to get an aging drug approved.
Jose 01:29:18
I’m not so sure about that. If you actually had a drug that worked, I think there would be a way to get it through.
Daniel 01:29:22
If we had something that treated aging, we would certainly find a way to get it approved, but it is incredibly hard to demonstrate. That is why companies like Retro focus on treating a specific disease and then extrapolating that toward rejuvenation.
It makes me wonder if there is a fundamental choice between rejuvenation and prevention. Is it possible that prevention would actually be easier, and we are simply barking up the wrong tree?
Jose 01:30:01
Prevention is potentially easier to achieve, but it is much harder to prove. If you give a treatment to someone with osteoarthritis and the condition disappears in a month, you can get approval very quickly. If you tell someone they are aging 10% slower, a trial might require 5,000 people and many years to show results.
Reversal is very appealing because you can see a step-function change. If someone has Alzheimer’s and then they don’t, that is much easier to measure than a change in the slope of decline. From a company-building perspective, reversal is more attractive.
At Retro, we have a treatment involving the replacement of microglia in the brain. In theory, applying this to someone with Alzheimer’s might stop the disease in its tracks. Personally, I prefer taking big swings. They may be radical, but if they work, the results are apparent quickly.
By aiming for large effect sizes, you avoid the need for massive clinical trials. If a treatment doesn’t work in a few people, you don’t need a large trial to tell you it failed; you just move on to the next big idea. While these radical approaches add scientific risk, they allow for smaller trials with much stronger signals.
Daniel 01:32:02
The concept of rejuvenation is very appealing. If we can figure it out, it would be better than anything else because we could save everyone.
However, I keep thinking about the damage of aging. Many issues are emergent from cellular aging. While atherosclerosis might be a mechanical exception, fibrosis is an example of an outward manifestation of the loss of function in individual cells.
Even if we rejuvenated individual cells, we might not solve these emergent phenomena. Conversely, if we could stop cellular aging in a 25-year-old, would we prevent almost everything from happening?
Jose 01:32:52
We don’t know how to do that yet. At Retro Biosciences, we focus on interventions that genuinely target aging, have a viable market, and are technically feasible.
We don’t know how to cure cancer yet, but we might be able to cure Alzheimer’s. If you had a potential cure for Alzheimer’s, would you not pursue it?
Our treatment involves creating young cells from iPSCs and placing them in the brain to replace old cells. Reversal is a powerful approach. We are applying this to the brain and to blood stem cells, replacing a patient’s entire blood supply with young cells.
The scientific risk is low because the research and patents already exist; we know how to produce these cells and administer them to patients. The real challenge is the grind of setting up a pipeline to manufacture and quality-control them cheaply.
We believe in the vertical integration of manufacturing. Cell and gene therapies are powerful, but people are often deterred by their million-dollar price tags.
These therapies can be much cheaper, but you must innovate on the manufacturing process yourself. You may need to build your own bioreactors or clean rooms if you want to make these treatments accessible to everyone. Much of the current high cost is driven by profit margins rather than the intrinsic cost of production.
1:34:23 Aging is the coolest hardest problem to work on
Daniel 01:34:29
I need to take a quick break.
Jose 01:34:30
I missed that poster: 160.
Daniel 01:34:34
That refers to Omri’s blog post about the raise to 160. It’s a cool room.
Jose 01:34:40
It is such an interesting problem to work on. While there is a lot of excitement about making this a reality, it is also a tremendous amount of work. It feels like we are building the track while running on it.
We don’t have all the answers yet. We have a potential path for Alzheimer’s, but we’re still figuring out how to grow the company, fund research, and solve the broader problem of aging.
I wish we had scaling laws like in AI, where more compute consistently leads to better results.
Daniel 01:35:16
We had Martin Jensen on the podcast recently, and he wrote an article with a point that really stuck with me: biology is harder than rocket science.
In many ways, that’s true. You’re trying to work with human biology, but you can’t even touch a human subject until you’ve validated your work in countless other ways.
