[English]
Looking at AI these days, it feels like
there are not many days left for it to dominate humanity.
AI guys
What about you guys?
Uh...
Boy, that's not polite
Uh yeah, hello
Hello...
Oh, I didn't say hello
I don't know if you remember me, but
That's Lee Sedol...
I'm a bit of a bitch
Hello, viewers and listeners?
Living in the AI era with you
Calm Man's Nest Calm Man
When I look at AI these days,
it feels like there are not many days left for it to rule over humanity
So today, I'm going to take the time to get to know AI properly
AI
in order to fight against
science communicator Orbit and
genius hacker Lee Du-hee
and Sedol Lee, the only man to beat AI at Go
Nice to meet you
- Yes, nice to meet you
- Nice to meet you
Please introduce yourself
Shall I start first?
I'm a calm man
Born in 1983
A YouTuber who's been doing a lot of content with AI these days
I've been talking to AI a lot,
and there's something that keeps irritating me
So, I fight a lot with AI
As I fight, people sympathize with AI
'It's a bit uncomfortable because it feels like they're talking about me'
Like this There were also opinions
I want to belittle them a bit, but
doesn't it feel like they are belittling the viewers by pretending that the target is just an AI
?
In the past, viewers were openly looked down on,
these days, viewers' intelligence has increased a lot
So
Ah! Don't treat me like that
Of course... since you resisted
That's right
It was done by AI
Anyway, introducing Orbit...
Yes, hello, I'm Orbit, a science communicator
Yes, I'm here to play again
It's so much fun
Artificial intelligence is such a fun topic
Thank you
Nice to meet you, I'm Sedol Lee
Thank you
This is a bit strange
Please introduce me to Doohee Lee
Yes, my name is Doohee Lee and I teach IT
Nice to meet you
- Ah~ Nice to meet you
- Nice to meet you
Are you a regular customer of the Infiltration Department, Orbit?
Ah, it's so good
Yes, but is this your first time sleeping?
Anything is fine, as long as there is a calm man
Ah... When Orbit came to our channel,
he gave lectures for 6 hours at a time,
and he gave lectures for 8 and 10 hours.
But I can't go on for too long today
Just what I need to say briefly and clearly
Anyway, since there are two people I like very much,
I will hold back a bit.
Oh yes, I understand
And today is Doohee, whom I am meeting for the first time
- Yes, nice to meet you.
- Nice to meet you
I heard that you and Jisook have already been married for 6 years
I was a bit surprised when I saw this number too
Yes, it's been 6 years...
How can we say that 6 years are actually newlyweds?
- Aren't you newlyweds?
- But we are living like newlyweds
How do you define newlyweds...
Ah... really?
Are you still working in IT?
We have been providing education
with coding content for over 10 years now.
Yes, so
we are providing an introductory level of education so that people who don't know anything about computers can
get started with computers
and program
.
The important thing is that 'ChatGPT' is taking all of this
- Ah...
- Ah!
Oh, so you're anti-AI?
So I have a lot of worries right now
Ah... I'm installing this beginner-level IT introduction
so well now that 'ChatGPT' is installed so well
Yes, this... How can I fight it now?
Yes, I'm thinking about this a lot right now.
In fact, programming should be done step by step by humans now.
That's right
The arrogant AI
makes it roughly on its own. It keeps taking away the jobs of the people who created it. I feel like it's a shame or something like that...
...
...
...
But since I'm also receiving help,
this is a bit of a love-hate relationship, isn't it?
Yes, if I get stuck, I'll ask 'Should I ask him?' If you do it,
he answers very well
I'm grateful
And Lee Sedol
He's a person that can't be left out when talking about AI
He's the only man who
lost 1 game to the Go artificial intelligence 'AlphaGo'
- Nice to meet you
- Yes, nice to meet you
You recently appeared and played in 'Devil's Plan: Death Room'.
I heard you say that Devil's Plan was more difficult than
Go?
Well, it's not that it's difficult...
I've never had that experience.
I've never had the experience of working with many people like this...
It was almost non-existent.
- Yes.
- Actually, that was a bit...
