Kind: captions
Language: en
We're entering into an era where life
itself is editable. One company has
already brought back the genetic code of
extinct animals. They've resurrected DNA
that hasn't walked the earth in over
10,000 years. Not as science fiction,
not as metaphor, as living organisms,
engineered, grown, and born. Their
founder isn't a biologist. He's a tech
entrepreneur. and his team is using the
same tools, machine learning, crisper,
synthetic biology to build something
we've never had before, the power to
rewrite evolution. Today's guest is Ben
Lamb, the CEO of Colossal Biosciences.
His company just unveiled the world's
first genetically engineered direwolves.
But this conversation isn't really about
animals. It's about what happens when
the interface to biology becomes a
computer. It's about what we design
next. I bring you Ben
[Music]
Lamb. Life seems so complex, so
mysterious. How is it possible that we
can manually edit it? Is there a
specific breakthrough that's made it
possible for you guys to deextinct
animals? So if you look at the uh that
DNA twisted ladder that you know
everyone thinks of when they think of
DNA to be able to edit one individual
rung and either knock something out or
you know eventually even to change those
into a different letter. So eventually
we can't do that now. We we sorry we can
do that now. But but when crisper first
was discovered, it then moved into
single nucleotide editing editing where
you could edit certain uh individual
letters, right? And then it came into
the uh DNA synthesis where we can
actually synthesize big blocks and swap
them in. What do you mean synthesize? So
we can actually like uh with you know we
do it internally at Colossal, but we
also have thirdparty partners that we
work with at some where we can actually
string a bunch of those letters
together. So instead of us wanting to go
in and make like 20 edits or a 100 edits
or even a thousand edits in a particular
order, we could it may be easier just to
sometimes synthesize that block. So this
block does this this block is this
letters in this order. And so instead of
us changing each one of those letters on
that kind of twisted ladder, we can
actually just uh design and build those
letters in the right order. So we don't
have to change them or swap them. We can
literally just insert, knock out that
same little block and insert a block
with the letters we want in the right
order. So, hold on. This is so crazy.
So, I love I love this is where we
jumped in, right? So, this is great.
When I say that, I'm literally obsessed
with how far this can take us. Uh I'm
obsessed, but I like to actually
understand this. I don't even think that
we're at the doorstep like where we are,
right? Cuz like if you look at even what
we're doing at Colossal right now, we
use a combination of tools, right? So,
we're not using just one tool. It's not
just like a knockout, meaning we're
knocking out pieces of the genome. We're
not knocking in pieces of the genome
where we're actually where we have
something we're trying to swap it out.
We're not editing just individual
nucleotides like making individual
letter changes. And we're not just doing
DNA synthesis. We're doing all of it.
So, we're combining all of it. And
that's what's called multipplex editing.
And so, one of the things that that
Colossal is trying to pioneer uh and
push forward is this ability to take
these tools that individually work.
Okay. in some cases great, but in some a
lot of times okay, and deliver it in a
in a mechanism where we can make a lot
of edits to the genome at the exact same
time. And we're still like we're we're
running victory laps on the fact that we
made 20 edits with one multipplex array,
right? Which is no which is unheard of.
No. Do you have to do that in one block
or you can literally do 20 separate
randomly placed all over the genome?
Yeah. Yo, which is amazing.
So we so we announced this woolly mouse.
So to give you context, we announced
this woolly mouse, which is objectively
cute regardless of what you think about
de-extinction. We made this boy mouse.
It it's it's amazing. And it had eight
edits to seven genes. So it just had
eight single nucleotide edits, right? So
it wasn't like we weren't synthesizing a
big block. We were literally making
eight uh specific edits in these in
these seven genes. But what was
interesting is they were the same uh
mouse equivalents of what we saw in our
mammoth to Asian elephant research. So
we do a lot of computational analysis.
We spent a lot of time looking at
ancient genomes, comparing them to their
closest living relatives, understanding
the difference, and based on the region
it's in, it kind of gives us a clue of
whether it'll, you know, make a a change
to the hair phenotype, fat metabolism or
whatnot. But these eight edits, just
these eight edits conferred all of the
uh uh hair type, hair length, coat
color, um uh hair density. Our chief
science officer calls it floofiness, but
uh basically how our uh how the hairs
grow in slightly different ways. So it
kind of gives it that uh that kind of
woolly texture. And then one was around
fat metal fat metabolism. And if you
look at like the control mouse, which
just looks like your average smooth like
gray mouse compared to this, you know,
this our our woolly mice look like, you
know, that they have a really cool fur
coat on, right? And we're still testing
whether the fat metabolism works. We
don't know if that one, you know,
confers cold in the mice. What's
interesting about that is um a lot of
people were like, "Whoa, that's
absolutely amazing." But it's only eight
edits. It's literally only eight eight
edits. And people had made eight edits
before, but they had done them
sequentially, right? So they like made
an edit in this line of mouse. That
mouse uh you know had offspring. Then
they made an edit in the next line and
the next line and the next line and they
stacked them over time. So eight kids
deep. You finally have the very
different. Yes. How are you doing that?
Is this writing on the back of a virus
that knows that it needs to go in and
clip? Like I have a vague understanding.
You've got it. You've got it right.
Right. So we So we design uh these
crisper guides and a crisper guide is a
virus. Is a virus. So we go in, it
actually delivers, it knows exactly
where to cut it. Um, but it's not 100%
right. So that's that's part of the
problem. It sometimes cuts in the wrong
place. It cuts in the wrong places.
Sometimes there's a lot of what's called
linear repeats in in the genomes,
meaning that there's places where
there's the same letters and, you know,
some of these tools aren't quite there
yet. I mean, we're still in the early
stages. Like I said, I think we're just
barely opening the doors. It's like
opening a book and there's a few times
where it says, "And the people went."
Yeah. And so if you tell it to look for
and the people went, you might be in the
wrong chapter. The wrong chapter or
whatnot, right? And so or if you changed
it, it may forget people, right? Or may
have knocked out people too. So there's
these things called offtarget effects
where you get these unintended
consequences, right? And so the way that
we fix it and the way that, you know,
uh, and this isn't where I think the
technology will exist long term, but
where it exists today is we do an
intense amount of like screening. So we
screen all of the cells before we edit
them, right? So we do sequencing on all
of them. So we're very heavy on compute.
That may not be a surprise to you given
my background in software, but so we
spend a lot of time on understanding uh
the existing genomes. Then after we make
uh edits, after those after we let those
cells kind of um go through a couple
passages, meaning they divide and and we
we then sequence the cells that that
fully divide before we even go into
things like the woolly mice. You know,
people were like, "Well, how did you
get, you know, all 36 of your mice that
were born were healthy?" It's like,
"Well, because we screened all those
embryos. We didn't take any embryos to
term." What are you looking for? Cuz
like when I think about human embryos
and screening like that, when people are
doing IVF, they're looking at like
morphological for a lot of time. They're
just reading the code in the DNA and
going, "We know what this means. This
means something bad." And so, we're
gonna I just went through IVF. So, um,
and yeah. And so we actually went
through it and and I actually have a
gene mutation that causes a truncated
protein uh on uh uh on the Titan gene.
And so we didn't want to pass that on
which has already been mapped. People
already know about this, which is crazy.
How much do we already know about this
stuff? I'll get push back on this, but I
think that we don't know very much. I
think we are learning at an incredible
rate, right? There's a ton of literature
out there, but things like large
language models and being able to ingest
a lot of research. A lot of times people
will call things uh different things in
research. So like if they're looking at
a mouse model, they'll say, "Oh, this
gene could have this effect, but that
effect may be uh described differently
in this paper than in this paper."
Right? So I think though there's been a
lot of raw data created, I think that uh
the kind of access to compute, AI, and
eventually quantum, which I hear is only
two years away, every two years. I think
that those things coming together will
be able to ingest a lot of this
historical data and uh existing data. I
think that we are at this moment where
you know in the next 5 years we will
move from scientific research and
scientific discovery to engineering and
I and I and I think that that is in
another way do you mean that um instead
of a bunch of people writing papers
about this we're going to know okay
we've scanned 46 million people this
guy's 6'2 this guy's 5'4 here's what it
looks like this guy has thick luscious
hair this guy has like really cheesy
thin hair and we just start mapping all
that stuff out and we go, "Oh, okay,
cool. I know these letters in this place
are going to result in this
characteristic." Yeah. And I think I I
think we will, but that's that's what's
called genotype to phenotype. Like that
that translation layer is pretty hard.
That's not what you mean. No, it is. It
is. So, but I think that's I think
that's very hard. There are certain
things like so one of the biggest
challenges that Colossal is working on
is size. So, there's not like a single
gene that you make like a single edit in
and it makes like and it's like, "Oh,
well, we can now 3X something. We can
now make a chicken that's 20 times
larger, right? And that's great for
farming, right? We're not there yet,
right? We as humanity, we're literally
going to have dinosaur chickens running
around here in 20 years. I'm not I'm not
saying that, but um we do get a lot of
dinosaur. That's our number the number
one requested thing is God bless
Jurassic Park. Yeah, we get a lot. We've
also heard that before, too. Um
surprisingly, shocking. But but I do
think that we're at this point where
we're going to start to understand uh
two major categories. one is expression
of genes into physical traits or
phenotypes right so that we know that
this gene or this combination of genes
actually will have this effect um and
and it's highly predictable right so I
think that that is the world so colossal
spends a lot of time in that genotype to
phenotype mapping that's like a large
portion of our time secondly though I
think that we're going to understand how
uh and there's been previous research to
do this but the application of these
technologies to human healthcare I think
are going to even more interesting than
than just like making woolly mammoth and
whatnot. So the ability to understand
that you have a predisposition to this
type of cancer or this type of of lung
of Alzheimer's or or diabetes unlike a a
single gene mutation that only requires
a single edit. You have to be able to
edit all of those genomes at the same
time with a all those genes at the exact
same time with a high degree of
efficiency. Right? Because you don't
want offtarget effects, especially if
you're dealing with humans. Can you do
this to somebody that's already alive?
You will be able to, right? And so
what's what's great about Colossal is
that we work in uh we do a combination
of editing in micro injection into
embryos, but also in editing cells that
then we use a process called sematic
cell nuclear transfer where we're taking
the nucleus of an egg or of a sematic
cell and putting it into that of a germ
cell or an egg cell, also known as an
egg cell. And so um so we aren't doing
that but a lot of the gene therapy work
and a lot of these other works are
looking at how we target different
tissue types for delivery. So you've got
the gene editing requirements, you got
the computational requirements, then you
have the the delivery requirements. I
mean for us it's not it's not as a as a
high priority but there's entire teams I
sat through a presentation at the V
institute last year where there was like
three teams that are just looking at
kind of like air traffic control of how
you push different targeted therapies to
different tissue types in the body. So
there are people working on it just
meaning uh we put it into the body but
we need to tell it exactly where to go
and what to do and what to do. Yeah. So
it target cells. Yeah. And people are
and people are looking at how that could
be a uh form of treatment even for
cancer. Right. like how do we how do we
target things better? So targeting is a
huge thing in in this field. Targeting
is a huge thing that makes a lot of
sense. It it affects colossal less,
right? So we aren't focusing on much of
because like we're not trying to take a
you know wolf and engineering traits to
it in a in a you know three-year-old
wolf. We're starting at the cellular
stage and the embryo stage. Okay. Uh, so
you're taking advantage of the fact that
viruses already do this. They go in and
they snip. When you say that you're
training a virus, how do you give it the
sequence of letters or whatever to look
for? I think I told you this before is
like I'm not a scientist, but I just get
like nerd out on it. I come from a
software perspective. So, we actually
have a teams that just do all of the
different guide design. So, they
actually design it. They actually put
the letters in. uh they'll actually uh
in some cases synthesize little pieces
of of the uh of the code that we want in
the right in the right order and then
and deliver that in the entire pack give
that to the entire package that then uh
goes into the cell to deliver it. Yo,
okay. What's the craziest thing that you
guys are doing right now with that
technology? the craziest uh I I'll give
you I'll give you three things that are
just super weird and I don't know if
we'll get there but uh we are trying to
synthesize a massive block of DNA.
There's a project that we're working on
without going into it uh has because you
can't because Lord knows with that smile
I want you to Yeah. Yeah.
