Home
Đăng nhập
Đăng ký
Loading...
The Intelligence Test Where Ants Beat Humans - Video học tiếng Anh
Luyện nghe
Nghe
/
Video
/
It's Okay To Be Smart
/
The Intelligence Test Where Ants Beat Humans
The Intelligence Test Where Ants Beat Humans
Chọn chế độ học:
Xem phụ đề
Chọn từ
Viết lại từ
Highlight:
3000 Oxford Words
4000 IELTS Words
5000 Oxford Words
3000 Common Words
1000 TOEIC Words
5000 TOEFL Words
Phụ đề (436)
0:00
- Thank you to AnyDesk for supporting PBS.
0:02
Every intelligent animal on Earth has a brain.
0:06
And bigger brains usually mean more intelligence,
0:10
more neurons, more memory,
0:12
more processing power.
0:14
So you wouldn't expect that ants,
0:16
each with a brain made of just 250,000 neurons,
0:19
could beat humans,
0:21
with brains of 86 billion neurons,
0:23
in a problem-solving task.
0:25
Except that's exactly what happened
0:28
when scientists gave them both this.
0:30
It's known as the piano mover's problem.
0:33
To solve this puzzle,
0:34
this oddly shaped load
0:36
must be maneuvered in a specific way
0:38
across an obstructed path.
0:39
Individual humans can solve this pretty efficiently.
0:42
- Uh-oh, uh-oh. - Uh-oh, uh-oh.
0:44
- Even if it takes a few tries.
0:46
- [Camera Operator] We ran into another roadblock.
0:48
What happened? - I had it, I had it
0:49
for a second.
0:50
Okay, we're getting somewhere.
0:51
- Oh, oh. - Oh, we did it.
0:53
Take that, ants.
0:54
- But give a group of people this puzzle,
0:56
tell 'em they can't talk,
0:57
and something weird happens.
0:59
The bigger the group gets,
1:00
the worse they get at solving it.
1:03
But now give that same puzzle to a swarm of ants.
1:06
Surprisingly, the more ants you add,
1:09
the better they do.
1:10
As a human, this is awkward,
1:12
because an individual ant is pretty dumb.
1:15
But it turns out that groups of ants
1:16
are capable of solving surprisingly complex problems,
1:20
like building bridges that adapt to traffic,
1:22
or choosing the most efficient path over an obstacle,
1:26
despite the fact
1:27
that no individual ant understands any of this.
1:31
This is evidence of a different form of intelligence,
1:34
swarm intelligence.
1:36
And it challenges much of what scientists
1:38
thought they knew about being smart.
1:41
Because if intelligence usually lives inside a brain,
1:46
where does an ant colony
1:48
keep its intelligence?
1:49
(inquisitive music)
1:55
Hey, smart people, Joe here.
1:57
All over nature, we see individual animals
1:59
come together in tight groups.
2:01
Herds, flocks,
2:03
hives or schools.
2:05
And without any leader
2:07
telling the swarm what to do,
2:09
stunning and complex group behaviors emerge.
2:12
It can sometimes feel like the swarm itself is alive,
2:16
like it possesses a mind of its own.
2:18
And many of the most impressive examples
2:20
are found in social insects like ants.
2:23
Now, when humans play tug of war,
2:25
something strange happens.
2:27
Individuals in a group typically pull less hard
2:30
than they would if they were playing alone.
2:32
We tend to slack off a bit.
2:34
And the group is less than the sum of its parts.
2:37
But in weaver ants,
2:38
the opposite is true.
2:40
They use their impressive strength
2:42
to bend and glue leaves into giant enclosed nests
2:45
high up in the tree tops.
2:47
Their secret,
2:48
the ants divide labor in a unique way.
2:51
Some act as anchors,
2:53
while others only pull.
2:55
When researchers measured the forces created,
2:57
each ant on average contributed more pull
3:00
than it could if it were working alone.
3:02
Unlike in human tug of war,
3:05
the whole is greater than the sum of its parts.
3:08
A phenomenon scientists call super efficiency.
3:12
Now, in the piano mover's puzzle,
3:13
most humans can solve it after a few tries.
3:16
- [Camera Operator] Oh, we hit a roadblock.
3:17
- (laughing) Phew.
3:19
- [Camera Operator] We did it, okay.
3:20
- Pretty unsurprising
3:21
for the supposedly most intelligent animal on Earth,
3:24
but when more humans are added
3:26
and individuals are prevented from talking
3:29
or seeing facial cues,
3:30
we seem to resort to selfish movements and pulling,
3:34
which quickly results in worse efficiency
3:37
compared to people solving it alone.