Jose 01:35:40
Even in vitro cell culture is somewhat insane. You can take cells and keep them alive in a broth, inside a plate and an incubator, completely removed from their natural environment.
Daniel 01:35:52
The environment where we conduct these experiments is highly unnatural. We were discussing recently how mice in cages are exposed to harsh fluorescent lighting, and the mice in the top cages are constantly shaken by the movement of the mice below them.
Jose 01:36:06
We try to control those variables; for instance, our cages have individual lighting. However, they are still mice, not humans in a cage.
1:36:09 What does it take to cure aging
Daniel 01:36:16
What do you think will be the major unlocks that make biology more like engineering? You mentioned scaling laws, and the AI models we apply to biology certainly have them.
Jose 01:36:33
There are papers on scaling laws for protein design models, though they aren’t as robust as those for text models. Many people are currently trying to create virtual cells by measuring outputs after various perturbations.
The problem is that these models are often built on RNA-seq data, which is only one modality, and they typically rely on cancer cells. We have to ask how much we are actually learning from that. Furthermore, if the solution to Alzheimer’s involves replacing cells, you wouldn’t necessarily see that path from a virtual cell model. You might see it from a virtual human, but not a virtual cell.
Cells also behave differently based on age. You don’t see age reflected strongly in RNA-seq, but you see it in the chromatin state. You also see it under perturbation. If you give alcohol to both a young and an old person, their livers react differently.
Is there anything like a scaling law in biology? If you plot lifespan extension against time, you don’t see a smooth scaling law. It looks more like a period of little change followed by a massive improvement. I suspect we will eventually figure it out, likely through cellular reprogramming.
We have had most of the necessary tools for a while. It is primarily a matter of executing the process in a fast iteration loop. We need to try something, learn why it didn’t work, and try again.
NASA spent ten years planning a single rocket launch, and if it failed, it was devastatingly expensive. Elon Musk’s approach is to build a prototype, fly it, learn, and iterate until you have a functional spaceship. We can apply that same fast loop to biology by testing in mice and keeping that knowledge within the organization.
We need long-lived entities with enough capital to run this fast loop. We already have CRISPR, billions of molecules, the entire genome mapped, and viral vectors to target various cell types. The workshop where we will build the cure for aging is already set up. We just need to dedicate the time to it.
Daniel 01:39:43
A few companies like New Limit are doing perturbation experiments, testing transcription factors to see the results on aging. That data could create an AI that identifies all the various programs that can be run in cells.
Jose 01:40:05
New Limit is different from the models I described earlier. They use primary human cells rather than cancer cells and focus on transcription factors. This is a very good strategy. We are doing something similar with proteins, though even that has limitations. If you are testing transcription factors, you still have to deliver them to the cells.
New Limit is currently focused on rejuvenating the liver. As I mentioned, delivering HNF4-alpha or FOXA1 can rejuvenate the liver. But unless a person is obese or has hepatitis, few people actually die of liver disease. Most people don’t even notice their liver is aging.
While a younger liver might provide some systemic longevity benefits, we pivoted from liver rejuvenation to cell replacement. There is a much larger market for treating conditions like Alzheimer’s. We want to build a large research entity that can run these iteration loops faster, and the liver wasn’t the best path for that.
Our current approach involves full rejuvenation ex vivo before putting the cells back in. This doesn’t scale for everything—you can’t simply replace neurons—but the transcription factor approach used by New Limit can be done in vitro to find the right factors for every cell type.
Getting those factors to the right cells throughout the body is the real challenge. Everyone in the field assumes someone else will solve the delivery problem. You can deliver to the liver, spleen, lungs, or skin, but systemic delivery remains incredibly difficult.
1:42:25 Delivery mechanisms for genetic therapies
Daniel 01:42:31
Is there a new delivery mechanism or technology you’re most excited about that might solve this?
Jose 01:42:37
Most people use lipid nanoparticles or AAVs for the delivery of nucleic acid medicines, both of which tend to go to the liver.
Daniel 01:42:48
The viruses go to the liver too.