So most people know we were working on
the mammothin and do people now know
that we're working on or we have worked
on direwolves. Uh but there's other
projects but like for example with the
when you say just for anybody that
doesn't know you when you say worked on
direwolves you mean they exist? Yeah
they exist. So we have three direwolves
right now. Yo yeah which I'm sure we'll
talk about which are amazing. So Ramulus
Remis and then Khesi. So crazy. They've
been extinct for 10,000 years. Uh 10,000
years. A little over 10,000 years. The
we used a 73,000y old skull in a 12,000
year old tooth to do all the
computational analysis to kind of build
our genome. Yeah. Right. Okay. So
craziest thing. So the craziest so two
crazy or three crazy things that we're
doing right now. number one uh is uh
patterning both in the thyloine and in
other species is really important right
so like getting how the stripes for
example you know we can't just like make
up like if if direwolf if we did the DNA
analysis and saw that direwolves were
striped there's no preserved direwolves
so it's like okay well the stripes don't
have to be 100% from a fidelity
perspective but we know exactly what the
thiosene looked like there's pictures
there's color corrected videos there's
hides so So to get that right, we're
actually synthesizing large blocks of
DNA uh larger than anyone's ever
synthesized. And we're working to to
safely deliver that in the cell. And one
of the technologies that we've developed
is we we built it really large and
assembled it. We're then cutting it at
specific points, putting it in, and then
trying to stitch it together all at the
same time. So that's one really insane
hard thing that we're doing. That's
really hard because of the size of the
block. You're trying side of the block
because it's pretty fragile, right? It's
going into the cell. There's only so
much. So, there's like a physical
manipulation problem. There's a there's
a Yes. But it's all done with like
chemicals. So, we're not It's We're not
doing it with like tiny robots. Um that
the virus isn't That's what I had in my
head. Yeah. Yeah. But we aren't doing it
right. The the system the virus is doing
the system. But does the virus literally
carry it like a tugboat? It it it
carries it like a
tugboat and so that's one. A separate
thing that we that I think is uh uh
pretty interesting is um we want to
create a universal egg um which I know
sounds a little weird but uh because
when when you look at doing sematic cell
nuclear transfer uh in this one okay do
tell what's a sematic cell okay so a
sematic cell is basically any cell in
your body that's not a egg or sperm cell
so whether that's a tissue we actually
have started using what's called EPC's
uh we did this on on the direwolves
which is pretty cool. Uh indothelial
progenitor cells and we love them
because they're only partially
differentiated. So if you think about um
like pur potent stem cells or all these
stem cells they're in the most naive
state so that you can use a bunch of
different chemical factors to uh
convince them uh to turn into different
tissue types, right? And so you can use
these transcription factors to convince
them to turn into different tissue
types, right? Which is pretty
interesting. So we're using that with
like hair. It's kind of creepy but
interesting. Um, so in addition to the
woolly mouse, we actually built hair
organoids for mammoths. So we are
growing uh hair follicles just to make
sure it works. Oh yeah, to make sure it
works, right? So making sure that the
edits that we are making in our elephant
cells uh will actually deliver uh the
phenotypes that we're looking for.
Right. Hold on, hold on, hold on. So we
know enough about the soup it would need
to be in to grow that we can grow the
hair by itself. So that's a whole
another thing. So if you look at so this
is what's interesting about like why I'm
super fascinated with what we're working
on is we don't just have to get
multiplex editing right we don't just
have to get computational analysis right
we don't just have to have the screening
for offt targets right or the embryology
right or the semaxure transfer which
we'll talk about the cloning right we
actually have to do things like media
conditions like the soup right like we
have to get that right and then we have
to also create different soups or medias
in order to make them uh change uh in in
certain tissues do certain things. Damn.
So, so it's not just adding. So, you
have to build the whole system.
And so, it's funny like when we launched
the boy mouse, people were like, uh, you
know, there we have a lot of people that
love us. We have some people that uh
question us, which is great. You know, I
think it increases the likelihood of us
having great conversations about
science, which is awesome. But some of
the people that have questioned us and
and and and they push back on us, one of
the things is they're like, "Well, it's
only eight edits." I was like, "Yeah,
but it's eight edits in this system."
Like, we're focusing on this, right?
Where when in reality, we need to look
at the whole system. Yeah. I find
anybody that uh says this isn't
interesting completely uninteresting.
Like, I don't even get that line of
questioning. I hear you talk about it a
lot. Like that one to me is just like
what? I don't have time for that. I
don't But but I think look, I think that
I believe in personal choice and freedom
and as long as you're not hurting
anyone. And you know, I think that
what's interesting about this is that
people uh sometimes don't respect like
like the woolly mouse is like a
technological triumph. I know that
sounds insane because it's just this
like mouse, but it it is it's amazing.
We literally m we took we took 60 uh
genomes uh in varying lengths from a 1.2
2 milliony old step mammoth all the way
down to a 3,500y old uh woolly mammoth
built a kind of this pang genome model
compared it to Asian elephants African
elephants by the way which genomes
didn't exist we had to go create those
genomes like coming from software I just
assume they just hadn't been mapped yeah
no one had mapped so no one had one made
the genomes and then no one had
annotated the genome so no one had gone
out and said this is this region of the
genome this is this gene this region of
the genome no one had done
coming from software I just you know I
knew that like 23 and me existed
somewhere yeah I just thought like oh
we'll just log on to the GCP of species
and be done with this thing right like
we'll just write you know uh comparison
algorithms and call this a day yeah but
none of that none of the data on either
side had existed which was crazy so we
had no data sets to start from a
comparative side so once we did all that
and we identified all of these core
regions we start growing things like
hair organoids but before we grow hair
organoids we had to actually figure out
how to turn elephant cells back into pur
potent stem cells so that we can then uh
turn them from uh pur potent stem cells
reprogramming them and turning them into
organoids. Jesus man. So there's a lot
that goes into this and you know we've
only been around for three and a half
years. I was going to say there's not
only a lot that goes into it but there's
a lot you've already solved. Yeah. And I
mean we're a long ways away from
everything we want to do but like that
that just little model right there I
think is like borderline magic. And uh
and so if you take all of that and then
you map all of those edits uh that drive
the hair phenotype, which is what we're
looking for in our woolly mouse to a
mouse equivalent model, right, of that
specific gene edit. Um because we didn't
want to just put mammoth genes in a
mouse. We thought that was unethical.
And we thought that, you know, there's
200 million years of genetic divergence
between the two. So it probably wouldn't
have been compatible with life. We
actually had to engineer it. Um so we
want to do this safely and responsibly.
Then you sequence it and all of that had
to work to produce a mouse and we
produced 36 and they're all healthy.
Jesus. And objectively cute. They
definitely are cute. There's no doubt
about that. But the like you said the I
don't know that it's borderline magic.
Sounds like it it spills over into
proper magic. Yeah, this is by far the
most sci-fi episode I've ever done. This
is bananas. Okay, so uh we know the
soup. We know the areas of the gene that
we can edit. We were going through the
three craziest things. That's one you
didn't. Yeah, the second one was going
back to uh that sematic cell nuclear
transfer where we actually take the
nucleus of a cell of a sematic cell and
put it into an egg cell. Well, you have
to have egg cells. So, you kind of have
like three choices for egg cells. Choice
A or four choices. One, you can uh
harvest egg cells from animals, right?
you know, and do it in a humane way
where you uh actually have to in some
cases develop devices to like we do this
on the Northern White Rhino project
where we actually have to go extract
eggs so that we can bioank them and
fertilize them so that we have them in
case that species goes extinct because
it's functionally extinct. There's only
two left. Um, and you'd rather start
with a full-blown egg than like a skin
cell. You you need a germ cell. So, you
need a germ cell. Secondly, you can make
germ cells through a process called
gametogenesis because they're they're
gameamtes. And how you do that is you
have to figure out you have to take a
cell like a skin cell turn it back into
its most naive state pur potent stem
cell and then from there you actually
have to uh reprogram it and get it to
turn into an egg cell that you can
add reprogram it what are you doing
you're adding a combination of these
chemical factors uh your your aminaka
nyaka factors yeah that okay so it's
basically a chemical sequence yes that
does this yes I I can't believe somebody
figured this out. Okay. Yeah, it's
crazy. But they're all but they're all
different for different species, right?
And so Colossal works in non-model
species, right? So like most people will
study mice, pigs, non-human primates,
right? And so um we're working in wolf
cells, we're working in fat tail dunard
cells, we're working uh in pigeon cells,
like we're working in elephant cell,
we're working in these cells that just
don't get a lot of attention. And so and
they just don't have a research. So we
have to figure out those soups. we had
to figure out the uh chemical um
transcription factors, the right
combinations, and they all have
uniquenesses, which is crazy. Like um so
wait, so closing out the three crazy
things and then because I we can go on
tangents for this forever because it's
fun and interesting, but um uh one of
the things that we're trying to do is uh
is to make a universal egg. But if the
egg donor will have a mitochondria,
right? So, we all have these like little
powerhouses of of our cell, these
mitochondria. And there is a
is an ability the further you go
genetically in in time that you could
have mitochondrial rejection, meaning
that that the mitochondria in the egg
may not be compatible with what's in the
nucleus, right? And so I'm not I'm not a
I don't get that deep into the
epigenetics side of it or in the
mitochondria rejections out of it, but
one of the things that we're trying to
do is if we could make a universal egg,
that would be incredible for like all of
conservation and obviously all of our
projects, right? Because then you would
never have to harvest oytes or egg cells
from a species, especially an endangered
species. When you say universal, do you
mean any species? Well, we're starting
in mammals, right? So it'd be mamillian
work to begin. Is that all? That's what
we're working on right now. Jesus, it's
hard. Okay, so failure of imagination on
my part. I never would have thought that
that's even possible and we don't know
if it is yet. But the theory is that
there is a point at which we all share a
set of things. The the egg cells for
these different species are are so
different, right? Um, but if you could
get the right housing structure and if
you could actually build uh the right uh
ways to either change or manipulate the
mitochondria or do mitochondrial swap in
the egg cells, then you'd have kind of
the the the building blocks like the
Legos for uh building a universal leg.
So then because there are species so
gtoenesis taking taking cells back to
their uh pur potent stem state their
most naive state to then program from uh
and re or sorry reprogram from is hard
it's very very hard hasn't been done
it's not like once again there's not the
GCP of that there's not a book that just
tells us how to do all this right uh
secondly um then going through the
gamtoenesis process of making eggs and
sperm that's insanely hard that's that's
that's a harder than making like a skin
tissue. So, if you could, you know, not
have to do that for every single species
and maybe AI and quantum just magically
figures out, who knows? But, um, I I
highly doubt it. Um, if you could make a
universal egg or at least have the
building blocks for universal egg, when
you're working a non-model species like
the northern white rhinest or elephants,
you could just use that. So, that's the
second thing that that we're working on
that's that's super weird. And then the
third thing is the artificial womb. So,
we are working on a full end to-end
system to grow everything exudo. Yo,
yeah, we'll get back to the show in a
moment, but first I want to talk about
building wealth. Crypto ownership surged
by 40% last year, but the real
opportunity isn't just owning crypto.
It's putting it in your retirement
accounts. The problem is most financial
institutions are moving at a glacial
pace to offer this option. That's where
iTrust Capital comes in. They've cracked
the code on making crypto investing
simple, secure, and most importantly,
tax advantaged. iTrust Capital has built
a system that's incredibly easy to use,
backed by bank level security, and
designed specifically for retirement
accounts. You don't need to be a tech
wizard or crypto expert to get started.
Just someone who wants to potentially
supercharge their retirement strategy.
Visit iritrust
capital.com/impact and use code impact
when you sign up to fund your account to
get a $100 bonus. Again, that's
itrustcap.comimpact and use code impact.
And for any financial professionals
listening, iTrust Capital is now
selectively partnering with
forward-thinking institutions to bring
this technology to your clients. If
that's you, email
partnerships capital.com to learn more.
This is a paid advertisement. And now,
let's get back to the show. Tell me more
about the artificial womb. That's like
one of the in my prep I was like, "Okay,
I've got to know how far along we are
with that." If we are successful by the
end of 2025
um or probably end of 2025, we'll show
some implantation in a model species
like a mouse. By the end of 2026,
um you know, you say implantation, what
do you mean? So, uh, what's interesting
about this is that, uh, the the placenta
is this like awesome amazing thing, but
it comes from the embryo. So, we don't
have to engineer that, right? We just
have, I'm going to massively
oversimplify it, but all we have to do
is to create the right environment with
the right chemical queuing in the right
place. If it is a placental mammal and
it needs to implant like 99.9% of of
placental mammals like a mouse it then
we have to give it a uterine wall or a
synthetic wall that it can implant into.
