00:00:01
If you say that evolution began
at the origin of life,
00:00:03
then you're discounting all of the
chemical and cosmic complexification
00:00:07
that brought us to that point
in which you could have the first
00:00:11
living cell.
00:00:12
How did it get there?
00:00:14
I always had this unsettled feeling
that there was something
00:00:17
missing from the laws of nature,
because none of the laws explained
00:00:22
all of the wonderful complexity
of the universe around us.
00:00:26
For us, all evolving systems are
conceptually equivalent
00:00:29
whether they're the evolution of atoms
and isotopes, the evolution
00:00:32
of minerals, or the evolution
of life and technology.
00:00:35
And so therefore may be described
by a single unifying natural law.
00:00:41
So the 'Law of Increasing
Functional Information' is an effort
00:00:44
to describe a kind of universal
characteristic of the cosmos.
00:00:48
The idea that systems over time, appear
to get more
00:00:53
and more and more complex, more patterned, more diverse,
00:00:56
more interesting, if you will.
00:00:58
And we characterize this increase in order
00:01:02
through a metric
called functional information.
00:01:05
We think information is as fundamental
a variable in the cosmos
00:01:09
as mass or energy or charge,
00:01:13
which may be a little out there?
00:01:19
So over the last 400 years or so,
scientists have come up
00:01:22
a set of ten laws,
00:01:23
that's the canon of physics.
00:01:26
And that it describes virtually everything
we experience
00:01:29
in our day to day lives.
00:01:30
The motions, the forces,
the energy,
00:01:32
electromagnetism,
the law of gravity.
00:01:35
That's it.
00:01:36
The only law of those ten
00:01:39
macroscopic laws of nature
that has an inherent arrow of time in
00:01:43
it is the so-called second
law of thermodynamics,
00:01:46
which says that over time,
a closed system's entropy should increase.
00:01:51
Basically means that the disorder
of the system should increase over time.
00:01:55
But we see the second law manifest
in our daily lives.
00:01:59
Our new shoes get scuffed.
00:02:00
You can break an egg and scramble it,
but you can't unscramble an egg.
00:02:05
As I'm thinking, as electrons are moving,
as molecule bonds are are formed
00:02:10
and broken, every one of those activities
causes an increase in entropy.
00:02:14
But that doesn't mean
that entropy is driving complexification.
00:02:19
There is some extra description
that is required to explain
00:02:24
the marvelous creativity
of everything that we see around us.
00:02:28
And also the forces of complexification
in everything else in the cosmos,
00:02:35
Go back to the Big Bang,
00:02:37
there was really no structure of any kind.
And then you started
00:02:42
forming protons, neutrons.
00:02:44
And the protons and neutrons formed atoms, and then the atoms formed molecules,
00:02:48
and the molecules formed stars,
and the stars created planets,
00:02:52
and minerals, and atmospheres,
00:02:55
and oceans, and ultimately life.
00:02:58
And now life's creating language and art
00:03:01
and various kinds of social structures and technologies.
00:03:04
And now we have computer programs
that themselves evolve.
00:03:08
I mean, that's an incredible range
of evolving systems.
00:03:13
But they all share
these three common aspects.
00:03:17
First, it must be composed of numerous,
diverse interacting components.
00:03:22
They could be atoms, they could be
molecules, they could be cells,
00:03:26
they could be individual people.
00:03:28
And those different components
can be arranged
00:03:31
in just countless, vast numbers of ways.
00:03:35
Second, it has to have mechanisms
00:03:38
for generating numerous configurations
of those components.
00:03:42
You have to have a way of mixing things
up, sampling new configurations.
00:03:47
You have lots of different ways
of arranging atoms or molecules or people.
00:03:52
And then finally you have to have
some kind of selection pressure,
00:03:56
selecting for a function.
00:03:59
In other words, they evolve.