3:39
What surprised researchers
3:40
was that not only can groups of ants
3:43
solve the puzzle at all,
3:45
the more ants you add,
3:46
the more efficient they become.
3:49
In fact, in head-to-head competitions,
3:51
a swarm of just 80 ants
3:53
can beat a group of 16 people.
3:56
In the wild, there's likely no organism
3:58
that demonstrates this group super efficiency better
4:01
than army ants.
4:02
Army ants are wandering hunters
4:04
that live in colonies
4:06
that can easily reach hundreds of thousands of individuals.
4:08
They're almost constantly on the move
4:10
in search of prey,
4:11
only stopping periodically
4:13
so the queen can lay more eggs,
4:15
surrounding her in a giant 3D structure
4:17
made of alive ants.
4:20
But when they are on the hunt,
4:21
these ants can travel hundreds of meters a day.
4:24
Covering those distances is a challenge
4:27
when you are the size of a raisin,
4:29
especially in the jungle,
4:30
when a fallen leaf can create an obstacle
4:33
dozens of times bigger than you are.
4:35
But army ants have a clever solution.
4:37
They build bridges with their own bodies
4:40
for their comrades to walk on top of.
4:43
Not only do these extend across gaps,
4:46
they also automatically widen
4:48
when traffic is heavy
4:49
and narrow when traffic is slow.
4:52
Human engineers have yet to build a bridge that can do that.
4:56
So how can these ants achieve
4:58
such complex engineering
5:00
when each individual is so dumb?
5:03
- I'm trying to understand why ant colonies
5:06
are smarter than human societies sometimes.
5:09
And the example I like to give
5:11
is that there is no traffic jams
5:12
in an ant colony,
5:14
despite them having very, very tiny brains.
5:17
- Like Simon, humans have been intrigued
5:19
by social insects for millennia.
5:22
2,300 years ago,
5:24
Aristotle reasoned that ants and bees
5:26
were political animals.
5:28
His term for social creatures,
5:29
they create and share some larger common good.
5:33
For the next 2,000 years,
5:35
people believed these social animals
5:38
achieved their goals by following the orders
5:40
of a single leader,
5:41
like a king or queen,
5:43
which was surely influenced by the fact
5:45
that most humans lived under those political systems.
5:49
But in the 1700s,
5:51
experiments by a Swiss beekeeper named Francois Huber
5:54
revealed that a hive's actions
5:56
are not governed by the queen.
5:59
She's just one of many individuals
6:01
with a specialized task.
6:03
Instead, group behaviors emerge
6:05
from individual members of the colony
6:08
interacting with each other
6:09
and interacting with their environment
6:11
according to simple rules.
6:13
Instead of the old idea
6:15
where one wise individual
6:17
determines the actions of the whole group,
6:19
group behavior emerges from individual actions,
6:23
despite no individual even understanding
6:26
what the group's goals are.
6:29
The group's problem-solving ability
6:30
is spread across thousands of insects,
6:33
each following simple rules
6:35
in response to its local environment.
6:37
What scientists now call swarm intelligence.
6:41
- When we say, "Intelligence,"
6:43
when we say, "A collective behavior is intelligent,"
6:44
what do we mean by intelligent here?
6:46
We studied it from the point of view of problem solving.
6:49
How does a system solve the problem?
6:51
And the problem for a system
6:52
is typically understood as something
6:54
that limits its adaptive value.
6:56
I need to reach food
6:57
because if I don't get that food,
6:59
my adaptive value is gonna drop very quickly
7:00
because I will starve
7:02
and not be able to survive and reproduce.
7:04
- According to this definition,
7:05
if an individual can solve a survival problem
7:08
better than chance,
7:10
they qualify as having some form of intelligence.
7:13
If a group can solve a problem better than chance
7:16
and without a central leader
7:18
telling the group what to do,
7:20
that group has swarm intelligence,
7:23
flocks of geese self-assemble into a V shape
7:26
that slices through the air,
7:27
providing a lift advantage
7:29
that's more efficient than a single goose flying alone,
7:31
despite no individual telling the others what to do.
7:35
That's swarm intelligence.
7:37
Schools of fish dazzle and confuse predators
7:39
by sensing their neighbor's movements
7:40
and synchronizing their motion in a large group.
7:43
That's swarm intelligence too.