So that means that the so you'll implant
an artificial womb. So we'll implant we
will I always thought this be like a
plastic bag. So we will will hatch. Uh
so I think there's a part of it that is
a plastic bag. That's the most outer
component. But I do think that you need
um uh internal component. We're doing it
with a com combination of hydrogels like
we're we're using these like super
sticky membranes but we're also using
bioprinting. So we're actually having to
print insanely small because it's
massively vascularized, right? Because
if you think about how it works in like
a human body or in a mammal, uh the the
embryo will hatch out a placenta. the
placenta will attach or implant into the
uterine wall. Uh that uterine wall, the
uh uh the placenta, which is almost like
this awesome alien. So, it's amazing.
Literally. Yeah. It the more you look at
this, the more you're like, oh god.
Yeah. Yeah. But what's great is like I
was like, we don't have to solve that
part. Biology does that part for us.
Like it it does want to to work and want
to live. And so, it has to then uh
connect and invade into the placental
wall, but it has to be able to, you
know, obviously stay implanted. So, it's
got to be sturdy enough, but it can't be
too sturdy that it can't implant. It's
then has to be able to be vascularized
enough so that it's getting the right
level of oxygen, nutrient, extrem
exchange, and chemical queuing. Uh,
because some of that chemical queuing
comes through the placenta. Uh, some of
it also comes from the uterine wall in
just that kind of amniotic sack. So, you
have to figure out. So, but once again,
it sounds more like to me like a data
problem and a tissue engineering
problem. Um, but it I mean it doesn't
sound like magic. It sounds like just
it's just data. It's like if we track
every single day of pregnancy and and we
and we test uh and understand you know
what that chemical queuing is then kind
of like the transcription factors for um
uh for uh reprogramming stem cells we
kind of know what the variables are
right is it glucose what are these
different knobs and then we just have to
tune them for every different species
and I think over time if you get enough
data then you can probably get 80% there
so if this goes great uh by the end
2026, I predict we will have the first
mammal born fully exutera, which isn't
that long from now. Uh, not going to be
a mouse. Yeah, it's not going to be a
male. It's not going to be a wolf, but
yeah. Do you already have one in mind?
It'll be a mouse or a marsupial.
Interesting. Are they just higher
turnover rate? Yeah, there's a uh our
fat tail dun art, which is our models
fat tailed. Fat tail done. It's
literally the cute it's it g it will
give the woolly mouse a run for its
money on objectively cute factor. So um
but if you look at the uh if you if you
look at them they only have about a 13
and 1 half day gestation because
marsupials have so much gestation that
happens outside the womb. Mice have a
little north of 20 days. Got it.
But mice uh unlike mice uh the uh the
placenta uh technically marupials are
placental mammals but the placenta only
shows up for a cameo for like the last
day of gestation. So it's not really
needed uh in terms of like gestation for
uh the fat tail done art. Right. Okay.
So when you say something will be born
exudo would I be able to walk into the
lab and see it like in its placenta just
growing one if we achieve it and then
two if we achieve it and it's highly
replicatable then yes you would be we
would probably stage gate it. So you'd
have it at every single like you'd be
able to say, "Oh, here's device one and
it's at day three. This is at day seven.
This is at day 24 or in this case day 20
and it's birthing." Bro, this is so
sci-fi. No, it's crazy. Legitimately,
but I actually think that artificial
wombs, if we're successful, um, will
have way more impact to conservation
than anything else. Like we work on the
Northern White Rhino project and, um,
we're our our job is to help them do Is
it just because you can do so many?
Yeah. Yeah. So, you can do it at scale.
That's what that's the beauty of the
artificial womb. If we're really good at
this, uh, instead of it's great for
animal welfare perspective, it's also
great for conservation because, you
know, instead of having to do IVF in
rhinos to help save the northern white
rhino, right, and put them in southern
white rhinos to birth them. Uh, you
know, this is how the technology stack
immediately applies at scale to uh to
conservation. We could take a blood
draw, isolate EPCs, endtheloper cells,
which is what we did on the wolves. uh
edit those engineer in genetic diversity
right from from a population map that
we've done a population map that we've
done based on like taking tissue samples
of northern white rhinos in uh frozen
zoos in in uh in museums and other
collections to overcome the cloning
problem to well no to overcome the
genetic bottleneck problem. So, so then
they're not, you know, they're not all
clones. They're they're all they have
enough genetic diversity. So, they all
become their own founder lines, right?
So, then you take that, then you do
sematic cell nuclear transfer. You make
uh an embryo, put it in a universal egg
or use gimtoenesis to make it, and then
you take those embryos after you've done
sematic cell transfer and you put it in
a excuter device or an artificial womb,
and you grow it to term. So the system
once again the system model that we're
applying to make mammoths or direwolves
or thiocines or whatever you know has a
direct application where we could become
a uh a technology partner for
governments and NOS's around the world
to go take our technology into
conservation at a scale we've never
seen. This is a world that I don't think
people fully understand like how
transformational this is going to be.
Um, we have a lot more to cover, but I
do want to touch on the ethics of this.
So, I unfortunately and am the unhinged
guy that wants to see this stuff like
pushed to the absolute max. Y, um, but
for people who their toes curl at the
thought of us playing God, um, how do
you think about this? Well, I think that
we have a moral I I don't think that I
think that we have a moral and ethical
responsibility to do it. So because you
know we actually have developed all of
these tools and technologies as a
species. We've become the apex predator
on this planet. When we started the
company it was forecasted that we're
going to lose up to 15% of biodiversity
between now and 2050. That was three and
a half years ago. It's now 50%. Those
aren't our studies. Those are
independent studies with why why is it
going so much faster? Well because we're
accelerating climate change. We're
acceler accelerating over fishing, over
hunting, eradication of habitat, right?
And so, uh, polluting, cutting down the
air, the the rainforest, all of that is
a form of playing God, right? And like
we go into the rainforest and like slash
it all down and burn it. Well, that is a
form of playing God. We are changing
that environment, right? I would even
argue that taking cholesterol medication
on a micro level is a form of playing
God, right? Because you weren't in you
weren't designed or engineered or got
the random roll of the dice to have a
specific uh, you know, cholesterol
problem, right? And so, I think all of
these choices are a form of playing God.
And so I think that we have a moral
obligation to do something that you know
with these technologies that can not
only help and benefit human healthare
but could help species. Okay. Um talk to
me about human healthcare. Is this going
to be a uh dear parents everybody is
going to have kids via IVF. Don't worry
it's going to become very cheap. It'll
be day like everybody's going to be
doing this. Don't worry it's not just
for the rich. But um you're going to
want to do that so that you can select.
Here's the beautiful thing for Colossal
is we only work on animals and
conservation. Are you going to dodge
like that? No, I'm not going to do I
will answer the question, but I'm saying
Colossal takes the stance that we as a
company aren't going to work on humans
because we already get so many crazy
conspiracy theories about the company
given that Incutel and others are
investors. They're like, "What are they
doing? What is the secret programs they
don't tell us about?" But we're pretty
transparent about what we're working on.
I was going to say the things you're
talking about already are like they're
already weird enough. Like I mean
they're I think people give us too much
credit and like that we've got like
crazy stuff even more crazy things
hiding. Um uh but but what's interesting
is is for us is like we at Colossal draw
that line so we won't we aren't going to
make those decisions. Why not like Ben
Lamb why do you not want to do it? Well
I mean so it's too big of a headache. No
I mean I I think that the separation for
Colossal is is good between animals and
and humans. Like we won't even work on
anything in the non-human primate world.
So anything in the monkey world, we
won't do or ape world. We won't do any
of that because like there was a species
which I don't think there's any DNA
called giant epipythecus which was not
King Kong but much larger than uh you
know our orangutans and apes. We
wouldn't touch that project. We've made
a decision that we're not going to work
on anything cuz people get asked us
about Neanderl, right? And so they're
like oh could you do a neand like it's
just a slippery slope. So, we've made
the decision. It's a slippery ethical
slope because it's too human. It's it's
just a slipper. Yes. It's just too too
human. It's too close. There's enough
problems for Colossal to solve without
doing it. With that said, all the
technologies that we develop that have
an application to human healthcare, we
will license out. We'll spin them out.
So, other people may use our
technologies to eventually grow, you
know, uh humans and artificial wombs and
whatnot, even though Colossal will not
do that, right? Um but I do think that
you know other people thank God it's not
us uh but other people will have to
think about what what is the line
between um making uh uh selections for
the betterment of of a society from a
genetics perspective um before you go
all out Gatka or you look like or it
looks like eugenics right and so those
are really hard things like I'm glad
that I don't have to like make those
decisions but I will say I had a kid and
went through IVF and there's now have
you seen this whole genome sequencing
stuff that's come out. So, um, most
people when they when they go through
IVF, they're great doctors out there
will look at a morphological grade like
how is the embryo doing on day one, day
two, day, day three, day four, whatever.
And so, how is it developing? How is it
ready? And they a lot of times they'll
grade them based on the
morpholog or or be behaving like this at
this stage at this stage development.
So, that's one. Well, then they started
saying, "Okay, well, let's look for
specific things like extra chromosomes
or extra things that could that could be
could be could be really bad, right?"
And so, um, they started looking at
those types of things, right? And at
least giving that data to the people,
right? Like, you know, this is this is
an issue. A lot of times people have um
miscarriages because some of the
something was broken in the uh genetics
of it and of the developing embryo and
it just wasn't compatible with life even
though it was developing on day one, day
two or day three. Well, so now there's a
new company, full disclosure, I use them
as a service. I was so impressed with
what came out of I did invest in them.
So I'm not I just always want to be
transparent about my investments. But
Orchid Health, which is pretty awesome,
and what they do is they take a tiny
little uh uh couple of cells from the
outer ring. So not in the developing uh
embryo but in the in the outer ring they
do full genome sequencing. So not
they're not doing uh they're not doing
like 23 and me. They're doing like
really deep uh uh sequencing because 23
and me is too surface too surface,
right? They they didn't look for
everything, right? So they look for the
the core stuff, but they didn't go I
think there's a lot of reasons they went
out of business, but they um uh they
went uh uh they they do the full genome
sequencing. So they can do they can look
for their specific markers like I had
this Titan gene mutation and so we did
not want to pass that on because there's
risk. so that we could screen for that.
Right. Well, so you literally check the
embryos. This one does, this one
doesn't. The one that does don't
implant. Yeah. I mean, we still Yeah. So
you you build So I built a spreadsheet
to kind of and I built a relevance
ranking system of it. Embryos did you
have? Uh so we went through two rounds
of of IVF and so we ended up getting uh
I think at the end we ended up having we
down selected down based on this uh kind
of distributed kind of bell curve that
we that we made in a spreadsheet. Uh and
we ended up having nine. Whoa. Okay. So
nine that you personally were like I'm
going to give these a grade before we
even did the morphological rating with
the thing. So, so the other thing that's
interesting, so you're talking about
other things other than the Titan gene.
So there's also and so this is this gets
into um and so this goes back into your
question about like you know and once
again this isn't colossal making these
decisions. This is me personally with
you know in my with my family making
these decisions and we started looking
at um uh they actually do this like
polygenic risk score which some people
don't what's polygenic. So they
basically look at at like we do know
that certain genes and certain mutations
um in addition to environmental factors
could lead to higher uh diagnosis of
early onset Alzheimer's or diabetes or
certain types of cancers. Right? So one
of the things that orchid does which I
found pretty interesting is it kind of
gives you like the bell curve where most
people fall and then it can tell you
like this embryo fell here and it has
this mutation falls here right and so
you know I think that you know the the
easiest grade is like if you're looking
at this what the technologies currently
allow for the easiest grade is you know
is it compatible with life does it look
more morphologically right okay are
there major things that we want to
screen so screen for like extra
chromosomes or chromosomeal anomalies
Um, and then and then from there you can
look for specific things like in the
case of me looking for this Titan gene
mutation that we didn't want to pass on.
And then it goes into this like area
that's I wouldn't say it's gray but it's
not like if you have this exact gene
variant you will get Alzheimer's. It's
off white. It's off-white. Okay. But
it's still interesting, right? It it's
it's not based on like, you know, why
don't you want to push it all the way?