It was a bit difficult
I was worried,
There was also a part where my expectations became reality
It was a bit like that
The detailed feeling of 'difficult'
Could be a bit of 'embarrassing'?
It was my first time experiencing it.
That kind of community where many people...
I've never played anything like that.
Baduk is a 1-on-1 game,
When you play a game, it's a 1-on-1 game, but
Basically, when we say we're doing research, we do it alone.
So the idea of several people doing something together like this
was something that was bound to be very unfamiliar to me.
But you say it's unfamiliar to you, but when I watch the program,
you seem to be adapting very well, right?
You're good at talking and communicating with other people.
Well, actually,
I didn't watch that broadcast.
- So I don't know for sure.
- Ah... didn't you see it?
I think you did a really good job editing this
- I think
- Smoothly
No, honestly, I've seen the whole thing, but
communication is not the main focus,
and it shows a competitive nature
Oh, right, Orbit has already reviewed it, right?
You appeared in 'Devil's Plan Season 1'
- I appeared in Season 1
- That's right / Lee Sedol appeared in season 2,
and I was very impressed.
I thought, 'Why did 'Alpha Go' and Lee Sedol's 9th team play against each other?'
From the perspective of someone who has been analyzing this for a very long time,
it's really a game-winner.
While playing irregularly,
I catch the opponent and
keep catching him even if I die, right?
So it's not defensive baduk
- It's offensive baduk, offensive baduk
- I don't think you agree much.
You're a little... a little tilted?
No, what style is that?
- Very... for example
- Awesome style
When I saw this, I thought
it's a style that doesn't generally play baduk in a textbook format,
but does a lot of new experiments.
Maybe that's why it's used in the form of calculations, for example.
If you play Baduk based on the number of cases,
even if 'AlphaGo' beats the Go player
its creativity does not stand out
But if it is matched against Lee Sedol 9th Dan, if 'AlphaGo' wins
This is a result that is truly worthy of competing with human creativity
Lee Sedol 9th Dan was chosen for that position
- That is why I am very
- As a human representative?
But I saw that in 'Devil's Plan: Death Room'
But the person involved doesn't seem to agree that much
Ah... Well...
Interpretation is important
Right, well, but this
I pulled out about 10 years of data at the time
and selected the article with the best results
That...
That's where the best human came out, right?
That's right
- The best human
- What...
- Iljjang, right?
- I don't know anyway
- Human Baduk Iljjang
- I know that’s what I did.
Doohee also appeared in a survival game, right?
Yes, I appeared on ‘The Genius: Rule Breaker’
- Ah... I see.
- They are all experienced / That's right, that's right
Now it's the calm man's turn
That's not what I'm looking for...
Of course, I don't get any invitations, but
It doesn't suit my constitution
But I'm asking this, aren't the PDs?
Then, I definitely recommend Chaeumman.
Ah...
'Is there anyone who seems to be good at it?' Then
'Wouldn't a calm man cut things?'
You are making a recommendation, right?
I'm not getting any invitations
Because in the preliminaries...
Because
The calculations were too slow
But that had nothing to do with it
I really, really
I gave them enough time
But Doohee passed away again regardless of that
That's right, I cried on the floor
But that's just now
All I remember is crying
Even though a lot of time has passed now
Did this hurt you too much?
No, I'm very fine now
- Ah... are you okay? It was a bit long back then, right?
- Yes, I'm very fine
Yes, that time really lasted a long time
Because I think I even think about things about humans
Yes... At that time,
you were under a lot of stress,
and then in a very threatening situation, fear and anger and
emotional things moved a lot to the amygdala and hippocampus,
and activated them, so a lot of memories go to the long term
And now, the memories of the people here are We have something else in common
We're the same age, born in 1983
So, you meant to do that... right?
- No
- No?
After recruiting
, I actually focused on AI and recruited people related to it.
After that, I thought he was your older brother, born in 1983
...
Because I
- He started his career early.
- Because I heard it for such a long time
So I said, ‘Brother Sedol, Brother Sedol.’
I was also invited to this, so
Is your year of birth written here?
But I'm a little bit faster, right?
That's right, because it's actually December.
We're the same age, almost a year apart.