7:45
Even the way as humans,
7:46
without anyone telling us what to do,
7:48
unconsciously organize into lanes
7:50
when walking in crowded spaces,
7:53
that is also a form
7:54
of unplanned emergent swarm intelligence.
7:58
Of course, not all group behavior
8:01
is swarm intelligence.
8:02
Sometimes animals just hang out,
8:05
which isn't particularly smart or dumb.
8:07
And sometimes we even see cases of swarm stupidity.
8:11
Ants, for example, follow pheromone trails left by scouts
8:14
in order to locate food.
8:16
But if this system glitches,
8:17
it can cause what's often called a death spiral,
8:21
where ants will be trapped
8:22
following chemical instincts
8:24
until they die of starvation.
8:27
Evolution and natural selection
8:28
reward species that can solve problems
8:31
which help them survive.
8:33
Swarm intelligence does this,
8:35
but it wasn't until recently
8:37
that scientists figured out
8:38
how it actually works.
8:40
In the 1980s, computer programmer Craig Reynolds
8:42
argued that we can't claim
8:44
to truly understand a swarm's behavior
8:47
until we can accurately reproduce it
8:49
in a computer simulation.
8:50
For his first experiment,
8:52
he decided to tackle flocking behavior in birds.
8:54
The strategy used by many species
8:57
to deceive and confuse predators.
8:59
It's a behavior so beautiful and impressive
9:02
that it led some bird watchers
9:03
to claim it was evidence that birds were telepathic.
9:08
Needless to say, this is not the case.
9:11
Watching birds fly, Craig asked,
9:13
"What are the main rules each bird follows
9:16
that end up giving rise to the flock?
9:18
What cues do birds use to stay in formation
9:21
and move in unison?
9:22
What dials can be tuned,
9:24
either by evolution
9:26
or by swarm members learning to change their behaviors?
9:29
And what happens when those dials are turned?"
9:31
Craig tested his hypothesis
9:33
by making a computer simulation called BOIDs
9:36
or the bird-oid algorithm.
9:39
What he found is that flocks form automatically
9:42
when each individual follows simple local rules
9:46
governing just three different variables.
9:48
First, how badly the birds want to be near other birds.
9:53
Second, how far they want to be
9:55
from their nearest neighbor to avoid collisions.
9:57
And third, how badly individuals
9:59
want to fly in the same direction as their neighbors.
10:02
He later added a fourth variable,
10:04
visual range,
10:05
since the better each individual can see,
10:07
the easier it is for large flocks to form.
10:10
But rather than publish this work
10:11
in a prestigious biology journal,
10:13
Reynolds presented it at a computer graphics conference
10:16
in Anaheim, California.
10:18
But somehow biologists took notice,
10:21
and his 1987 presentation
10:23
is now hailed as the birth
10:25
of a new field of scientific study,
10:27
swarm intelligence.
10:29
Today, Simon applies the tools of time-lapse photography
10:32
and BOIDs-like computer modeling
10:34
to understand swarm intelligence in army ants.
10:37
As we glimpsed earlier,
10:38
when these ants encounter an obstacle,
10:40
they have a special trick to keep moving.
10:43
They shorten their path by building bridges
10:45
with their own bodies.
10:47
Here they're patching a gap between leaves,
10:49
and here they're on a track with an adjustable corner.
10:53
And to most humans,
10:54
it's obvious that a straight line
10:55
is the shortest path between two points.
10:58
But do ant swarms know this too?
11:01
When Simon introduced this zigzag pathway
11:04
for the ants to navigate,
11:05
he knew that a bridge from joint A to joint C
11:08
would be ideal,
11:09
but this gap is too wide for the ants
11:11
to construct a bridge.
11:13
In response, the ants did something
11:15
surprisingly intelligent.
11:17
They started filling in the elbow with a small bridge
11:20
at joint B,
11:21
and then they worked their way down.
11:23
This is a brilliant solution to a rather complex problem.
11:28
But how do they do it?
11:30
What rules do ants follow
11:31
to decide when and where to form
11:33
and disassemble a bridge?
11:35
(inquisitive music)
11:37
Careful observation by Simon and other scientists
11:39
revealed the ants' tricks.
11:41
If an ant follows just four simple rules,
11:44
that seems to be enough
11:45
to create this complex group behavior.
11:48
Rule number one,
11:49
slow down when the terrain gets rough.
11:52
Rule number two, if the ant in front of you slows down,
11:56
walk on them.
11:57
It's a bit rude, but it's what they do.