Why do I want to push it all the way? So
I mean I did, right? I mean that that's
as far as the technology goes today. So
I did that like we chose we chose you
know we looked at all of the data and we
made a selection based on the data but
we didn't make the decision based on was
it compatible with life and how did it
morphologically look we went through all
the way to the
distrib and then we built a spreadsheet
our oh god I want to see that
spreadsheet so bad. Um I promise I will
never show anyone that spreadsheet. I
figured that was going to be the answer.
Wherever we are today, I have a feeling
we're going to be able to read more and
more things.
One human response that while I get it
because humans are like that, strikes me
as very strange,
you've you've said this a few times in a
very sort of polite way about looking
for extra chromosomes, down syndrome.
Now, when you say, "Yeah, I wouldn't
implant an embryo." This Tom Billy
speaking, I would not implant an embryo
that had Down syndrome. Uh but then
you're going to get I will get backlash
on that because
people in my family we have people with
Down syndromes like you know in in in my
extended family right and so I'm not
saying that you should or shouldn't but
that but it should be your choice that
that's I'm a big believer in your
choice. There's too many um
morphological characteristics like uh
this kid is going to be tall or
beautiful or whatever that people think
or smarter let's get really well and so
people have already said that right like
we at and there's a general moratorum in
in the world including the United States
on germline editing meaning making edits
that will get passed on to the gene pool
but then you got certain places like in
China at the Beijing genomics institute
where they are like we are sequencing
everyone and we are looking for genes
and gene clusters that can drive
intelligence and so they're they're very
open about that right and so it is a
really hard thing right it's like if we
could remove I'll give you an example
like cickle cell anemia which affects a
percentage of the population uh it's one
gene edit it's just one so if we could
eradicate that for all humanity should
we I mean I would make the argument yes
but I also don't have cickle cell anemia
right and so these things go beyond ind
you know science in terms of
capabilities but go into not just ethics
but go into philosophy you know it's
like I would argue that we have some
form of directed evolution as a as a
species right you know we are no longer
just hunters and gatherers already we
we're already doing it right it's like
when people say that like what colossal
is doing uh is could go too far or that
you know we one of the things we get
back you know people love to argue
whether it's a dire you know whether we
can classify it as a direwolf or not but
what people which I'm sure we'll get
into But what people don't uh realize is
that like these these and what people
don't argue is that the technologies
that we have that we're now deploying
here exist and they're only going to get
better. And so we have to be thinking
about what the applications are for
humanity. But I would argue I always
tell people that you know we've been
doing uh gene editing for a long time.
We've just been really shitty at it,
right? Like we've been m we've been like
crossbreeding stuff forever. like
whether it was weed or corn or dogs, you
know, like or who you pick as your
spouse. We have been doing this the
oldfashioned way for a long time. Now,
let's just use data and let's just do it
faster and more efficient where we have
a high degree of predictability of the
outcome. Dude, this is inevitable. So,
I'll put my cards on the table. I have
no idea how you're going to react, but
um we are in a cold war with China.
China is going to do everything that
they can to beat us in AI. I agree. I
agree. Uh, so we can't play around.
China's going to do everything they can.
I saw a video today of a robot dog with
a gun on its back and training it to
autonomously do its thing. Yeah. If we
know, so right now, as far as I know,
the only place that has done edits on
humans is China. Is China. Yeah. Uh, you
will often say the person that did that,
the doctor that did that was allegedly
punished. I always find the use of the
word alleged. It's interesting. Yeah.
It's like you do something that big, it
gets out of jail pretty quick. Yeah.
Because you have at least running in the
back of your mind the possibility that
that was sponsored or at least like, uh,
well, if you're doing it, it's it's a
China is an information state. So, I I
find it hard to believe that people at
various degrees of government didn't
have the information.
So, it's a very delicate way to put it.
Uh, also I feel like forget editing for
a second, selection of the embryos that
you have, I would fully expect people to
go as hard in the paint as they possibly
can to get the ones that meet whatever
characteristics that are desirable. Can
I add something to that? This is crazy.
This for my own personal journey. The
morphological embryo that looked the
best based on, you know, this side, like
not going all the way out, but just this
side. The one that actually looked the
best was had a genetic defect and was
not compatible with life. Interesting.
So, more than likely, our incredible
doctor would have put in an embryo that
would have led to a miscarriage. Yo,
which is which is emotionally taxing on
the family, specifically on the female.
And it's also it's also economically
taxing. And he wasn't wrong based on the
morphological grade. Like there's a
grading system you can look at it and
but it it was not genetically compatible
with life. M but that's that's what
would have been implanted right if we
wouldn't have done these other things.
So so I do think that data can help us
be informed to make uh you know we just
make we make more informed decisions
with that data right doesn't mean there
are always going to be the right
decisions but at least we have the
personal choice to make those decisions.
Do you think that there's any of the
gray in selecting from
unedited but IVFcreated
uh embryos? Um, no. I think I I think
that that's, you know, it's it's like
the like I don't know if you have you
seen the gallery test, the grail test,
the blood test. I just got one. You did?
Literally like a week ago. You get your
results back? I haven't got my results
yet. That's a little nerve-wracking,
isn't it? For me, it is. I do I do it
every year. I went in because literally
as we're recording this, I had skin
cancer removed from my face. So, I was
like, uh, I want to know everything.
But, you're one of those people. I'm one
of those people, you know. I do I do the
function test and I do four quarterly
blood tests. Uh in addition to the
function test, which is already insane
and incredible and and comprehensive as
it as it is. Then I do four blood then I
do four quarterly blood tests. Um I do
the the Grail Gallery test, whatever you
don't know, gra.com it whatever it is
they they actually call it. Um I I do
all I do all of that because I want the
data. But a lot of people like they
don't want that data. So going back so
but that but that apply it's sometimes
it's hard for I think people like you
and I to understand that but if you and
I uh but because we're wired a certain
way but I think that goes the same way
with embryo selection. I want the best
amount of data to make the best decision
but not everyone does. Everyone some
people are like I I just want to trust
nature and trust I just want to trust
the universe and trust the chance
system. Of course I don't understand it
but they're entitled to it. I definitely
think people should be able to make
their own decisions. Yeah. Um I just
don't understand the push back on the
selection. There's been push back on
Orchid and I'm like here's this
incredible woman that developed a
technology uh and just wants to give
people more data so that they can make a
a a decision. Right. Say more about
this. What is it? This is the Orchid
health. This is the orchid orchid test
that gives you that that risk score.
This is what the system that we use. But
it's like they got pushed back when they
launched and I was like it's just data.
Yeah. You don't have to use the data.
You you don't have to take the test. So
can you steal men their argument? I
think that their argument is uh with
more data you'll make a better decision
and you will lead to healthier babies
that have longer lives that will also
have a longer term impact. I mean the
people that hate it. Oh that hate it. I
think they think it's a form it goes
back to it's a form of playing God,
right? It's like it's or it's eugenics,
right? But I mean, I think that
selecting for I think selecting for a
certain like skin color or hair or eye
color, that's eugenics, right? Seeing if
you're seeing if your kid has like my
gene mutation that has a my that has
like the issue that I have with my um
truncation on my Titan gene, that's not
eugenics. That's just like I don't want
to pass that on. I would argue that
things like orchid are a service to
humanity. They're not just an individual
service because your those genes will
persist in the gene pool long after
you're gone. It's a moral obligation to
do this, not just to you and your family
and to your child, but to humanity as a
whole. and and because of the ripple
effects of the gene pool, this will get
so bizarre if I'm a kid and my parents
chose an option that they thought was
cool or kind or there needs to be a
distribution. Like, wait a second, I
could have been the embryo that was
three times smarter. Like, I'm now
suing. Yeah, that's going to get
certainly very bizarre. It's going to be
get bizarre. And Orchid doesn't test for
intelligence, by the way. Yet. Yet. I
mean, I'm not a part of I and maybe they
never will, but that strikes
me to your to your comment. BGI and this
isn't like some like deep state like
dark web secret like they've been like
the CEO of is like yeah we do that like
they're they're very open about it.
They're like yeah they're like why
shouldn't we? Yeah that that is quite
literally my question. Um okay so uh by
what mechanism do we decide what is
eugenics and horrible and what is uh
yeah it's my kid I want to fine-tune a
little bit here. Yeah I mean the good
news is there's no germline there's this
moratorium on germline editing and uh
and so we don't have to make that
decision today. But let's pick an easy
one. So, I was just, for whatever
reason, this came across my feed and my
wife happens to have a very good friend
of hers, uh, mixed race guy, white wife.
They have twin daughters and one looks
light-skinned African-American, but not
American cuz she's British, but she
looks light-skinned, black, and then the
other one is ginger freckles. Like, how
are these two twins? Obviously, they are
not monozygotic twins, but twins
nonetheless. Yeah. And if you have that
and you're looking at that and you get
to pick like instead of you know two or
you go through and you want every color
of the rainbow or whatever like oo where
do we go? It is in the realm of sci-fi,
right? And it's like leaning on this to
as it relates to government. What is
government's decision in that, right?
Like what is their right? You know, we
have a general moratorum on not being
able to is that considered germline
though. You're just selecting because
that's already like this is happening
right now. I don't know if they were IVF
or not. But I don't I don't think anyone
out there has or is surfacing that type
of data like eye color. Like no one's
surfacing that type of data.
So people are just like, "We're not
going to tell you." Yeah. I don't think
that lasts forever. I don't think it
lasts forever. And and you know, one of
the things that makes America different
than China is our ethics is what we'll
choose to do, right? Like Yeah. I I
mean, one of the things that like
Colossal, certified by American Humane
Society, the oldest humane organization
in the world. We run everything by them
from where our animals live to how we
treat them in the lab to every single
thing, right? It's hard. It's very, very
hard. And if we didn't care about that
and we just wanted to do something in a
country like in a southeastern Asian
country and you know be like, "Oh, we'll
just go kill a 100 elephants to to do
this and we'll have a mammoth in two
years." Like we we we choose to do
things uh in an as an American company
in America by American standards,
America's ethics, and America's rules.
And you know, not all countries, you
know, in this weird complicated
geopolitical world we live in, you know,
play by the same rules and play by the
same ethical standards. They also look
at things. They also look at like if
you're if you're looking at a lot of the
Asian countries, they're looking
generational, right? So they're making
they're thinking about what is the
generational thing and we're more of a
society that looks at today and
tomorrow. So I I think there's cultural
and and um kind of generational uh views
that go into this that have a um very
very big very like it's it's like like
the old adage about like how people
steer like a cruise a cruise boat,
right? If you want to go over there, you
make a small turn. now and it'll
eventually go there. And I think that
some countries think like that. And I
think that affords them a a at least a
perspective where they don't look at the
ethics the same way we do. Yeah. We'll
get back to the show in a moment, but
first, the faster you can get your idea
to market, the better your chance of
success. And right now, there's a window
of opportunity that won't stay open
forever. Shopify's AI tools are
literally doing the work that used to
require entire teams. Need product
descriptions? The AI writes them for
you. Web design, choose from hundreds of
templates and customize them instantly.
Product photos need enhancement.
Shopify's AI handles all of that as
well. Shopify is not just an e-commerce
platform anymore. It's like having a
marketing department, design team, an
operations expert, all working for you
24/7. And that's why Shopify now powers
10% of all e-commerce in the US. From
giants like Mattel and Gym Shark to solo
founders just getting started. So turn
your big business idea into reality with
Shopify on your side. Sign up for your
$1 per month trial and start selling
today at
shopify.com/impact. Again, go to
shopify.com/impact. And now let's get
back to the show. As you um look at what
is going to be possible, everything has
to be done in order. You can't do
everything all at once. What you can
actually you can parallel path a lot of
it. Say more. Yeah. So, so in this whole
system, right? So, we're doing while
we're doing the computational analysis,
we know we're going to edit Asian
elephant cells. So, we have to get Asian
elephant cells, right? We have to get
the soup or the medium to grow them and
grow them happily. We actually So, one
of the things that's interesting about
uh we're having this I had this
conversation last night over drinks with
Bob Nelson uh about p-53. So, do you
know anything about p-53? Yeah. Having
just had cancer removed from my face.
Yes. Yes. So, um not everyone does,
right? And so, what's interesting about
p-53 is uh us and mice have one copy of
it. And uh the uh in elephants and blue
whales which are harder to study than
elephants for the whole breathing thing
is has a um uh they have about 20 times
the expression and what we know about we
don't understand everything about p-53
but what we know about it in
overexpression is when a mutation occurs
cancer is a mutation what it does is the
cell scinesses it like auto it it like
autoterminates itself just kills itself
right so then it doesn't turn into a
tumor and and then you know go crazy.