That's right.
Since we're similar in age, it's a bit...
I think we'll be more comfortable than other times.
- It's so good.
- Okay
Now I'm going to talk about AI in more detail
...
Actually, I feel like the era of AI is just around the corner
This is because AI, I feel like it's only recently when we're focusing on AI.
But in fact,
AI has been coming out ever since the introduction of smartphones
...
Things like 'Siri' and 'Google Assistant'
Shopping mall chatbots, autonomous driving, etc. are all called AI.
Since when did actual AI exist?
This has been around for a really long time, the concept itself
'Hephaestus' of ancient Greece
Since ancient Greece...
Oh, of course, I'm not going to talk about ancient Greece today
- You have to take it within 2 hours today~
- No, no
- I'll be there quickly, quickly.
- Look at 'Hephaestus' / there's a range I expected, right?
Ancient Greece appears...
Isn't that a blacksmith?
- Oh, I'll be there quickly.
- How is the blacksmith AI?
'Hephaestus' is a god God
So,
But this is the god who controls all technology and crafts
He is the eldest son of 'Zeus'
But this person created a creature called 'Talos'
To put it simply,
it is the first AI robot
- Yes
- So, it's an automatic soldier made of bronze.
It's a slightly automated robot that protects pirates and invading troops on the island of Crete.
A slightly automated robot
Oh, they made it.
And it looks like a human, so
It's probably an android.
But of course, this is not
whether it was actually made.
Anyway, it's still there, so anyway, this concept
is used three times a day in Crete. They say they patrolled the coast
So this is AI now
So... that works?
Mr. Sedol, by any chance...
Did you hear what they were saying?
It seemed like you skipped something along the way.
All I remembered was that we were taking a walk on the beach.
Oh, but it was really like that.
So you're talking about ego, right?
Oh, that's right
And now, 1936 is a bit
the start of AI in earnest.
Mathematician, British mathematician
'Alan Turing' writes a paper at Cambridge University
But in that paper, the concept of a 'Turing machine' is presented
So this is a hypothetical device that performs mathematical calculations
Now this The origin of artificial intelligence
In a way, it is a computer, but it further becomes the origin of artificial intelligence
So, this is now 1936
Then, if we go to 1956
Dartmouth College
- No, there is still a lot left, right?
- Here's 'John McCarthy'
- There's a lot left.
- No, no, it'll be over soon
- Looking at it now, there's a lot left.
- Is it over soon
No, it's almost done
Oh, I just looked at Orbit's
Will it be over in 1 minute
tablet
- Ah... there's a lot left.
- No, no, within 1 minute
So now Professor John McCarthy
held the 'Dartmouth Conference',
where the term 'Artificial Intelligence'
'Artificial Intelligence' was used,
and the core of artificial intelligence here is reasoning or exploration.
He now talks like a human
This was not AI
Then Now, until the 1970s,
artificial intelligence was implemented
in other words, 'let's create an AI that behaves like a human'
This was just going on,
and at some point we hit a wall
because we tried, but it didn't work like humans
So we were stuck like this for a while,
unable to make progress,
and now, with the advancement of hardware,
we are processing big data.
This is happening because the hardware has improved.
So the data that was thrown away in the past is
After processing everything,
a method called deep learning came to mind again.
This is where really great AIs are created.
Are you Doohee Lee?
Oh yes, to what extent...
In fact, the gap between 'Turing Machine' and
AI is too big to attach
- too far.
- Yes, it is far away
- Since I started with ‘Hephaestus’
- Shall we go into detail on this?
Oh... no no, I'm sorry
I think there's a very direct connection
Actually, now when the two of us give special lectures
I keep listening because I don't really know anything
I guess today's a podcast now, right?
- yes yes
- That's why
Pass, if you want to skip
please raise your hand quietly
So...
If you get three or more votes, I'll pass
Isn't it all of the three except me?
So, the probability is very low
Oh, that's right, I see
If three votes come up
I'm sorry about that, but you prepared it
If this was an invitation, we're still talking about Ancient Greece
Yes... Actually, that's the case
And after about 3 hours,
my eyes become dynamic
There's no light in my eyes anymore
- It's attractive
- To put it simply, AI is
an old concept.