11:59
Rule number three, if you get stepped on,
12:02
freeze and brace yourself.
12:04
You can even hook legs with a neighbor if needed.
12:06
Rule number four,
12:08
when you're no longer being stepped or pulled on,
12:10
pause for a moment,
12:12
then march onward.
12:14
That short pause,
12:15
the fact that army ants
12:16
don't crawl outta the bridge
12:18
as soon as they're not being stepped on
12:19
turns out to be surprisingly important.
12:21
Without it, ant bridges usually disassemble too quickly
12:25
to really be useful,
12:27
along with physical traits,
12:28
such as hooked feet, powerful mouth parts,
12:30
plus strong joints and exoskeletons
12:32
that can withstand heavy foot traffic,
12:35
these four rules are essentially
12:37
all that army ants need
12:39
to start building living bridges.
12:42
It's estimated that at any moment,
12:43
20% of a colony is locked in bridges.
12:47
And since any ant stuck in a bridge
12:48
isn't available for other tasks,
12:50
this is a costly habit.
12:52
But the advantage it provides to the colony
12:55
and to the egg-laying queen
12:57
is clearly worth the sacrifice
12:59
in the eyes of evolution.
13:00
So how did those ants
13:02
solve the piano mover's problem?
13:05
These ants seem to communicate
13:06
simply by how hard an individual is pushing
13:09
or pulling in one spot.
13:11
That's one rule.
13:12
And when the ants run into an obstacle,
13:14
they slide along it
13:16
rather than reversing course
13:18
and bouncing around randomly.
13:20
That's rule number two.
13:21
Simply by being persistent
13:23
and responding to force,
13:25
this results in a super efficient method
13:27
for exploring many possible solutions,
13:30
and allows groups of ants
13:31
to quickly converge on the correct path
13:34
rather than bouncing around inefficiently.
13:37
So where does an ant colony store its intelligence?
13:41
(upbeat music)
13:42
It's not kept by any one individual.
13:44
Instead it's spread through thousands of individuals,
13:47
each following simple rules
13:49
in response to local cues.
13:52
These rules can be figured out
13:53
through careful observation
13:55
and tested in computer models
13:56
to make sure that we can replicate
13:58
what we think we understand
14:00
about the natural world.
14:01
And scientists are applying this knowledge
14:04
to solve human problems.
14:05
Simon's working with city planners
14:07
to design better traffic systems
14:09
that will create fewer traffic jams,
14:11
and designing a self-assembling robotic conveyor belt system
14:15
for use in construction sites
14:17
and disaster relief situations.
14:19
When you think about it,
14:20
the human brain itself
14:22
is a collection of individual neurons,
14:25
no single neuron aware of its purpose
14:28
or any larger goal,
14:30
just simply responding to local signals.
14:33
But the result is you,
14:35
the greatest intelligence in the universe,
14:38
at least that we know of.
14:40
Stay curious.
14:42
And thank you to AnyDesk for supporting PBS.
14:44
AnyDesk was created to provide fast,
14:47
secure, remote access,
14:49
whether you're working from home,
14:50
on the road, or managing systems across multiple locations.
14:53
With AnyDesk, you can remotely access
14:55
computers and servers directly,
14:57
enabling secure file transfer
14:58
without relying on consumer cloud storage.
15:01
And if you're working on servers or older systems
15:03
with remote server monitoring and CMS access,
15:06
you can control dashboards and monitor web servers
15:09
directly from your phone or tablet.
15:10
Plus it's available across all major platforms
15:13
and has backwards compatibility with older systems
15:15
in case your data's running on legacy systems.
15:17
AnyDesk is available for personal use.
15:19
But if your research requires a lot of global coordination,
15:22
AnyDesk offers custom-tailored business plans.
15:25
To learn more about AnyDesks,
15:26
go to anydesk.com/besmart,
15:29
or check out the link in the description.
15:31
And as always, thank you to everyone
15:33
who supports the show on Patreon.
15:35
We could not make these videos
15:37
without your support.
15:38
In a world full of AI slop,
15:42
we want you to know that "Be Smart"
15:44
is and always will be a handmade product
15:47
of real human brains.
15:50
Pretty intelligent ones too.
15:52
If you'd like to help us keep making this show
15:54
and join our super intelligent hive mind,
15:57
well, you can check out the link down in the description.
15:59
All over nature, we see. (babbling)
16:03
Uh.
16:05
My phone's ringing.
16:06
I can't talk to you right now.
16:08
There we go, all right. (table smacking)