And so um we know this about blue whale
whales. We know this about uh elephants
specifically. And if you look at the age
that elephants live and the body weight
that they have, it's called I think it's
called Pedto's paradox. And you look at
this distribution curve, they actually
don't get cancer. Uh they get cancer a
fraction of what they should like a a
minuscule fraction. and they basically
don't get cancer compared to what we
would if we had at the same we have
roughly the same lifespan if not a
little bit more now but had the same
body weight and had the same mass and so
we have the same number of of cellular
divisions and so what's interesting
about that is like you know to your
parallel path point we had to figure out
if you start editing a
cell looks like a mutation because it is
a mutation it's a force mutation looks
like cancer so we had to figure out how
to regulate p-53 how you turn it down so
you can make the edits what animal did
you find this on in Asian elephants. Yo.
And then you had to figure out like how
you turn it back up. This is in the
woolly mammoth experimentation. Yeah. In
the in the woolly mammoth project. Got
it. So this goes back to your par
parallel path is like while we're doing
computational analysis and your data
like more data is more data right. So as
we're getting more and we're looking at
more of these as we're running like our
woolly mice experiment as we're growing
organoids over here as we're creating uh
induced pur potent stem cells so that we
can create organoids in the first place
and then eventually gametes. Then we
we're also working over here on the
actual editing of this of the uh targets
that we made in the woy mouse of the
mammoth equivalents. We're also working
on uh uh the we had to figure out before
we could even do that how to regulate
p-53 and turn it down but then turn it
back up because you don't want you know
super cancer elephants. And then we also
have a whole team over here that's
working on elephant IVF. How on earth do
you turn p-53 down for a bit and then
back up? It's a great question. I don't
know. That sounds like one someone on
our science team that's much harder than
me can tell you that. That is crazy. Uh,
so are you turning it back up in the
same creature? No, in the same cell
line, right? Because what you don't want
it to do is you want to be you want to
make sure that you can turn on um it's
like it's kind of like the reverse of
like immortalization constructs. If you
put an immortalization construct so that
the cell just keeps dividing, keeps
dividing. Well, eventually it it would
turn into cancer. You got to eventually
sometimes uh when you're making lots of
edits, take out those immortalization
constructs. And so it's a similar kind
of thought process on the P-53.
That that is so
I don't know what to do with that. I
don't know how that's possible. That's
but it goes back to your core question
is like you know we're not if we did all
this linearly we would have animals it
would take us a decade per animal. But
you know the reason why we you know we
have four labs we have 172 scientists uh
including you know people a part of the
you know national academy of sciences
our chief science officer we have 17
academic partners uh around the world we
fund 40 posttos and we have 95
scientific advisors with all of that you
know we we we've tried to pull all that
together uh and then we've raised a lot
of capital raised roughly half a billion
dollars uh to to do this and you have to
do this because you have to build the
system and if you want to do it you know
and have like all these species back,
you have to do it in a parallel pathway.
Okay, why do we want all these species
back? So, uh you know, a couple core
reasons. Number one, I think that and I
think you probably get this from your
background. anytime you look at things
as a system versus a point solution is a
higher it's known through like every
scientific and every engineering feat
like the Apollo uh program that you
innovate and build technologies that
have farreaching application than you
have any idea versus a point solution
where it's like oh I'm going to have a
hammer and a nail and I'm going to just
I I just want this piece of wood to go
here right whereas a system model is I
want to build a sustainable house that
can survive all of the elements right so
you have to build all of those systems
right where you have clean water, get
rid of extra, get rid of all, you know,
be able to do air filtration, have
power. So, it's a very, very different
way of looking at things versus like I
just want to solve X. And sometimes I
get negative feedback on this when I
talk about it as it relates to
scientific papers. Uh, because I'm a
very big believer in looking at the
system and documenting it and showing
the science over time, but that's not
the priority, right? The priority is,
you know, can we can we do this? So
anytime you develop a system uh you know
and specifically one that has an
application to conservation you know I
think that's interesting and it and it
allows us to innovate. So number one is
it's a system model that drives a lot of
uh potential for both uh support for
conservation as well as uh technologies
that could apply for human healthcare
both monetarily and monetize but also
that could also just help humans right
so so that's one reason why I think we
do this a second reason why we do this
is I think that you know doing things
that are hard teach us more they make us
ask harder questions like we had a whole
conversation about eugenics today would
we would we have had that conversation
if we weren't talking about gene
engineering ing and pushing gene
engineering to the to the to the limits
because we made a direwolf and a woolly
mice. I don't know. And so so I think it
also helps us from a human condition
perspective think bigger and then
expands kind of what like how does it
widen our aperture in terms of like what
is possible? And then the third thing is
kids. Like we people like love to just
like skirt over this when we're talk the
people that don't love what we're doing
love to skirt over this. Every week we
get stacks of like letters and drawings
and paintings and stuff from teachers
and parents and kids saying my kid wants
to go be a scientist or my kid wants to
be a biologist or my kid and I like we
had a a woman from Florida send us a
note saying that her classroom won't be
quiet except when they're talking about
direwolves and woolly mice and mammoth
and all this stuff and they're obsessed
with it. And so, you know, I think
inspiring the next generation like the
Apollo program and the effects of the
Apollo program, I think created, you
know, I think it outside of the success
of going to space and opening up space
and and and kind of that, you know,
monumental triumph for humanity, I think
it it changed the human consciousness on
a level of what was possible.
Yeah. When I look at
um gene editing, when I look at particle
physics, I feel the same way.
that you see these things that just
exist. They seem so fundamental. They
seem unalterable. The the process by
which they come into being feels like
magic. Yeah. And then to hear, oh, no,
no, no. Like you can you can read the
DNA, you can go in, you can make edits,
you give the virus a little tugboat
thing, it drags it in, it puts it in the
right place. Like it's just like what?
Uh it makes me realize the world does
not work the way that I thought it did
and that there are levels to how much I
can engineer it. Yeah. And I and I think
that we're just gonna get better at it,
right? Like we we as a thing like this
is the trash version. Yeah, this is
that's a good point. We are we are the
trash version, but I I do think we're at
this point where, you know, over the
next 5 to 10 years, we are going to be
moving more and more into using AI and
uh big data to uh and compute to
essentially be able to do simulation of
design, right? So, it's like how do we
like we have a tool this is kind of
weird uh and this is a very this is a
trash version of it. We have a tool
where um because there's so many
different editing modalities based we
continually feed back into this um into
this model what works the efficiency
what offtargets were like unintended
consequences happen we keep feeding back
in the model and so we've got this like
u uh so this internal software product
that we built where it's
basically give so we can now put in
those guides and put in the designs that
we're trying to to design and it'll it
will recommend what different editing
modality which check which will and
predicts what what is the highest level
of efficiency it thinks that we will do
based on that combination. Did you guys
train the AI yourselves? Yeah, so we
trained it all ourselves. Crazy. Yeah.
Uh are you we're doing a lot of editing.
Do you think that AI will be able to
master the wrong word but will it be
able to understand biology in the way
that it understands language? Yes.
How close are we? How long is that going
to take? I think it's um it depends on
focus and funding, right? Like, you
know, people wanted to write, you know,
jokes and and essay papers, right? It's
like if the same if the same amount of
effort were going into AI um and
training large language models around uh
genomes, you know, I think that that
would be something that could be done in
the next 5 to 10 years. Whoa. Yeah.
Uh, and once we once we do that and we
have uh the ability to uh do full DNA
synthesis, like be able to write
everything that we know we need to write
with a high degree of confidence that
it's going to work or multiplex edit on
a scale where we're making thousands of
edits at once with 100% efficiency. Once
we're able to do all of those things and
do all those things really really well,
then I mean I think that that the field
of synthetic biology, the ability for us
to do accelerated directed evolution,
the ability for us to, you know, truly
guide biology the way we we want and not
just based on, you know, environmental
factors and random chance, um,
completely changes. Put your sci-fi
writer hat on for a second. So, we have
the five to 10 years of training the
model. The model is now for people that
don't know, it's probably worth going
into a little bit of alpha fold, protein
folding, how far we've already come on
that. Um, but on the other side of that,
so we map all this stuff out,
what starts to happen? Are we like, oh,
we uh take a 25-year-old and we make
sure that they can withstand radiation
and we send them to Mars. We take a
78-year-old and we make them look 26
like Yeah. So all the above. So I I
think that we will look at being able to
to engineer things. First of all, give
me uh protein folding. What makes you
believe that this is going to happen?
Why I'm going to why I believe because I
think we're we're getting really under
we we are we are getting so good at
understanding like how a protein how a a
specific thing that that is made from
from our body how its actual shape is
and how it binds. You you've probably
heard about how different blood types
and other and other uh uh reasons why um
uh certain diseases like COVID there's
there's different ways that that based
on how it actually folds, how it fits
together. So think of it as like a
complex puzzle, right? So it's not a
two-dimensional puzzle, it's a
threedimensional ple puzzle and how you
put those things together. They can only
bind or fit together certain ways and
it's based on how the folding how the uh
protein is folded. We've gotten so good,
we not we as Colossal, we as humanity
have have gotten so good um uh with
Rosetta and Alfold and all these all
these tools that we are now
understanding the shapes of proteins so
that we can predict what it will bind to
and how it will bind. What do the
machines look at? Are they looking at a
sequence of letters and they say, "Oh,
this will create a protein that looks
like this." you're looking at a sequence
of letters to that uh that then will do
that prediction and then I don't know
what the uh functional assays are that
they then do it and the molecular assays
are to to verify it but it's highly
accurate. So we did we went through um a
process with a with a gene called L
coral specifically for which which
drives size um or it's known to drive
size. It's not the only thing in size.
So it's not like once again some of
these things are multigenic in nature
meaning that we can't just like oh we
can't just make the super eloral and
then you you have like an elephantized
chihuahua right like you just can't do
that it's it's it's a lot of these genes
working in concert together that
produces some of these phenotypes but we
know for a fact that a truncation in el
coral um in and how it folds looks uh uh
uh in great danes uh is different than
that in chihuahua for example right and
we know that drives
some percentage of size. We don't know
if it's we don't know if it's like 100%
or 200% or 500% or 20%. But we know it
we do know um from looking at mice from
looking at at doing a a canid analysis
we know from years of scientific
research that that gene specifically
with this truncation creates a larger
sized uh canid like we know that. And so
when we went to look at the phenotypes
for the direwolf, we looked and said,
"Okay, what what is the L coral variant?
Does it match the greywolf? Does it
match the Chihuahua? Does it match a
great dane?" It looked a lot more a lot
more like a Great Dane than it looked
like a wolf or whatnot. So, you know, so
we put that variant in, right? And so
that truncation in that in that gene uh
you know, increased, you know, we have
much larger wolves than typical wolves
at the size and the age that they are.
So we think that you know that was um
that truncation helped lead to that. Did
you pull the one the exact Great Dane
version of it just to make sure that it
would be compatible? The direwolf
version of it. So because direwolves are
because there was debate on this which
um once again this is an educational
opportunity right so like when we there
was a paper that came out in 20 uh it's
like 2021 2022 because direwolves were
popular because of Game of Thrones and
they were only able to get and B Shapiro
our chief science officer along with you
know another like 30 incredible
scientists were on this paper that
basically that that the results of the
paper because they only got 0.15x so
they only they didn't get a full read of
the genome So they they only got about
15% of of the genome. They didn't need a
full pass based on the sample they had
at the time and the technologies that
they were employing to do it. They they
looked at this and the conclusion that
they came to uh in where a direwolf fell
was it was somewhere between a wolf and
a jackal.
I don't know if you've seen though
sometimes the results of those
discoveries when they make it to media
are well and this was not us. This was
years ago. Direwolves weren't wolves.
They were jackals. That's not what the
paper said, right? I was involved like
like I was running another business. I I
didn't even know about this at the time.
Didn't didn't read the paper, but that's
what the headline in a moderately
reputable uh uh news magaz news outlet
of today said and is like that
direwolves jackals. And so still to this
day, so so now, so fast forward,
Colossal has 500 times, not 500, like
500 times more data about two different
variants or two, two different samples
of of direwolves that have 60,000 years
of genetic divergence between our two
samples. Whoa. That's a lot, right? What
people don't realize is that there's
more genetic divergence between sample A
and sample B, that colossal sequence,
than than sample A and today because
it's 12,000 years, right? And so it's
like it's very it's it's very it's very
very simple, right, to to to see that.