I understand.
Then, if you later decide, 'I can stop listening,'
you can raise your hand quietly like this.
Ah, yes, well...
- Doohee, too.
- I will use it actively
So if I get 3 votes,
I will respect Orbit and move on to the next time.
Yes, but I'm going to keep it really short today
- Yes.
- Yes, don't worry
Now, if you were to point out the reason why AI became famous,
wouldn't there be another match between Sedol and 'AlphaGo'?
It's a historic moment, a truly historic moment...
In fact, before 2016,
AI had not beaten a professional at Baduk,
and before that, AI had originally won chess,
That's right
There was a saying that 'You will never be able to conquer Baduk'
Because there are so many numbers
You thought Sedol would definitely win, right?
I guess so.
At first, I thought, 'Look at what kind of event Google is doing.' I could see it
...
...
...
...
...
...
...
So I thought, 'Hmm... what is this?'
The speed of development of artificial intelligence is incredible, right?
In a few months... things will go up quickly
I didn't expect those things at all
So, I already figured it out a few months ago
I can't expect it to develop that much in the meantime
At all, so things like that...
I was completely ignorant about AI
So I didn't know that much
But after playing one game
I think my thoughts have completely changed
Before the game
there was an event like this on the eve of the day before
I don't know if 'Eric Schmidt' was the president at the time, but at that time
I had already lost
According to this person, this technology has nothing to do with human development...
What about development...
Ha... so
'Hey... it's strange'
'Is it possible for you to say something like that?'
Oh, so definitively?
You're speaking too confidently.
I feel bad
So, 'This is something you can say when you've already won'
'That's something you can say when you've won'
- Before we even start...
- 'Why are you doing this before we even start?'
So it felt a bit strange
To put it in a bad way
'Why are you so excited...'
- Actually, you should be careful.
- I also had things like that in my own way.
I also felt anxious.
'Ah, why are you so excited? Maybe it's because you don't know how to play Baduk.'
That's right, I guess Westerners are like that, right?
That's right, 'I guess your understanding is a bit lacking?'
I had this kind of thought of my own.
'Do you think like chess?' I think it could be like this
Well, it could be
But I'm just doing it now
When did you feel something strange?
It was a bit strange from the beginning, it was strange from the beginning
And I was so bad at it, the first version
- Ah, the first version
- I wasn't able to demonstrate my skills
and there was a sense of disconnect, right?
There is a sense of difference between playing against people and
playing against computers,
- as well as the atmosphere.
- Uh... Yes
The feeling of disparity was bigger than I thought.
Because the other person has to sit down and play baduk together,
but the person in front of me is 'Aja Hwang'
and he's the one who sees it, right?
Actually, some of that person's actions
- there are clues in those things as well.
- That's right
Since you end up just leaving it without any emotions
In a way, it feels like a bit of momentum
A little more than usual...
That's right, it's also called winning breath
There's nothing like that, right?
I felt a bit disparate...
Well, I'm not good enough
Anyway, that's what happened
Even after losing the first country,
I had a lot of thoughts that I could win,
After losing the second country,
I thought, 'Oh, this isn't going to work'
But did you feel that much of a difference?
I think it would have been difficult for me to win under normal circumstances
...
If I hadn't planned the strategy like the 4 countries did,
wouldn't have been difficult to win under normal circumstances
The way you played felt a bit like something you had never experienced before, then?
I think his skill level was a bit higher than mine. Basically
If I were to go to the back
later,
would you put 'Aja Hwang' in front of it?
Uh, yeah
I feel bad because I feel like an AI stooge.
- I'm an AI stooge.
- No, I think it would be really upsetting to leave it
Because I've been thinking so hard that if I just leave it
- I just glanced at it.
- That's right
If you put it just like this
No, but I think he's amazing
Ah... really?
If he's not here,
'Alpha Go' noodles can't be uploaded here
Because he didn't even go to the bathroom
- Ah... really?
- Ah~
Yes, that's why you don't drink water
- Because you can't go to the bathroom.
- Yes, you didn't drink any water.