Jose 01:42:50
Everything goes to the liver. You can target viruses more easily than lipid nanoparticles, which generally go to the spleen, liver, and lungs. There is a paper called SORT that discusses changing lipid composition to improve targeting. AAVs can be targeted toward the brain, but they still hit the liver heavily.
Daniel 01:43:07
And you have immunogenicity issues with the AAVs.
Jose 01:43:11
They are viruses, so they have immunogenicity issues. They also have a limited payload capacity of about 4.7 kilobases.
Because of this, some researchers are using HSV, which allows for much larger payloads. There is already an FDA-approved therapy for the skin that uses HSV to transduce cells.
Daniel 01:43:33
That’s the herpes virus.
Jose 01:43:35
You could use various viruses for delivery. For example, herpes likes neurons, so it can be used to deliver to those specifically. You could also use cytomegalovirus (CMV), which is a herpesvirus that is quite large and very effective at hiding from the immune system.
There is also a parasite called Toxoplasma gondii that travels to the brain. A student at the Boyden lab modified Toxoplasma as a delivery mechanism to the brain. You could also engineer cells, which are essentially nanobots; you can put anything in them and send them out to perform specific tasks.
With an entire cell, you have much more to work with. T-cells, for instance, can recognize cancer or infected cells and target them specifically rather than killing everything. When they fail at this, you get autoimmune diseases.
You could imagine building circuits where a T-cell circulates, identifies a hepatocyte, and delivers a rejuvenation factor only to that cell. You could build a cell therapy for rejuvenation where these cells move around the body, providing factors to whichever cells need them.
Daniel 01:44:49
Is that happening now? Who is building that?
Jose 01:44:52
The closest thing I have seen was a paper from Guan Kui Liu regarding FoxO3-engineered mesenchymal stem cells (MSCs). They engineered FoxO3—a factor that boosts DNA damage repair—into these cells and injected them into monkeys.
The issue in that study was that the cells were allogeneic, meaning they came from a different donor. When you put them into a different being, you face immune rejection. The cells release their cargo, rejuvenate the surrounding area, and then die, so you have to keep dosing.
However, if you were to do it autologously, where the therapy is made from your own cells, there is no reaction. People tend to shy away from autologous therapies because they are expensive; you have to make a separate therapy for every person. But we believe that with sufficiently advanced bioreactors, we can make it cheap enough for everyone.
Daniel 01:45:44
I imagine we will eventually figure that out. I wonder what technological advancements we would need to reach a point where we have fully programmable T-cells.
Can we run code in the T-cell so that when it is around the liver, it excretes liver rejuvenation transcription factors, and so on for each tissue?
Jose 01:46:05
To some extent, we already have CAR T-cells which operate on a basic logic: if you have a specific antigen, then kill. We even have AND gates and OR gates now. There are labs working on pathway engineering where entire signaling pathways are designed.
The building blocks are already there. You would have to find receptors in the given cell type of interest that your cell binds to.
Normally, T-cells detect a target, create a tunnel, and throw in granzyme and perforin to kill the cell. You could imagine them throwing in something else instead. That might require significant directed engineering to achieve, but the concept is sound.
Daniel 01:46:51
I imagine having a receptor for the right antigen—or perhaps several different receptors for different antigens—where that receptor on the inside of the membrane attaches to different payloads depending on which one you want to deposit.
Jose 01:47:03
If you were able to have a system that can deliver a specific cargo to a specific cell type using a cell therapy approach, the problem is basically solved because these things are fairly modular.
In the CAR T case, there is a paper on “FibroCAR,” which is a chimeric antigen receptor targeted against the fibroblasts that produce fibrosis. You can clear fibrosis with a CAR T. You can just swap the CAR and put anything you want in there.
We have ways to kill any cell type in a very programmable way, just like CRISPR. The question is programmable delivery—having the cell secrete a specific factor if and only if a certain condition is met. We don’t have that yet.
Daniel 01:47:43
Why not? It seems like the obvious solution.
Jose 01:47:47
I haven’t looked into it deeply. We are currently busy making cells to replace old ones, but that could certainly be a pathway to universal delivery.