00:04:02
The most fundamental selection
forces for us are selection
00:04:06
for static persistence, dynamic
persistence, and novelty generation.
00:04:10
And in different systems,
the selective force that matters the most
00:04:14
could be different.
00:04:15
Static persistence, talks about the spatio-temporal continuity
00:04:19
of arrangements of matter.
00:04:20
So you can think about stable
atomic nuclei or stable crystals.
00:04:26
And eventually
a collection of those molecules
00:04:29
may form a network
that self-reinforces.
00:04:32
In science, we call this an autocatalytic network,
00:04:35
one with positive feedback loops.
00:04:37
And that collection of molecules,
taken together is selected
00:04:41
for its dynamic persistence,
which applies to open systems.
00:04:45
Systems that are constantly exchanging
00:04:47
matter, energy, and even information
with their environment.
00:04:51
Human beings, we're always exchanging
molecules as we breathe and as we eat.
00:04:56
And it's not the constituents
of those systems that are persisting
00:04:59
through time, but rather their activities,
their functionality.
00:05:05
There is also selection at times for novelty,
for doing something new.
00:05:10
To be able to see or to fly,
or to swim or to walk on land.
00:05:14
And those new characteristics
allows you to persist and explore
00:05:18
literally new spaces
that you never could explore before.
00:05:23
But a critical distinction
between life and minerals,
00:05:26
the evolution of the atoms, is life
appears to be open ended in its evolution,
00:05:31
constantly exploring new parameter
and possibility spaces.
00:05:37
There is this kind of yin-yang
00:05:39
between the idea of increasing disorder,
increasing entropy,
00:05:43
and at the same time increasing
information, increasing patterning.
00:05:48
At the most basic level,
I think of information
00:05:53
as being described
by something called Kolmogorov complexity,
00:05:57
And that just means
how many bits of information
00:06:00
do you need to completely describe
a system.
00:06:04
Whether that's the words in a book
whether it's the genetic code,
00:06:09
whether it's the information
you need to make a mineral.
00:06:13
No matter how you jumble
the components of that system,
00:06:16
over time, the more Kolmogorov
complexity should remain the same.
00:06:20
But we propose that over time,
a different kind of complexity,
00:06:24
a different kind of information,
which we call 'functional information',
00:06:28
actually increases over time for systems
that are evolving.
00:06:33
The metric 'functional information'
was first introduced
00:06:37
by Nobel Prize winner
Jack Szostak in 2003,
00:06:41
thinking about the functionality
of biological molecules.
00:06:45
This was later applied to additional
systems like artificial life and language.
00:06:51
And then when we were thinking
about what metric
00:06:54
to use for a law of evolution
00:06:58
that can apply to both biological
and non biological systems,
00:07:02
we decided that functional information
might be a useful metric.
00:07:06
The functional information of a system
is characterized by two numbers,
00:07:10
the total number of configurations
that that system can partake in,
00:07:15
and the number of configurations
of that system
00:07:18
that actually perform
the function of interest.
00:07:21
And so you get a fraction
when you divide one number by the other.
00:07:26
So there's a little bit of math involved,
but the math is pretty straightforward.
00:07:29
And you just think of it the rarer
something is
00:07:32
the higher its functional information.
00:07:34
As you're ratcheting-up
00:07:35
more and more function,
the functional information goes up.
00:07:40
You can
00:07:41
see the functional information
increase over
00:07:44
time in numerous natural
and artificial systems.
00:07:49
You can actually calculate it.
For instance, we can watch
00:07:52
as the complexity
in terms of functional information
00:07:56
of Earth's mineralogy
increases over time.
00:08:00
So a mineral is just defined
as a naturally-occurring solid,
00:08:04
that has a well-defined
chemical composition
00:08:07
and a well-defined atomic structure.
00:08:10
The first minerals
would have occurred in very old stars.
00:08:14
And as that star's atmosphere
begins to expand and cool,
00:08:19
you start condensing out
about 25 tiny minerals.