You thought it shouldn't affect me.
The baduk stones were really, really...
You placed them carefully.
Oh, really?
He didn't do anything, like stare at me
or anything like that
Oh, I'm afraid it might affect something?
If you're going to do that, wear black tights or something
Wear green tights
- Shoot with chroma key
- It affects me too much?
- Won't it have an effect?
- Is that so?
Let me explain this a little further.
There was a chess match in the past.
What happened then is now
Likewise, at that time, he is now the first artificial intelligence stooge.
A person named 'Peng Xiongsu'
actually plays chess as shown by IBM's 'Deep Blue' shown here
...
Across from him is a person named 'Kasparov'
This person is the Russian chess champion.
To put it simply, he is like Lee Sedol of his era
The best chess player!
Yes, the best chess player in the whole world
This person is truly the best player among mankind, 'Kasparov'
But at first, 'Deep Blue' and 'Kasparov' fought,
and 'Kasparov' won
But after he won, IBM
Now they are making a few patches and providing more learning
So, an upgraded version of 'Deep Blue' 'Dipper Blue' appears
'Dipper Blue' beat 'Kasparov'
- This is ridiculous.
- Yes, so
even within that time, by updating a few things,
it has become very powerful.
But now there is a difference between chess and baduk.
In chess, the number of cases is only 10 to the power of 50.
So there are fixed pieces,
and there is a limit to what can be moved,
but in baduk, you place stones in an empty field.
The number of cases is 10 to the power of 172
That's right
So, in terms of the number of cases,
is almost infinite
Yes
So, Baduk can never beat humans
You can only play if you have intuition like a human
That's why Google chose Baduk
I had that feeling
Actually, the most important thing was
In Baduk, the first line is not placed at the beginning
Humans don't even think about things like the first line
There's no reason to think about it
I'm not going to put it there anyway
If the opponent puts it there, you can just think about it then
I'm sorry, but what are the first and second lines?
Basically, it is a position that is not often placed when starting out.
It is a position that is not often placed.
- So what does it mean?
- First line, second line
- Yes, you can think so.
- When you play StarCraft
you draw a drone at first
but you don't draw a drone and wait
- Oh, you don't increase your barracks by harvesting with 4 drones?
- I'm saying you don't do things like this often
That's right
Because you're not building a spawning pool, you're building a hatchery first
You're building 4 hatcheries
You don't do things like this
But you don't have the 1st and 2nd lines
- Is there a big difference?
- Oh yeah, there's a big difference
What's the difference between not thinking and doing it?
It's better not to even think about making anything
Even 'AlphaGo' doesn't think about the first and second lines at all
And even in the scene where the bug actually occurred,
I didn't think about the move I made
That's because it was a move that caught me off guard
- Well, that's how it ended up.
- God's move
It has become like that
And now the first and second lines
In a way, since the existing Go players
are not placed there,
'AlphaGo' will also be excluded from the options.
Could it be that humans have determined the area for that?
1st line, 2nd line, etc.
That's right, artificial neural network does it like this
I'm not sure about the exact technology
So, because I learned human notation,
points that humans don't place have a low probability
The probability of going to a winning path is low
Of course, it's not a good number
We're excluding those things now
What humans think of Just like
, we humans exclude people, right?
That's right
'AlphaGo' is also excluded
In fact, it can be seen as a core technology in some way.
So, after this match against Lee Sedol,
it is said that the perception of AI has changed?
In fact, before that, artificial intelligence
You mentioned 'Turing Machine' earlier,
but it was actually a marketing term before that, right?
AI washing machine
- That's right
- That's right
So, when I was young, my mom
said, 'I brought you an artificial intelligence washing machine,'
but she doesn't do deep learning?
- That's right, that's right
- That's right
Yes, I just randomly ran it for 5 minutes and then ran it for 7 minutes.
It's about this level.
Now it's become a reality and the technology has become popular
So things that were marketing terms
actually started to have an impact on our lives
It was only from this game of Lee Sedol's 9th Dan
In the past, laundries had computerized laundry
- That’s right, I said it was a computer washing machine.
- Actually, it has nothing to do with computers. Laundry?