Daniel 01:47:56
It’s amazing. You are so limited by the cargo you can fit into a virus, but you can fit so much machinery into a cell for coding whatever result you’re trying to achieve.
Jose 01:48:06
You also need regulation. For example, if you put Yamanaka factors on a T-cell and it finds an age-related marker on a cell surface and injects them, the cell might over-respond and create a tumor. You need a way for the process to stop at the right time.
The factors you are infusing into the cell should be self-limiting so that regulation is built-in. One of the biggest constraints for rejuvenation is that if you push too hard, you may get side effects.
Even in a non-Yamanaka approach, a transcription factor that is good for the liver might be harmful in a different cell type. You need to ensure things are going to the right place.
As a long-term goal, we want to rejuvenate in situ. In the short term, it feels like a gigantic hack to make young cells outside the body and put them in. You could swap 80% of your blood cells and your microglia. No one is really working on that, but we should. It seems very doable right now.
Sometimes, believing strongly in an idea is the moat. It’s not necessarily rocket science, but no one believed in it strongly enough to execute it until now.
1:48:53 Retro’s work to replace microglia and engineer them outside the body
Daniel 01:49:36
You are sidestepping the delivery issue by doing the genetic engineering ex vivo, outside of the body. But that creates a new issue: how do you deliver those engineered cells back into the body?
Jose 01:49:51
That is actually the easy part, or at least we chose to circumvent the difficulty. In the LNP or AAV case, targeting the liver is very easy. In our case, with hematopoietic stem cells (HSCs) and microglia, they know where to go once injected.
That is the beauty of these cells. Cells reside where they do because they have receptors that signal when they are in the right place.
We have known for decades that you can perform bone marrow transplants by injecting cells into the blood; the cells simply find their way into the bone marrow and engraft there. Similarly, if you inject microglia into the brain, they will spread and take over naturally.
Daniel 01:50:37
You just need to inject the microglia into the CSF or something similar? Since they are very mobile, they should move throughout the brain.
Jose 01:50:45
They are meant to move. Hematopoietic stem cells (HSCs) also know how to navigate the blood. We already know bone marrow transplants are possible; this is not science fiction. People do bone marrow transplants all the time.
Daniel 01:50:56
How do you kill the bad microglia cells?
Jose 01:50:59
There are a couple of approaches in the literature. Many cells, including microglia, have receptors that tell them to stay alive. These receptors are constantly being activated by surrounding cells that secrete substances like CSF1 or IL34.
Microglia receive those signals and recognize they are in the right place. If you block those receptors, the cells feel lost and die. That is one approach.
Another approach relies on competition. The incoming young cells might survive longer. Since cells in the microglia turnover by dying and dividing, you could imagine that young cells slowly replace the old ones. The old ones may be damaged, and the new ones will eventually replace them.
In the case of bone marrow, people currently use chemotherapy or radiation to ablate the marrow. However, you can also block receptors like CD117 or c-Kit to kill the old cells before putting new ones in.
We don’t know how to do this for every single cell type yet, but these two approaches seem promising. If all we had to do to cure Alzheimer’s was replace all the HSCs and microglia, that would be a massive achievement.
Daniel 01:52:18
There was a paper out of Calico from Oliver Hahn’s lab recently regarding microglia. He showed that if you put young microglia into an aged brain, the young microglia aged at an accelerated rate.
Jose 01:52:31
The issue with that paper is that those cells were not truly microglia. Microglia are unique; they are the tissue-resident macrophages of the brain that clean up debris.
One way macrophages are made is from HSCs in the bone marrow. If you inject HSCs into the brain, they become “microglia-like,” but the real microglia are different. They are made very early during development and populate the brain before the blood-brain barrier forms.
Once trapped there, they self-replace. There is no natural influx of these HSC-derived macrophages. They are a different lineage of cells.
That paper made its claim based on transcriptomics, noting that the cells looked inflamed. That makes sense because they are reacting to an old brain. However, many other papers show that when you inject microglia into Alzheimer’s mice, even if the microglia appear inflamed, factors like neuroinflammation, plaques, and behavior actually improve.
