Subtítulos (436)
0:00- Thank you to AnyDesk
for supporting PBS.
0:02Every intelligent animal
on Earth has a brain.
0:06And bigger brains usually
mean more intelligence,
0:10more neurons, more memory,
0:12more processing power.
0:14So you wouldn't
expect that ants,
0:16each with a brain made
of just 250,000 neurons,
0:21with brains of 86
billion neurons,
0:23in a problem-solving task.
0:25Except that's
exactly what happened
0:28when scientists
gave them both this.
0:30It's known as the
piano mover's problem.
0:33To solve this puzzle,
0:34this oddly shaped load
0:36must be maneuvered
in a specific way
0:38across an obstructed path.
0:39Individual 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.
0:48What happened?
- I had it, I had it
0:50Okay, we're getting somewhere.
0:51- Oh, oh.
- Oh, we did it.
0:54- But give a group of
people this puzzle,
0:56tell 'em they can't talk,
0:57and something weird happens.
0:59The bigger the group gets,
1:00the worse they
get at solving it.
1:03But now give that same
puzzle to a swarm of ants.
1:06Surprisingly, the
more ants you add,
1:10As a human, this is awkward,
1:12because an individual
ant is pretty dumb.
1:15But it turns out
that groups of ants
1:16are capable of solving
surprisingly complex problems,
1:20like building bridges
that adapt to traffic,
1:22or choosing the most efficient
path over an obstacle,
1:27that no individual ant
understands any of this.
1:31This is evidence of a
different form of intelligence,
1:36And it challenges much
of what scientists
1:38thought they knew
about being smart.
1:41Because if intelligence
usually lives inside a brain,
1:46where does an ant colony
1:48keep its intelligence?
1:55Hey, smart people, Joe here.
1:57All over nature, we
see individual animals
1:59come together in tight groups.
2:05And without any leader
2:07telling the swarm what to do,
2:09stunning and complex
group behaviors emerge.
2:12It can sometimes feel like
the swarm itself is alive,
2:16like it possesses
a mind of its own.
2:18And many of the most
impressive examples
2:20are found in social
insects like ants.
2:23Now, when humans
play tug of war,
2:25something strange happens.
2:27Individuals in a group
typically pull less hard
2:30than they would if they
were playing alone.
2:32We tend to slack off a bit.
2:34And the group is less
than the sum of its parts.
2:38the opposite is true.
2:40They use their
impressive strength
2:42to bend and glue leaves
into giant enclosed nests
2:45high up in the tree tops.
2:48the ants divide labor
in a unique way.
2:53while others only pull.
2:55When researchers measured
the forces created,
2:57each ant on average
contributed more pull
3:00than it could if it
were working alone.
3:02Unlike in human tug of war,
3:05the whole is greater than
the sum of its parts.
3:08A phenomenon scientists
call super efficiency.
3:12Now, in the piano
mover's puzzle,
3:13most humans can solve
it after a few tries.
3:16- [Camera Operator]
Oh, we hit a roadblock.
3:19- [Camera Operator]
We did it, okay.
3:20- Pretty unsurprising
3:21for the supposedly most
intelligent animal on Earth,
3:24but when more humans are added
3:26and individuals are
prevented from talking
3:29or seeing facial cues,
3:30we seem to resort to selfish
movements and pulling,
3:34which quickly results
in worse efficiency
3:37compared to people
solving it alone.
3:39What surprised researchers
3:40was that not only
can groups of ants
3:43solve the puzzle at all,
3:45the more ants you add,
3:46the more efficient they become.
3:49In fact, in head-to-head
competitions,
3:51a swarm of just 80 ants
3:53can beat a group of 16 people.
3:56In the wild, there's
likely no organism
3:58that demonstrates this group
super efficiency better
4:02Army ants are wandering hunters
4:04that live in colonies
4:06that can easily reach hundreds
of thousands of individuals.
4:08They're almost
constantly on the move
4:11only stopping periodically
4:13so the queen can lay more eggs,
4:15surrounding her in
a giant 3D structure
4:20But when they are on the hunt,
4:21these ants can travel
hundreds of meters a day.
4:24Covering those
distances is a challenge
4:27when you are the
size of a raisin,
4:29especially in the jungle,
4:30when a fallen leaf
can create an obstacle
4:33dozens of times
bigger than you are.
4:35But army ants have
a clever solution.
4:37They build bridges
with their own bodies
4:40for their comrades
to walk on top of.
4:43Not only do these
extend across gaps,
4:46they also automatically widen
4:48when traffic is heavy
4:49and narrow when traffic is slow.
4:52Human engineers have yet to
build a bridge that can do that.
4:56So how can these ants achieve
4:58such complex engineering
5:00when each individual is so dumb?
5:03- I'm trying to understand
why ant colonies
5:06are smarter than human
societies sometimes.
5:09And the example I like to give
5:11is that there is no traffic jams
5:14despite them having
very, very tiny brains.
5:17- Like Simon, humans
have been intrigued
5:19by social insects for millennia.
5:24Aristotle reasoned
that ants and bees
5:26were political animals.
5:28His term for social creatures,
5:29they create and share
some larger common good.