00:08:24
They're stardust,
00:08:25
and those go on then,
to seed the universe with the materials
00:08:30
that form planets.
00:08:32
And planets can then do new things.
00:08:35
And when Earth becomes alive,
that opens up
00:08:38
whole new ways, whole new processes
to make more minerals.
00:08:42
So while the number of minerals
is increasing, the combinatorial
00:08:46
possibilities are increasing
much, much faster.
00:08:50
There are 72 different kinds of
chemical elements that help form minerals.
00:08:54
So how many different combinations
and permutations are there of those atoms?
00:08:59
And it's something like
ten to the 46th power.
00:09:03
An unimaginably large number.
00:09:05
It's finite, but it's a very, very large number.
00:09:08
And of all those different possible
combinations,
00:09:11
only a tiny, tiny fraction
actually form stable minerals.
00:09:16
Only 6000 minerals are actually found.
00:09:19
And it's because nature selects
00:09:23
for those configurations
00:09:26
that persist,
and most configurations fall apart.
00:09:31
So now
we have the negative log to the base
00:09:33
2 of 6000 over 10 to the 46th power.
00:09:36
And that gives us the number of bits
of functional information.
00:09:40
The functional information of minerals
on Earth today is about 142 bits.
00:09:47
But we can also see
the functional information
00:09:49
of artificial life in computer simulations
increase over time.
00:09:54
We can see the functional information
of organic molecules
00:09:58
increase over
time in laboratory experiments.
00:10:01
What our framework does allow us to do
00:10:04
is identify and explore
environments in our solar system
00:10:09
that may have been selected
and evolved in a similar manner.
00:10:13
that give us these three essential
ingredients for an evolving system.
00:10:19
Places that we're looking
include Jupiter's moon Europa and Saturn's
00:10:22
moon Enceladus.
00:10:23
And of course, Saturn's moon Titan,
where there are these interactions
00:10:27
between hydrospheres and geospheres
that can provide the raw ingredients,
00:10:33
the free energy
to sample different configurations
00:10:36
of those ingredients,
and potentially also selection-pressures
00:10:39
that will cause those systems to evolve
in functional information over time.
00:10:47
If there's going to be pushback for
our ideas,
00:10:51
it's always going to be
the point of the second law is enough.
00:10:55
The second law of thermodynamics is all
00:10:58
you need to explain time.
But we tried that game.
00:11:02
Let's take the second law
of thermodynamics and try to explain the
00:11:07
origin of life.
00:11:07
And it just doesn't work
as far as we're concerned.
00:11:11
There has to be something else.
00:11:12
And that's because systems
ratchet themselves up with this increase
00:11:18
in functional information
through selection.
00:11:21
It's consistent with but it's different
from the second law of thermodynamics.
00:11:25
This is very tempting
to think about this idea
00:11:29
in terms of, well, we're selecting for function,
00:11:32
and this ratcheting up
is something that's being driven
00:11:36
in some way with a purpose,
but we don't think of it that way at all.
00:11:40
And it's really important
to make this point, think about gravity.
00:11:43
Without gravity,
we wouldn't have stars and planets.
00:11:46
But the purpose of gravity is not stars
and planets,
00:11:50
the purpose of the law
that we're proposing,
00:11:52
is not to create conscious brains
that can think about the cosmos.
00:11:56
That is just one of the outcomes of matter interacting in this way.
00:12:02
But it's not a complete theory yet.
00:12:05
I mean, we could always be
completely wrong.
00:12:07
Maybe some scientists will see
the tragic flaw that we have.
00:12:10
And and we'll say, here's why it's wrong.
00:12:12
And then it'll go into the dustbin
of history.
00:12:15
And we'll see our laws going to be subject to the same kind of selective pressure.
00:12:19
Our theory itself will evolve,
00:12:21
perhaps according to the law of increasing functional information.
00:12:24
This is just the beginning.