Those are the things we arguably care about most. It remains to be seen what happens when you put them in a human. The theory is that the aged brain is inflamed and damaged, and that inflammation comes largely from the microglia. Replacing them should help lower the inflammation in that brain.
If microglia replacement fails to do anything, we can try replacing astrocytes or oligodendrocytes next. At some point, if you replace enough components, you have to fix the problem.
If even that doesn’t work, we could edit the microglia to make them more resilient so they can take over the brain without becoming dysfunctional. With this approach, if the wild-type cells don’t work, we can create “super microglia” and keep trying until it works.
Even though replacing cells in the brain sounds extreme, it is in many ways easier than developing molecular therapeutics. For molecules, each one is unique with its own targets and toxicity. Microglia are very safe. You can edit and change them, and there is no known cancer associated with microglia. We know that if we make a gene edit and test it, it will likely be safe. The cells won’t migrate elsewhere, and if it’s not effective, we can just try again.
Daniel 01:55:15
When do you think you will have Alzheimer’s disease cured?
Jose 01:55:20
That is unclear. We are testing another molecular drug for Alzheimer’s this year, but the microglia program is further out. I think we will take it to trials in 2027 or 2028.
We are currently focused on quality control and getting it into FDA-ready shape. We are also testing it in Alzheimer’s models, which is tricky. You cannot simply put human cells into Alzheimer’s mouse models, and the mouse models where you can put human cells don’t naturally have Alzheimer’s. You have to create strange transgenic mice.
Daniel 01:55:58
Right. Mice don’t naturally develop Alzheimer’s.
Jose 01:56:02
You have to alter them significantly to give them Alzheimer’s, and even then, it isn’t exactly the same as the human disease. It is important to remember that just because something works in a mouse doesn’t mean it works in a human. Similarly, if something doesn’t work in a mouse, it might still work in a human.
Daniel 01:56:15
A mouse is not just a small human. It is quite different.
Jose 01:56:18
Exactly. This is why it is important to test in aged mice and across many different models. It is possible to “cure” Alzheimer’s in mice by removing amyloid plaques, but that doesn’t work as well in humans. Perhaps that is because of inflammation or other factors of aging that weren’t present in the model.
Our guiding idea is that if you have a therapeutic that works for aging itself, it should work for many diseases simultaneously. The same molecule should work across many models.
This helps us avoid overfitting. If you try too hard to fix a specific mouse model, you overfit to that model. For example, these mice often have much more amyloid beta in their brains than humans do, and it appears much earlier.
If your drug fixes amyloid-driven Alzheimer’s, tauopathies, regular aged mice, and Parkinson’s, you are likely identifying a more fundamental mechanism. That gives us more confidence that it isn’t just a quirk of a specific mouse model.
Daniel 01:57:31
Are all your microglia results currently in mouse models?
Jose 01:57:35
We have indirect human data from the literature regarding the risks and feasibility of this approach. For example, can we get these cells into the brain to begin with? The answer is yes.
When patients receive bone marrow transplants and later pass away, autopsies can reveal donor cells in the brain. In cases involving a male donor and a female recipient, we can identify Y chromosomes in those cells.
This evidence shows a double-digit percentage replacement of microglia in the brain following transplants. It proves that these cells can successfully enter and remain in the brain.
Daniel Shur 01:58:08
Wow.
Jose 01:58:09
There is additional evidence from a condition called clonal hematopoiesis of indeterminate potential, or CHIP. In this condition, marrow cells proliferate faster, which can lead to blood cancer. However, these cells also migrate into the brain more easily.
As people age, blood vessels become leakier, allowing these cells to enter the brain. People with CHIP actually seem to develop less Alzheimer’s, possibly because this natural cell replacement is occurring. This suggests that the replacement process is actually helpful.
While no one has precisely replicated the injection method we are proposing, this natural evidence from population and GWAS studies—where many Alzheimer’s genes are microglia-related—is as close as we can get before clinical testing.
1:59:10 Consciousness
Daniel Shur 01:59:17
Since we are talking about brain replacement, we have to address consciousness. When you start replacing tissue in my brain, at what point am I no longer me?