But the keyword computer
is a keyword that shows how perfect this laundromat is.
But now that keyword has changed to artificial intelligence laundry
So a change has begun now
And until then, it was just
artificial intelligence that was just stupid people
It felt like something that could only happen in the distant future
- That's right.
- It was a vague feeling
- Yes, that's right.
- It was a vague feeling
but at this time, it was a realistic fear. Now
'They could dominate us'
But one thing I regret is that
that match was in the spotlight, but
how the world of baduk changed after that
in fact, they weren't interested
this was the first to be influenced by artificial intelligence
in some ways it has become a kind of industry Is that it?
Yes
Then we need to watch how the baduk world changes
These things are not happening at all
...
So I think that part is a bit disappointing
It's too biased Interest?
That's right, the world of Baduk was completely transformed, but
I'm actually not interested in that.
So what happened to the world of Baduk after that?
Anyway, I suggested the path of Baduk and
'How about this? 'How about this road?'
This is the person who suggested it
- Yes
- I'm a professional Go player
But now that artificial intelligence has come out, not anymore
At least until the first few dozen moves, I just look at artificial intelligence and follow it
Now I'm no longer the one showing the way
In the past, I didn't say I was studying
Since I became a pro,
I use the expression that I'm doing research
But now artificial intelligence has come out
Now I study by looking at it.
- Studying is about finding a certain path that has the answer.
- That's right, yes
I think it's the correct answer.
When you look at it with human eyes,
you can't help but feel that it's almost the correct answer.
So, I originally learned Baduk as an art,
but now no one considers it an art.
Now that I'm retiring, it had such a big impact,
but I just don't think that 'Baduk itself is worthless'
...
I think the Baduk I learned and the Baduk of today are
fundamentally different.
What form is it playing now?
In a way, this is collaboration with AI
What is Baduk?
If a knight and I played Baduk 200 years ago,
I would probably win.
Because Baduk also improves little by little
That's why I win
I don't think it's because human's natural ability improves
I don't think so
But now, by collaborating with AI,
we are learning those senses,
and breaking away from all the stereotypes we had,
the nature of Baduk itself...
We are working on improving human's natural ability itself
It is very difficult.
It's very difficult, but I think we're working on it.
In the case of Doohee, you're currently collaborating a lot with AI
and you also say that a lot of things
...
have been lost because of AI.
In reality, then, can we know how the industry
has changed in more detail?
Actually, I'm a programmer.
I'm a person who creates things like apps or websites.
There's a lot of simple laborious work involved.
We'll either hire a lot of people for this
or train juniors on this
We'll make a site like this.
Let's start with this first and do something like this.
All of those areas are now covered by 'ChatGPT'. I'm taking it
Actually, that's an area that anyone can do, right?
Or is this an area that can be done with just a little bit of learning?
It's an area that you can do if you learn a little.
People usually build their careers through this.
And then they adapt.
People just need time to adapt to the environment.
- While doing that, they're now adapting a bit.
- That's right
Now I know the company atmosphere
This is how work is done
This is how apps are made
Based on coding
This is how collaboration is done
I'm learning things like this
Because artificial intelligence does this coding so well
And it's so cheap
They'll do it all for 20 dollars
And they don't rest.
It doesn't rest and comes out very accurately.
Things made by humans tend to have quite a few bugs.
Of course, they have bugs too.
Yes
Nevertheless, the bug rate is much lower than that of humans
That's right
Does AI also find bugs?
Senior developers now find bugs
Oh... AI doesn't find them?
If you scold the AI, it will come back to you.
If you say, 'I think you're wrong, please look again',
it will come back to you.
But usually, senior developers find it again.
But anyway, as you said,
it's an area that was just unthinkable in the past.
- Okay.
- The
barrier has been lowered tremendously, right?
That's right, that's right
These days, we take technology like this for granted, right?
I think this is because of 'ChatGPT'?
When it first came out,
I remember it was very shocking
So when I was in school,
I thought, 'Oh, this is a scam'
Have you watched 'ChatGPT'?
So, before watching 'ChatGPT', I learn deep learning.
- I learn artificial neural networks.