5:33For the next 2,000 years,
5:35people believed
these social animals
5:38achieved their goals
by following the orders
5:41like a king or queen,
5:43which was surely
influenced by the fact
5:45that most humans lived under
those political systems.
5:51experiments by a Swiss
beekeeper named Francois Huber
5:54revealed that a hive's actions
5:56are not governed by the queen.
5:59She's just one of
many individuals
6:01with a specialized task.
6:03Instead, group behaviors emerge
6:05from individual
members of the colony
6:08interacting with each other
6:09and interacting with
their environment
6:11according to simple rules.
6:13Instead of the old idea
6:15where one wise individual
6:17determines the actions
of the whole group,
6:19group behavior emerges
from individual actions,
6:23despite no individual
even understanding
6:26what the group's goals are.
6:29The group's
problem-solving ability
6:30is spread across
thousands of insects,
6:33each following simple rules
6:35in response to its
local environment.
6:37What scientists now
call swarm intelligence.
6:41- When we say, "Intelligence,"
6:43when we say, "A collective
behavior is intelligent,"
6:44what do we mean by
intelligent here?
6:46We studied it from the point
of view of problem solving.
6:49How does a system
solve the problem?
6:51And the problem for a system
6:52is typically
understood as something
6:54that limits its adaptive value.
6:57because if I don't
get that food,
6:59my adaptive value is
gonna drop very quickly
7:00because I will starve
7:02and not be able to
survive and reproduce.
7:04- According to this definition,
7:05if an individual can
solve a survival problem
7:10they qualify as having
some form of intelligence.
7:13If a group can solve a
problem better than chance
7:16and without a central leader
7:18telling the group what to do,
7:20that group has
swarm intelligence,
7:23flocks of geese
self-assemble into a V shape
7:26that slices through the air,
7:27providing a lift advantage
7:29that's more efficient than
a single goose flying alone,
7:31despite no individual telling
the others what to do.
7:35That's swarm intelligence.
7:37Schools of fish dazzle
and confuse predators
7:39by sensing their
neighbor's movements
7:40and synchronizing their
motion in a large group.
7:43That's swarm intelligence too.
7:45Even the way as humans,
7:46without anyone
telling us what to do,
7:48unconsciously
organize into lanes
7:50when walking in crowded spaces,
7:54of unplanned emergent
swarm intelligence.
7:58Of course, not
all group behavior
8:01is swarm intelligence.
8:02Sometimes animals just hang out,
8:05which isn't particularly
smart or dumb.
8:07And sometimes we even see
cases of swarm stupidity.
8:11Ants, for example, follow
pheromone trails left by scouts
8:14in order to locate food.
8:16But if this system glitches,
8:17it can cause what's often
called a death spiral,
8:21where ants will be trapped
8:22following chemical instincts
8:24until they die of starvation.
8:27Evolution and natural selection
8:28reward species that
can solve problems
8:31which help them survive.
8:33Swarm intelligence does this,
8:35but it wasn't until recently
8:37that scientists figured out
8:38how it actually works.
8:40In the 1980s, computer
programmer Craig Reynolds
8:42argued that we can't claim
8:44to truly understand
a swarm's behavior
8:47until we can
accurately reproduce it
8:49in a computer simulation.
8:50For his first experiment,
8:52he decided to tackle
flocking behavior in birds.
8:54The strategy used
by many species
8:57to deceive and
confuse predators.
8:59It's a behavior so
beautiful and impressive
9:02that it led some bird watchers
9:03to claim it was evidence
that birds were telepathic.
9:08Needless to say,
this is not the case.
9:11Watching birds fly, Craig asked,
9:13"What are the main
rules each bird follows
9:16that end up giving
rise to the flock?
9:18What cues do birds use
to stay in formation
9:22What dials can be tuned,
9:26or by swarm members learning
to change their behaviors?
9:29And what happens when
those dials are turned?"
9:31Craig tested his hypothesis
9:33by making a computer
simulation called BOIDs
9:36or the bird-oid algorithm.
9:39What he found is that
flocks form automatically
9:42when each individual
follows simple local rules
9:46governing just three
different variables.
9:48First, how badly the birds
want to be near other birds.
9:53Second, how far they want to be
9:55from their nearest neighbor
to avoid collisions.
9:57And third, how badly individuals
9:59want to fly in the same
direction as their neighbors.
10:02He later added a
fourth variable,
10:05since the better each
individual can see,
10:07the easier it is for
large flocks to form.
10:10But rather than
publish this work
10:11in a prestigious
biology journal,
10:13Reynolds presented it at a
computer graphics conference
10:16in Anaheim, California.
10:18But somehow biologists
took notice,
10:21and his 1987 presentation
10:23is now hailed as the birth
10:25of a new field of
scientific study,
10:29Today, Simon applies the tools
of time-lapse photography
10:32and BOIDs-like computer modeling
10:34to understand swarm
intelligence in army ants.
10:37As we glimpsed earlier,
10:38when these ants
encounter an obstacle,
10:40they have a special
trick to keep moving.
10:43They shorten their path
by building bridges
10:45with their own bodies.