But what what people don't realize so we
we took all that data we went back to
that initial author list on that first
paper and we built a phoggenetic tree
and we said on the phoggenetic tree
where dos fall and it turns out that
they were like everything in life they
were a a hybridization of extin of an
extinct lineage of canids that were
closer to wolves. Mhm. That doesn't mean
that the paper that came out years ago
that said we don't the result was we
don't know if they're closer to wolves
or jackals. It's inconclusive, but it's
somewhere between there. Uh it doesn't
this is just more data so that we know
oh it's skues more towards the more
towards greywolf lineage than jackal
right and that's kind of what everyone
thought but you know the misguided
headlines from five years ago because
Game of Thrones everyone got so excited
but then they were like oh direwolves
were like these are these are fantasy in
the show because they're really jackals
and they're jackals and that got into
like the cultural like knowledge base of
of the world and it was just it was just
you know misinformation information.
Yeah. And so now and so we we literally
get comments we're like, "These guys
don't know what they're doing. They
started with the greywolf. Grey wolves
aren't closer." It's such a weird
reaction. Look, I fully understand it's
what humans do, but it it was a weird it
was it was fun. What a strange way to
respond in the face of what is possible.
Yeah. Um Okay. So, why not clone? Why
not go grab like I get maybe a direwolf
there's nothing where you've got 100% of
it but with things that you could get
100% of the sequence is cloning better
at that point. Well, you need a living
cell to clone, right? Because you're
taking a living nucleus and taking it
from a living cell and putting it into
That's the only way to clone. You can't
clone off of a DNA strand. No, but what
you can do this goes into the um this
goes into the artificial egg and um uh
you know with mitochondrial figuring out
mitochondrial rejection and then DNA
synthesis. So instead of synthesizing a
big block, what if you could synthesize
the whole block, right? And so I think
eventually we'll get there. It's
unclear, but but this is where this is
where it's funny that we talked about
eugenics. I made a I made a joke on a
podcast that wasn't wellreceived by the
scientific community. Um, which is like
I've never seen the scientific community
so eugenicsy about a wolf before because
they all wanted to argue over hilarious
because it's true. It's like, you know,
if you look at the phoggenetic tree and
you look at all life that's ever been
like like the tree of life, right?
Better known as the tree of life and you
look at all of it, 99.9 to like the nth
degree is based on where it lived, when
it lived, and what it looked like. That
that like we don't have DNA from
anything. We don't have we don't have as
I mentioned they don't we don't have DNA
from the animals that live today and so
this idea of purity of direwolf is just
a psychotic perspective that's just
weird. Okay. Well, let me steal man
their argument for a second. This is a
Ferrari
uh chassis on top of a Toyota engine.
How is it not that? try to go pet that
uh that Toyota engine and see how it
works out for you because just the way
that they I mean if you look at just uh
their size and where they were located
one our genetic donor was where we we
got it from where they're located. So we
took greywolves which were American
greywolves. We didn't import them,
right? So we kind of met that standard
and we looked at uh and if you looked at
just the uh so that that kind of fits
where they came from. We we chose the
same wolves from the same place that
they came from, right? We didn't import
magic wolves. Uh number one. Number two,
if you look at kind of the phenotypes,
here's what we know about direwolves.
This is all we know about direwolves.
They were 20 uh based on the fossil
record. I'm gonna tell you something
else that we didn't know until Colossal
did it, but no one seems to care. uh one
is that they were 20 to 25% bigger. They
had a larger cranial facial morphology
which implied that they had a stronger
bite and based on their bone density and
their uh bone specifically in their
shoulders and in their and in their legs
most likely uh you know they were they
were heavier, right? So they they were a
stockier animal. That is what we know.
So if we just did that, I would argue
that from a functional deextinction
perspective and the same way that we
classify 99.9 literally indefinite
species in the tree of life, they would
be classified as direwolves. But we
didn't. What we also did is we said,
"Oh, this is interesting. Their coats
were white." And we know that because we
looked at the data and we found in the
DNA, but there's no direwolf. There's no
frozen direwolves. There's no frozen
samples of this. People thought because
of a paleo artist's rendition of that
paper from uh 5 years ago that they were
red because they were like, "Well, we
got a reddish brown because they're
like, we got to make them look more
like, you know, jackals because they're
closer to jackals because someone read
an article that was inaccurate at the
time." And even people in that paper
will tell you from that paper, that
paper didn't say they were jackals. I
know it because these the people that
wrote it because they were like, "We
wrote it. we're the ones that that
didn't that wasn't in the conclusion of
the paper. And so um so we t actually
took it we actually took it a step
further in two categories. One we we
found uh kind of like in the woolly
mouse the hair variant that drives you
know the hair and so it they have a much
thicker hair. I mean when they were born
they kind of look like baby polar bears.
They're amazing. They're they're like
these cute little fluff balls. Um and
then which there's pictures all over the
internet of them. And then if you look
at their hair now it's super thick. It's
it there's like waves to it. It's it's
amazing. They almost have this like
ridge line to them and like mane across
their back which is amazing. We had no
idea of that from the fossil record. We
also didn't know they were white. We had
only misinformed the public and said
that they were red or brown because of a
conclusion of a paper that that the
paper didn't come to that conclusion
anyway 5 years ago. And so I would argue
that not only are they direwolves uh
based on how
99.9% of every species is currently
classified on on the planet, but you
know they're even more direwolves than
anyone could predict based solely on
those phenotypes because we identified
things in the in the DNA.
And
so you guys, how many like DNA blocks?
So we edited you put in So we edited uh
20 uh uh uh 14 genes with 20 edits.
Okay. And starting with a greywolf.
Starting with a greywolf. This is our
genetic donor because they're 99.5% the
same genetically. Got it. Comparable to
what Asian elephant is to mammoth. H
okay. And most people don't realize
this, but Asian elephants are closer
related to mammoths than uh Asian
elephants are to African elephants.
How do we not end up in a Jurassic Park
scenario where uh we made something and
it has an unintended con
unsequence in the ecosystem. Well, it's
a great question and for all these
species we go through like people think
about um and I don't think that was the
exact permissive Jurassic Park because
they weren't rewalding the Jurassic
Parks, right? like they didn't have a
they did not have an ecosystem
restoration or a um conservation subplot
that unless that just got like you know
on the cutting room floor but so I don't
think it's exactly the Jaspar scenario
but uh but to the ecosystem restoration
uh unintended consequences is we have to
measure all this stuff right anybody
that tells you like hey you can
introduce this species back here we'll
have zero unintended consequences or
we'll have this this intended
consequence is just wrong but the
beautiful thing about rewing is that
colossal People kind of, I think, once
again, give us too much credit. They're
like, "Oh, they're just going to rewild
them." Like, we we have a foundation
that we raised $50 million for, and we
have 48 conservation partners, including
some of the top reing partners in the
world. Specifically on the direwolf
project, we actually work with some of
the gentlemen that ran the Yellowstone
reing program for greywolves back into
Yellowstone. So we, you know, if if if
the uh government, indigenous people
groups and uh ecologists and everyone
wants to rewire wolves at some point, we
would work with them to do that, right?
Like we built a rewing plan because I
think anytime you do this, you should at
least be like, what are the
possibilities, right? So you should
think about that. But they live on a
2,000 acre secure expansive ecological
preserve that are monitored 247. They've
got 10 uh that's certified by American
Humane Society. They've got 10 full-time
uh uh people to attend to them. Uh
security. Uh they've got uh storm
shelters, an animal. We built an animal
hospital in case anything happens to
them that we can treat them right there.
They don't have to like it's not like
calling 911 and then having to take them
to the hospital. We have a hospital and
a full-time vet there. Are you guys
going to make more direwolves? Uh so we
are going to make three to five more. So
we want to have a pack size of roughly
five to eight because will you
impregnate the ones that you have or No,
we'll engineer them. So, we're gonna
engineer them and they'll be birthed by
a greywolf or by No, no, by a domestic
dog. That's crazy. What kind of dog? Uh,
so they're just large uh hound mixes. Is
that dog like what just came out of me
or do they just No, they nurse them and
everything. They get the colostrum from
it. Everything they they nurse them and
take care of them. It's great. So, to
them it's like I just have weird looking
kids. Yeah. Or super cool kids. Depends
on how you want to look at it. But then
but then all of our uh you know um and
then we work this is kind this to me is
kind of fun is that uh you know because
we work with American Humane Society so
well uh all of our uh dog all of our um
uh dog mothers our seriates uh get ad
adopted through an anonymous adoption
program to their forever home. So that's
kind of fun. There's people out there
that have um you know uh moms of
direwolves but you don't want them to
know for some reason. We It's just like
I feel like that's like a um spectacle
of the animal, right? Like not saying
that anyone do that. Like I think that
if I if like you know people I know
adopted them, they'd probably be like,
"Oh, that's cool." And they may tell
their friends, but who know? I mean,
people are, you know, we're all weird in
our own ways. Someone could try to
exploit that, right? And be like, "Oh,
this is the mother, you know, like in
game like mother of dragons. This is the
mother of direwolves." And it's like, I
don't think that's good for the animals.
That's Oh, man. See, what what what
people sometimes don't think about is is
like if my job was to raise money and
hire an awesome team and let them just
do stuff in the lab, this job would be
very boring and easy. What's interesting
about this job is that we have
conversations like this, right? We we go
meet with the government and we go meet
with indigenous people leaders. We work
with, okay, how do we adopt our
surrogate moms, right? There's so many
com like nuances to how we think about
this that that's what sometimes is hard
that unless you like say you know when
we when we launched the direwolf story
we worked very closely with a handful of
people Rolling Stone Time magazine um uh
the New Yorker and others even though
the New Yorker broke embargo and kind of
screwed us for a while um they which was
very very painful week for us um because
we didn't have our website up we didn't
have our our scientific paper hadn't
been submitted did they apologize Or did
they do it on purpose? Well, they
definitely did it on purpose because it
was in print. You don't like It's not
like someone like spilled coffee and hit
the publish button, right? They're like,
"Oh, yeah, that looks good in in uh in
in design or whatever and and that's
going to the printer and that's going to
newsstands tomorrow." So, that doesn't
feel like an accident. Yeah. So, they
have not publicly uh uh apologized. Um
but we'll just never work with them
again, so it's fine. Yeah. Interesting.
Uh, okay. Let's bring this all back to
humans. Okay. So, you've got Brian
Johnson, the don't die movement. Yeah.
Longevity, Escape, Velocity, how Peter
Dandis, you actually have a lot of
people here in LA. Yeah. Yeah. No, Peter
I know well. Brian I know wellish. Um,
and I I don't think Brian gets enough
credit because like people like love,
you know, we live in a we live in this
like armchair Twitter X of people where
they just want to criticize. They don't
want to do anything, right? What's that
famous quote that's like, you know, uh
the the that you know, people are either
on the field or in the stands and
there's a lot more people in the stands
bitching with people on the field. A
lot. Yeah. And and like, you know, I'm
not going to do everything that Brian
does because it's a that is insanely
hard lifestyle, but I'm glad that
someone's doing it. Like, so that's the
way I look at it, right? It's like
Brian's not hurting you. Yeah. I don't
understand. Brian's but there's this
weird backlash towards Brian. I don't
agree with everything he's doing, but he
he's not like saying, you know, hey, you
have to do this. You have to do this.
You know, he's not showing up with like
ski mask and like making people do red
light therapy. And so it's so that I
kind of would get behind that. Let's go
buy a new business model for it. I love
red light therapy. But like I think it's
unfortunate when people like Brian get
backlashed. It's like this guy is taking
biomarkers to an extreme and he is
running an experiment that we that
someone in science should be we should
be lucky. we should be thanking him that
he's doing this. Yeah. Kick off as much
data as you can on and I'm not saying I
agree with all of his uh the results
from it. But, you know, if you sit down
with Brian, which sounds like you have,
he'll he'll tell you like the number one
thing is sleep. Like the first he's like
you don't even have to get weird on like
your eating schedule or workout
schedule. He's like you really need he
said the number one thing to me is like
what your resting heart rate is and
variable heart rate is going into sleep
in that transition mode to sleep. Yeah,
that makes a lot of sense. Do you think
we will actually hit longevity escape
velocity where science adds a year of
life for every year I live? Yes. How
soon? In the next 20 years, 15 years.