Jose 01:59:27
Microglia do not seem to be heavily involved in cognition. In mice, you can deplete microglia for months, and they continue to function normally. They likely remain conscious.
Replacing neurons would be a much different experience. There are currently trials for Parkinson’s disease that involve localized neuron replacement, but as long as enough of the original brain remains, you likely would not notice a difference in your continuity of self.
If I were to inject neurons everywhere and double your neuron count, you would feel different—perhaps like the days you wake up feeling particularly sharp—but you would still feel like yourself.
The brain is quite modular. We see this in Alzheimer’s patients; initially, the brain becomes slower, but the sense of self remains even as memory and word-finding abilities decline. Continuity persists as long as there is enough brain structure.
2:00:39 Jose’s Origin Story
Daniel Shur 02:00:42
What is your origin story? How did you end up in this field?
Jose 02:00:46
I was born in Madrid, Spain, and grew up in the Canary Islands. It was a wonderful place to grow up, very sunny and similar to California. Even the palm trees here in California are originally from the Canary Islands.
I eventually went to college to study mechanical and aerospace engineering. Rockets were exciting, but I couldn’t find a job in that field. I ended up working in a car factory in Coventry, UK, making London taxi cabs.
After six months, I realized I wanted something more. AlphaGo had just been released, which inspired me to get into machine learning. I started as a data scientist at a consulting firm in London. I was wearing a suit every day and visiting client sites, but I realized that the extroversion required for consulting wasn’t for me.
I transitioned to a startup called Aiden AI, where we built machine learning models for marketing analytics. We eventually sold that company to Twitter.
Daniel Shur 02:02:45
So many people destined to work on longevity get stuck in the B2B SaaS phase of their career.
Jose 02:02:51
I was working at Twitter in London before Elon Musk took over, vesting my shares and wondering if building SaaS products was all there was to life.
I started visiting San Francisco because the people I followed and admired online were all based here. I wanted to find a way to move, but I didn’t know who would hire a guy with a blog.
I managed to move here thanks to a benefactor who hired me on an O-1 visa. They gave me two years to work on whatever I wanted, which is when my path really shifted.
Daniel Shur 02:03:27
That’s amazing. Did you get their attention because of your blog?
Jose 02:03:28
Yes.
Jose 02:03:32
Someone slid into my DMs saying, “I like your blog, do you want to come hang out?” After we met, I switched to blogging full-time for about two years. During that time, I wrote extensively about technological progress and first became interested in longevity.
I wrote a longevity FAQ piece primarily because I wanted to understand aging. I didn’t have a specific strategy or a plan to work in the field. I just wanted to write it. Once I did, people began taking me seriously because they saw I had put in the time to think about these problems.
I began attending conferences and understanding the landscape. Eventually, after moving here, I was able to raise money for a project. When you read enough papers, you develop a sense of taste for what’s missing. To me, that was the idea of combining interventions.
I wondered what would happen if we combined treatments like rapamycin and telomere therapies at the same time. I hadn’t seen much work in that area. I remember talking to Laura Deming about this, and she confirmed that no one had really tried these combinations.
I wanted to try it, which led me to Rejuvenome and later to Retro. It wasn’t a master plan; things just made sense in the moment. It felt like the plot demanded that next action, and I followed it.
Daniel 02:05:07
Amazing. Where can people go to follow your work?
Jose 02:05:11
You can find my blog at nintil.com or follow me on Twitter @artirkel. You can also search for my full name; there is only one of me, so it’s easy to find.
Daniel 02:05:21
Thank you for joining us on the podcast.
Jose 02:05:22
Thanks for having me.
Daniel 02:05:23
Thank you for listening to this episode of the Free Radicals podcast. If you enjoyed the show and would like to support us, the most helpful thing you can do is share this with a friend you think might enjoy it too.
Please also leave us a five-star review on Spotify or Apple Podcasts, and like and subscribe on YouTube. It would really mean a lot.
I’m Daniel Shur and my co-host is Eric Dai. Thanks for listening.