10:47Here they're patching
a gap between leaves,
10:49and here they're on a track
with an adjustable corner.
10:54it's obvious that
a straight line
10:55is the shortest path
between two points.
10:58But do ant swarms know this too?
11:01When Simon introduced
this zigzag pathway
11:04for the ants to navigate,
11:05he knew that a bridge
from joint A to joint C
11:09but this gap is too
wide for the ants
11:11to construct a bridge.
11:13In response, the
ants did something
11:15surprisingly intelligent.
11:17They started filling in the
elbow with a small bridge
11:21and then they worked
their way down.
11:23This is a brilliant solution
to a rather complex problem.
11:28But how do they do it?
11:30What rules do ants follow
11:31to decide when and where to form
11:33and disassemble a bridge?
11:37Careful observation by
Simon and other scientists
11:39revealed the ants' tricks.
11:41If an ant follows just
four simple rules,
11:44that seems to be enough
11:45to create this complex
group behavior.
11:49slow down when the
terrain gets rough.
11:52Rule number two, if the ant
in front of you slows down,
11:57It's a bit rude, but
it's what they do.
11:59Rule number three, if
you get stepped on,
12:02freeze and brace yourself.
12:04You can even hook legs
with a neighbor if needed.
12:08when you're no longer
being stepped or pulled on,
12:15the fact that army ants
12:16don't crawl outta the bridge
12:18as soon as they're
not being stepped on
12:19turns out to be
surprisingly important.
12:21Without it, ant bridges
usually disassemble too quickly
12:25to really be useful,
12:27along with physical traits,
12:28such as hooked feet,
powerful mouth parts,
12:30plus strong joints
and exoskeletons
12:32that can withstand
heavy foot traffic,
12:35these four rules are essentially
12:37all that army ants need
12:39to start building
living bridges.
12:42It's estimated
that at any moment,
12:4320% of a colony is
locked in bridges.
12:47And since any ant
stuck in a bridge
12:48isn't available for other tasks,
12:50this is a costly habit.
12:52But the advantage it
provides to the colony
12:55and to the egg-laying queen
12:57is clearly worth the sacrifice
12:59in the eyes of evolution.
13:00So how did those ants
13:02solve the piano mover's problem?
13:05These ants seem to communicate
13:06simply by how hard an
individual is pushing
13:09or pulling in one spot.
13:12And when the ants
run into an obstacle,
13:16rather than reversing course
13:18and bouncing around randomly.
13:20That's rule number two.
13:21Simply by being persistent
13:23and responding to force,
13:25this results in a
super efficient method
13:27for exploring many
possible solutions,
13:30and allows groups of ants
13:31to quickly converge
on the correct path
13:34rather than bouncing
around inefficiently.
13:37So where does an ant colony
store its intelligence?
13:42It's not kept by
any one individual.
13:44Instead it's spread through
thousands of individuals,
13:47each following simple rules
13:49in response to local cues.
13:52These rules can be figured out
13:53through careful observation
13:55and tested in computer models
13:56to make sure that
we can replicate
13:58what we think we understand
14:00about the natural world.
14:01And scientists are
applying this knowledge
14:04to solve human problems.
14:05Simon's working
with city planners
14:07to design better traffic systems
14:09that will create
fewer traffic jams,
14:11and designing a self-assembling
robotic conveyor belt system
14:15for use in construction sites
14:17and disaster relief situations.
14:19When you think about it,
14:20the human brain itself
14:22is a collection of
individual neurons,
14:25no single neuron
aware of its purpose
14:30just simply responding
to local signals.
14:33But the result is you,
14:35the greatest intelligence
in the universe,
14:38at least that we know of.
14:42And thank you to AnyDesk
for supporting PBS.
14:44AnyDesk was created
to provide fast,
14:47secure, remote access,
14:49whether you're
working from home,
14:50on the road, or managing systems
across multiple locations.
14:53With AnyDesk, you
can remotely access
14:55computers and servers directly,
14:57enabling secure file transfer
14:58without relying on
consumer cloud storage.
15:01And if you're working on
servers or older systems
15:03with remote server
monitoring and CMS access,
15:06you can control dashboards
and monitor web servers
15:09directly from your
phone or tablet.
15:10Plus it's available
across all major platforms
15:13and has backwards compatibility
with older systems
15:15in case your data's
running on legacy systems.
15:17AnyDesk is available
for personal use.
15:19But if your research requires
a lot of global coordination,
15:22AnyDesk offers custom-tailored
business plans.
15:25To learn more about AnyDesks,
15:26go to anydesk.com/besmart,
15:29or check out the link
in the description.
15:31And as always, thank
you to everyone
15:33who supports the
show on Patreon.
15:35We could not make these videos
15:37without your support.
15:38In a world full of AI slop,
15:42we want you to know
that "Be Smart"
15:44is and always will
be a handmade product
15:47of real human brains.
15:50Pretty intelligent ones too.
15:52If you'd like to help
us keep making this show
15:54and join our super
intelligent hive mind,
15:57well, you can check out the
link down in the description.
15:59All over nature,
we see. (babbling)
16:06I can't talk to you right now.
16:08There we go, all right.
(table smacking)