Yeah. What's going to be what are the
sequence of breakthroughs that we need
to get there? Um I think the biggest
thing that we need is so there's been a
lot of stuff on cleansing blood. A yo, I
didn't see that coming. Yeah. So this is
this is a a newer thing. Super kidneys.
No, you you remember the uh Silicon
Valley Blood Boy episodes? Yeah. Where
they'd like circulate the blood? Bro, if
that was real there. So, get your boy
hooked up. So, get ready. So, there is a
not that service, but there is a service
that's not currently available in the
United States. I have not done it and
nor do I endorse it because I haven't
done the research on it. Sure. But, uh
Gary Brea has talked about it. A handful
of other people have talked about it.
But where they basically do, it's like a
super dialysis machine, right? So it
it's a
transfusion, takes all your blood out,
runs it through a system that's
basically like this super mesh, and it
takes out like precancerous tumors, old
cells, just all this stuff, and then
pumps it back into your body. And it's
supposedly like your your the blood that
goes back into you is like the
equivalent of um of like a 26-y old. So
that's that's that's Do you know what
that's called? Uh no, we can find it. We
just called my wife did something like
that. Peter Demandis knows it too. I I I
forgot. So So I think you're going to
have like the cleansing side of that
side, right? Um and then and then I
think what you have and so I think
that's here today. Another thing that's
here today, if you look at like the the
uh uh big causes of of death in America
is like you have heart disease, uh you
have diabetes. uh uh many people think
that Alzheimer's and dementia are
basically type three diab di diabetes
and just led from inflammation right uh
which is diet and a lot of things can be
driven by that um but if you look at
things like cholesterol we have
medications now that you can absolutely
lower your cholesterol and not have
heart buildup so my Gary and I have a
site I'm learning more Gary's got really
good ideas but you know my my LDL is in
the 30s which is and what they've shown
is anything under 60 you start to
reverse any plaque or any buildup in any
of of your veins and arteries. And you
got there by taking something. Yeah. I
take a I take a Rapatha which is made by
Amgen. I don't an injection. Yeah. And
so so so you you have things like that.
Then you've got things like metformin,
right? And you got other things to lower
your A1C. So I would argue that we have
roughly defeated diabetes and major
forms of heart disease today, right?
Like I I I I truly believe that like you
know if you if you if you don't eat
sugar, if you do if you eat a moderately
healthy lifestyle and you take certain
therapeutics, we have done that. I think
the GLP1 some people hate the GLP1s, but
I think the GLP1s are really really
interesting. They're starting to show
the that uh because you have these GLP-1
receptors not only in your brain but all
over your body. You're starting to see
things like uh kidney function improving
uh muscle function and heart improving
uh from this. And these are all nonBBMI
related uh studies. So um there are
people that think that GLP-1s will lead
to uh uh lower uh um Alzheimer in and
dementia, but those are So do you think
the punchline is going to be a bunch of
exogenous stuff or do you think the
punch line is going to be AI reads your
genome, we get very good at this? I
think I I think that the punchline for
longevity escape velocity which is one
to one is a combination of therapeutics
and then I think eventually kind of like
the function blood test or others it you
will get a combination. So instead of
taking like a patha shot a you know
GLP-1 shot um whatever the frequency is
in whatever the doses is you take your a
metformin I think you'll have a single
like shot I think it'll probably be
administered via shot versus orally.
That's really interesting. So I think
you'll have like a combo shot that is
like your longevity shot that keeps all
your biomarker that gives you like Brian
Johnson level biioarkers. So I think
you'll have that. Um and I think you'll
have that before we have you know uh the
point that we are you know do doing full
kind of like AI and I think AI could be
applicable to that but I think
eventually we have um AI that you know
tells us like hey here's XYZ you know
edits to your body that that is because
you're close enough to that problem to
know how hard that's going to be. Yeah,
it's really hard. But I mean there's
people out there like, you know, um I
mentioned Bob Nelson. His he's got
numerous companies including like Altos
Labs. George has Rejuvenate Bio that's
doing insanely cool stuff. And these are
the there there's biotech companies
right now that are in pre-clinical
trials and some in clinical trials that
have, you know, um potential like
massive potential for longevity.
It's interesting. Remember longevity? I
I think that that there is a um a a a I
think that there is a and I think AI
potentially could help solve this but
there is a um uh very large drop off in
that transitional data between like the
world's expert. So if you take like the
world's expert in like epigenetics or
the world's expert in genome engineering
like George, right, and he dies, you can
go read all his research and you can go
but unless you've trained in AI to think
like George, be creative like George,
have every thought that George has had,
then then you're going to lose even if
you have the next best geneticist in the
world, there not like there's going
there's still like this generational
loss of data that occurs, right? Because
there's problems that you probably think
of or I think of that we don't write
down. doesn't mean that that on some
like background thread they're not
running all the time and you're not
thinking about them but you probably
just haven't like uh I'm going to go
solve this problem or I'm interested in
this and I've accumulated data based on
this interest but you haven't like put
that into something that if you were
gone that I or someone else could go
read and so so I think that I think AI
can help with that but I think that if
we can increase some lifespan and
increase even if we don't get to full
longevity escape velocity I think that
coupled with AI that coupled with hell
span so that you are more productive to
society longer, you eventually get to
the point that it's where it's just
immortality.
What are the roadblocks that you see
coming with what you're studying right
now? Like when I look at AI, I say,
"Okay, well, if there is no um
computation problem, meaning we have
enough chips, if there's no energy
problem, then I don't see why we don't
get to ASI."
Um do you see like are there similar
things like that with what you do where
you're like well we'd first have to be
able to solve for this problem. So, so
for us with what we are trying to do at
Colossal, there's no like I there's no
like FTL problem, right? Like it's like
we don't have to like solve faster like
we're not like to get for to us to get
to our alpha centuries, we don't have to
solve faster than light travel. We just
it takes us a longer time and it it
costs more money, right? So ours is an
efficiency ours is more of an innovation
uh play than an invention play, right?
Like we are trying to innovate these
technologies. That doesn't mean we're
not trying to invent new things along
the way. So we are doing some discovery
but you know our bank is that AI and
some of these other access to compute
will accelerate the efficiency of these
technologies because we are doing it
right now on a rudimentary scale and you
know uh when we started the business
people were making one edit at a time.
Our thyloine project we've made 300
edits in the cell line we haven't taken
that to term yet because we haven't
solved the cloning the smaxel nucleot
transfer process in uh in marsupkills
yet. We have a team that's working on
Why is it different? It's so once all
these on model species are are are
unique. So So for example, this outer
shell in a marsupial because you think
giant big elephant, right? Or or human,
right? And then you see these like
little bitty dun art and you'd probably
think, well, it's probably got the
weakest little cell ever ever. It's got
this zona palooa, this outer shell. Zona
Palooa Palooa, which is insane. That's
the literal name. Yeah, that's the
literal name. So we we we have created
this uh laser assisted uh system for
drilling in uh using a laser in so that
we can extract DNA. We're now doing it
with computers and and robotics and AI.
It's pretty in computer vision. It's
pretty awesome. Um but we we hit it with
the laser doesn't move the outer shell.
Whoa. Doesn't move it. We we we go up to
like 11 or whatever the highest dial is
and then some and it just the cell just
blows up because you just like hit it
with like the death star, right? And so
what and so and then if you use and the
reason you use that because if you use
like a a needle if you're going in and
you're and you're jamming the needle
it's like you're you are damaging the
DNA in that, right? It's like the fact
that like Dolly even worked with like
all the blunt instruments is a ma is
magic in itself, right? I don't think
the the the team at Edinburgh and the
the the team that did Dolly gets enough
credit like that is a miracle in itself.
Like people don't realize that. They're
just like, "Oh, there's moved to sell."
But given that they were like like
they're like flying using like you know
nails and hammers and wrenches like it
was incredible what they did. And so now
we it's come a long way. But so so for
specifically that uh we actually had to
invent a uh so we so a lot of times
we'll even make our own tech. So we
actually uh make our own needles in many
cases our own glass needles. So, we had
to make it out of quartz so that we
could so that we could vibrate it at the
right frequency so that the vibration so
that the because the glass couldn't take
it, but the quartz could because it was
hard enough and then that quartz needle
at that vibration could pierce into it
uh because the laser wasn't strong
enough and when we dialed it up too much
just destroyed the cell. That is
fascinating. But see, those are the
weird things that we have to solve. When
people are like, "Oh, it's it's it's
just a it's just the most genetically
modified animal on the planet." I was
like, "Yeah, but we had to do a bunch of
other stuff." What do you think is the
upper limit of the number of edits you
can make without without uh DNA
synthesis? Without just synthesizing it?
And you mean you mean through multiplex
editing? Yes. Thousands.
Maybe tens of thousands. Whoa. That's
all you really need because you're doing
target. I mean the the more I I would
argue though once again this is coming
from a software perspective. I all get
[ __ ] about this from a biologist but the
smarter you are on computational
analysis the more you know the less
edits you make because every edit you
make is going to uh to to have some
level of risk because things are
redundant or they're not necessary to
the final form. Yeah. And and and the
thing that people don't realize like
when you get a genome sequence of all
these letters, right? Uh and you run it
through the reason why you need so much
coverage. We talked about like uh I
think we ended up having 13 or 14x
coverage. That means that we had 13 or
14 full reads of the genome uh for the
direwolves. Whereas before they had
0.15x they didn't have even a full one
read. The reason why you need that is
because uh and you need to have at least
probably about six or 8x uh I think to
do the extinction properly because um
your b the machines aren't perfect. So,
and the DNA is degraded. So, it's giving
you at every single site where it's
giving saying this is a XYZ letter,
right? This is a C or a or a G or
whatever. When it's doing that, it says
these um it's giving you a probability
score that this is a C. It actually
isn't 100%. So, they're like, oh, it's
like 50%, 99%. So, the more you do, then
the higher probability is that you get
that, right? And so even when you we as
humanity get to the point that you can
synthesize a full genome and this is
where not to argue that the weird
eugenics points again uh about like what
makes a mammoth a mammoth even if you
get to an end toend in into mammoth
there is no end toend mammoth genome. So
right now we're doing an assembly of of
a mammoth genome, building a reference
genome, but we're using 60 genomes that
we're putting together, right? And so
then you've got to go look at that and
then you've got to go say, so even if
you synthesize all of that, you're still
going to have holes. So we have a we
announced in October of last year that
we have a
99.9% complete genome uh for uh for the
Tasmanian tiger orene, which is amazing.
How are you picking what animals you do
in what order? Oh, um, well, like a
thyloine. What the? Yeah. What? So, uh,
it sounds like a period in time. It does
not sound like an animal. Yeah, it's
it's it's awesome. There's a Jurassic
and we've seen a thyloine, right? Uh, I
think so. Can we show you? Yeah. Pull it
up. Yeah. Can you show It's amazing.
It's like a zebra. It looks fake. It's
like a zebra wolf. Looks like a kid was
drawing and had a seizure halfway
through. That's what that looks like.
It's awesome. And there's videos of them
and everything. Are they going extinct
or are they No, they went extinct in
1936. Yo, so they're awesome. They're
they're Okay, but why why this one? Oh,
so uh specifically why the thyloine is
we hunted it to extinction. So the
Australian government put a bounty on
its head and actually paid uh farmers
and hunters to kill the thyloine.
Doesn't that tell you this thing's a
pain in the ass? No, they got a bad rap.
They the sheep farmers which were
stealing each other's sheep, killing
each other's sheep for competition blame
the thyloine. There's no data. Like I
I've spent a ton of time in Tasmania. I
spent a ton of time with all the
thyloine experts and the researchers.
There's no data that shows that they
would ever attack a sheep or or eat
anything that size. They actually killed
mostly kind of those mezzanine
marsupials like Tasmanian devils and
whatnot. So they were Now is this a
marsupial? Yes. So yeah. And like the uh
and like the wombat it's got a backwards
pouch. So if you think of a lot of most
marupials have a patch like you know up
front and joy's in it. Um but it
suggested this was a and there there are
thyloine dens that have been found. So
it was a burring animal. So it actually
burrowed and if the pouch was facing
forward it would fill it full of dirt
and kill it just like the wombat. That's
so yeah cool. Yeah it's super cool. But
but so so we picked that species because
one um we as humanity made it go
extinct. two, there's great DNA. It this
is not an ancient there. We had to go
build the the genomes, but this is a
genetic engineering feat, right? So,
there's 70 million years of genetic
divergence between a fat tail done and
which is like a marupial mouse and we're
turning into a marupial wolf. H which is
crazy. Um that is crazy. But it's
amazing. Yeah. So but once again going
back to so you pull just sort of from
what's going to be the most fascinating
what do we have connections to that we
could rebuild what is possible right is
there a reason to it in this case we
removed it there's not been another apex
predator that has replaced the thyloine
in lower Australia or Tasmania and we
know is that creating problems yeah so
there's this there's this kind there
this concept called tropic downgrading
where you basically have a ripple it's a
ripple effect right so you have this
entire ripple effect on the ecosystem
when you have a a missing predator.
We've seen this in Yellowstone. We've
seen this all over the world. And so
what's interesting for the thyloine is
that uh have you seen this? We should
look it up because it's awful to look
at. But have you seen the facial tumor
disease? No. I've heard you talk about
it. I very aggressively resist. Don't
look it up. It's rough. It's like I It's
bad. But don't you dare. I could see him
over there typing already. It's so bad.
Yeah. It's like out of a horror movie,
right? I'm going to have my producer
Drew arrested. Yeah, it's it's bad. But
but what's interesting is that uh if you
look at predators, they're they have
some terminal internal calculus of
energy expenditure, right? So not every
pursuit of a kill results in a kill. And
we've seen this in Africa with cheetahs
and and with big cats, if they go so
long without making a kill, they're so
unsuccessful after making a kill,
eventually they're too tired to pursue
it again and they die, right? So there
is some internal biology, you know,
model that they use uh to decide whether
they're going to go after a kill. But
what's interesting is so they typically
pick the young, the old, and the sick
and weak. And so this goes back to the
survival of the fittest. This is great
from a genes perspective. And so there
there's been people like Dr. Andrew
Pasque and others that have said that,
you know, if we do if if the thyloins
were here, the the there would still be
probably the facial tumor disease, but
it'd be a lot more controlled because
though like if you see a devil, a
Tasmanian devil with facial tumor
disease, it can barely it's like
walking. It's stumbling like it it can't
see well. So, it kind of attacks
everything. That's how it transmits its
disease so much because when they're
eating, they'll actually attack each
other because they're pretty aggressive
when they're in these like swarm eating,
right? Woof. Um, have you read the book
1493?
No. Oh man, dude, you've got to read it.
It is all about how humans have
transformed the earth in ways that they
were never even aware of. Yeah. Like I
am almost certain the following is true.
Uh, earthworms are not indigenous to
North America. That's amazing. Oh, I was
like, what? What? Uh, and they were
brought over here because they would use
dirt as like a ballast in ships. And so
then when they got here and they
unloaded everything or when they loaded
things up. Yeah. So they would bring the
dirt over cuz they knew they were going
to load a bunch of stuff up. Uh, so when
they were about to load up, they would
dump all the dirt and in the dirt were a
bunch of worms and they have spread all
across North America and just completely
changed the soil and all of that. Uh,
same things with potatoes. Like potatoes
are from We're bad at this. We like you
probably heard me talk about like the
cane toads like we introduced Australia
and they're killing all the marsupials
because we because the marsupials didn't
evolve next to them so they're eating
them and dying. Yeah. Is that a big
driver for you guys? Just like
ecological balance. Yeah. So we think
it's interesting, right? So like you
know in the cane toad project we have
made a single edit so one edit and it
confers 5,000 times resistance to
cantoxin which is super cool right?
Because then you can make like super you
can make you like one genetic you can
make a handful of these of these uh
super quo which is what's eating them
and dying and then and other animals are
eating them too and dying but not as
much as the quals because frogs and
toads are primarily their diet. Well
then these super quals can get that
population under check and then the
coals can live and then it can also have
this halo effect of getting rid of the
cane toad. You have that even a bigger
halo effect of protecting other
marsupials. Have you come across any
cool things like that in humans? like
you could have an increased um tolerance
for oh uh plastic which I know you have
a great punchline for or cuz we've got
microplastic problems we've got
radiation yeah there I mean I I think
that I think radiation and cancer
suppression are two really interesting
areas that you know I think that data
that we see from some of our work and
that we see other people working on I
think that that could be minor changes I
think we'll have a a cancer vaccine in
the next not we as colossal But I think
humanity will have a cancer. No. Why a
vaccine? Because it's about 200
different types of diseases. Like cancer
isn't one thing. It's about making it a
vaccine work because then it just trains
your body to go attack it.
Interesting. Huh. There's a lot of
people obviously focusing on cancer
which is great. But um you know going
back to like longevity escape velocity
you know that and you know Alzheimer's
unless we can see if Alzheimer's and
dementia is directly connected to um
environmental accelerants
um uh inflammation. H so talk to me
about this enzyatic breakdown process.
The company's called break or breaking
breaking. Yeah breaking.com.
This is great. Yeah. So um Sucana and
Vascar and her incredible team at the V
institute actually found this microbe
that uh when put in uh with only adding
salt water naturally occurring or
naturally naturally occurring naturally
occurring uh it could it would uh break
down and degra visibly degrade at the
time we didn't know why any type of
plastic that was in there. So, if you
had a vial of salt water and you put in
uh and you put the enzyme in or put the
micro in that creates this enzyme and
you put nylon, fishing nuts, anything in
there, right? Uh I guess that's mostly
fish nets, any other types of industrial
plastics at different rates, but it
would take plastics that have never
degraded or only degraded a little bit
um or would degrade over 800 years and
you know in a couple years uh completely
degrade them. M. And what it was doing
was it's breaking the chemical bonds.
That's why I named it breaking. It
actually breaks the chemical bonds in
plastic. So it turns plastic into just
normal biomass into normal biomass. That
is awesome. It's so it's really really
cool. And so what we did is we we uh one
of the things that's been this halo
effect of colossal that's been pretty
interesting is that because we have a
computational biology core right and
software and we we know how to look at
at at at genomes in this type of data
and because we have genome engineering
capabilities a lot of researchers are
bringing us really cool projects where
they need AI computational analysis and
genome engineering and so they came to
us and said hey we have this micro we
think it does this can we do an analysis
so we did a lot of analysis work and we
looked at it and we
Yeah, like it's creating this enzyme.
It's actually a couple things working
together to create this specific enzyme
that breaks down and everything we threw
at it, it broke down. So then we're
like, how do we make it hungrier? How do
we make it where it exudes more of that
enzyme? That's a genetic engineering
challenge. That's not a um that's that
that's not just a directed evolution
challenge, right? Like we're not going
to be like breeding this, right? We're
going to we're going to be accelerating
it through synthetic biology and
actually editing the genes that that
make some of those enzymes. That sounds
like a huge breakthrough. Is it is it a
big breakthrough? Yeah, it is. I mean,
right now, you know, our goal is, you
know, I mean, my goal is to get it to 24
hours. They can break down any plastic
based on the amount of distributed
surface area that you need, which is
really interesting because a lot of the
plastic degradation tools out there are
just making smaller plastics. They're
not actually disintegrating the
plastics, number one. They're also not
um a lot of them are having you they
have to be heated or pre-treated with
the chemical. Some of those chemicals
are worse than plastic. Great. Yeah. So,
it's not helpful. And so, our thought
process was um so we did this. We we
incubated it. We got it where we wanted
it to be. Now, it's a it's it was 22
months. Now, it's about 18 months. Um
but our goal is to get it to 24 hours.
That's still massive. It was 800 years
to two years is already huge. And we're
and it's we currently have um uh uh 14
it's either 14 or 16. I should know this
because we just had a board meeting on
it but 14 or 16 pilots. So we're working
with I can't say the names of I mean I I
could but I don't think I'm supposed to
but screening water screening things
that come through water uh and then and
then treating it. Number one. Another
area is textiles. Like a lot of the a
lot of the stuff that we make,
especially like in fast fashion and
others, has um you know, huge has
there's a lot of plastics in our
clothing that most people don't realize.
Yeah, that it's really starting to make
my radar now. Yeah. Um yeah, that one
feels like it's massive. And we're going
to we're going to rate um Yeah, it's
going well. So like the So So I think I
think about these companies in kind of
like, you know, three categories, right?
Like is there a general desire for them
in the world, right? Is it is it is it
solving a use case and is there a
general desire because those those can
be the same or those can be different
right I think everyone generally agrees
that bio loss of biodiversity is bad and
everyone generally agrees that the
accumulation of plastic everything from
the oceans to our bodies is bad right
and I think that everyone agrees that
that doing really cool things and
inspiring kids for science and in both
those cases do it right then those
things are good and so so we in both of
these projects or companies you know we
we those all being all science. Yes.
Well, then the science has got to work,
right? So, like we have to be able to
create a woolly mouse. We have to be
able to create a direwolf. We have to be
able to create, you know, uh degrade
plastic from 800 years to to to two
years. And then it's really just the
business modeling side, right? And so,
you know, um I've kind of fallen coming
from software and not being a biologist
at all and kind of being good enough at
learning it. Uh, I've just fallen in
love with this idea that the combination
of AI, comput, and synthetic biology, we
can solve a lot of really cool
challenges and um and and I think we can
do it in cool and interesting ways that
gets people excited, right? Like if you
have a plastic degradation company,
that's awesome. If you then take the
plastic degradation uh company uh work
in a bunch of textiles behind the
scenes, that's awesome. But then if you
do that and you bring in like you know
people you know from like uh Noah or
from like Jacustoe's foundation or
celebrities that care about the ocean
then you're going to bring awareness to
science in these problems. And so you
know one of the things that I think that
that we do with our companies which is
interesting is like we're solving really
hard things but we also try to do it in
a fun flashy way that gets people
excited. What's the coolest challenge
you want to solve um with these
companies or ever? Let's go with ever.
Um I I think it'd be really interesting.
So I I don't want to be in the
therapeutic space. I think there's
enough people in in the therapeutic
space, but I would like to if I could
have two solutions is I would like to be
able to engineer uh plants into doing
whatever we want. Um George and I joke
about this treehouse concept of like,
you know, not to get too weird and
hippie-ish, but it's like why can't a
tree grow in the form of a a dwelling?
And why can't it like we weren't on
mushrooms when we did this? This is
true. But it's like but then why can't
it have bioluminescent fungi and it like
it light everything and like you know it
would be like the ultimate kind of like
cleansing for the environment, right?
Cuz you see redwoods that live in all
kinds of environments. So So I I'd love
to be able to engineer plants in a way
that that like we're not just like
chopping down forest, but that like the
plants could build our cities. That's
like the weirdest most I'm not working
on that, but it's a really weird thing.
Uh that's one. Two is I'm very
interested in the oceans. Uh we're not
working on this either right now, but um
I I think that there is a huge
opportunity if we look at, you know, uh
the most stable uh patterns of of
weather globally and where life persists
consistently, it's somewhere between 30
and 70 ft uh under the water and it's
pretty stable there regardless of like
the ocean churn above it. So if you
could uh build tools and technologies to
to make coral more resilient and through
synthetic biology, not by just macro by
micro fragmentation, that all of it grow
faster, but it's still going to die. So
you have to do genetically modified
corals. Um if you do that and you build
closed systems for uh underwater living,
I think that's really interesting.
That's very interesting because I mean
if you if you look at if you look at the
ocean and you look at the surface of the
earth like like you could build I mean
what is the economic value of the
California coast trillion dollars? I
don't know it's a lot right if you owned
all of it. Well there's no wildfires 30
ft under you have to build closed
systems. So the same technologies that
you're developing without needing ma
major heat uh and cooling inversions and
radiation tolerance that you're spending
a fortune on cost a kilogram to put to
space. you could apply those same
technologies to Earth um and and you
know work on underwater cities. Do you
think that's going to be something that
happens in our lifetime? I mean I think
that we're going to achieve uh longevity
escape velocity. So the answer to any of
those questions is yes
because our lifetime's indefinite. Yes.
Oh god. I hope that that really comes to
pass my man. I cannot thank you enough.
Where can people follow you? Uh I'm just
on Twitter or X just binlam. There it
is. Two M's. All right, everybody. If
you haven't already, be sure to
subscribe. And until next time, my
friends, be legendary. Take care. Peace.
If you like this conversation, check out
this episode to learn more. Due to AI
and a changing global order, the world
is in the middle of the greatest period
of change ever. But because we're in the
middle of it, it is nearly impossible
for us to accurately see what's going
on. We are in the fog of war. While the
world panics over job loss in Kib.