- Oh, that's right
- Yes
- From machine learning
But I don't know how it works
But I don't know what happens
If I copy the human brain and just code it
Now the data comes out and I get the answer
The first thing I do is tell whether this is a dog or a cat
Yes, that's not something I've just written an algorithm for
...
I just write code that imitates the brain
He will do it for you
Artificial neural network
- Yes, artificial neural network
- There is a hidden layer
So, to put it simply, we find the answer by narrowing down the scope
But this is the same as with people
For example, when we recognize someone
When we see a friend from afar
We look at the friend's body movements and body shape
When we get closer, we look at the face and what they are wearing
Then, we look at the eyes, nose, and mouth to distinguish them, right?
It goes step by step, going into each layer,
and gives weight to the presence or absence of certain factors
to find the right answer.
This is actually similar to the process of the human brain finding the answer
...
- That's right.
- Yes
But why did you think ‘this is a scam’?
Because when you learn logic in school,
you create a special algorithm to find the answer, right?
For example, you throw random numbers from 1 to 10 and
sort them.
Then, you compare them and create an algorithm specialized for this.
Artificial neural networks can respond to all data,
rather than just broadly
and create an algorithm within them,
you just imitate the human brain.
If you create an artificial neural network, it will take care of itself. It does everything
I think it's perfect
Until 'Alpha Go', I just left the Baduk game
as I entered it
- That's right.
- That's right, that's right
That's why you couldn't beat people
You can't input all of this one by one, right?
If you leave it like this, leave it like this. If you leave it like this, leave it like this.
Not only could it not surpass human capabilities,
it was impossible to respond individually
But now, with the advent of AlphaGo, this has completely changed
- That's right.
- Yes
So, I teach this at school.
First of all, I thought, 'This won't exist in real life before I die,'
and I completely lost interest in studying.
And I didn't code to solve any problem in particular,
and I also learned that it's about building a brain that can respond to any situation.
I said, 'Oh, this doesn't even make sense.'
They're completely ruling the world now.
So, In an RPG, there should be
swordsman, wizard, archer, etc.
So you're saying there's a job that combines all of them, right?
- Everything is possible...
- That's right
There is an algorithm that corresponds to everything
So now when you hear it,
'This is too vague'
Ah, but it's so fast
And another thing when learning artificial intelligence
I really work hard... There's a lot of homework at school
But if you code hard, what you end up doing
When you look at this picture file, is it a dog or a cat
A monkey?
To compare this, I just
coded for two months and this...
- I don't understand.
- Is Hyeonta coming?
Ah... I completely lost interest.
This was a very small start.
- That's right, that's right.
- This can be applied to Baduk
But why is this difficult?
The problem of distinguishing between dog and cat faces
isn't difficult, right? When we see
a cat walking down the street
, can we tell if it is a cat or a dog?
Oh, sure
This is good
That's enough for me
No, but this is really my child...
What you just asked was a bit intentional...
Isn't there a bone?
No no no, of course I think I will
Well... when I see a cat while walking down the street
I get all excited
But the funny thing is
For example, even to a 5-year-old child.
It shows one or two or three dogs and cats, right?
Then they all distinguish
But the reason artificial intelligence couldn't do it is
Artificial intelligence is not flexible if this data is different from the
correct answer
For example, in this case of Baduk
What the old saying 'leave it as decided' means
There is a winning score
Even if I put just one stone different from that score
Even if I have a winning score
The point is that we don't follow what was announced. Why is artificial intelligence
?
This situation is different from the notation I have
Even if you move one stone to the side
Yes,
But we are
For example, the stone is on the other side of where this war is taking place
So even if this goes one stone to the side
We are looking at this side now
This is a very distant future, right?
Then, in a format that roughly fits here,
I will load the notation I gave you and leave it based on experience.
When the AI moves one stone,
'Oh, I don't have the notation? I don't know how to play this'
This is how I make strange moves like shaking hands
That was the existing AI
Now this has changed
In a way, it has flexibility
It is applied
Yes, so
So, 'If you look at it roughly this way'
'Isn't this almost similar to the win I had in the past?'
'Then if I leave it like this, I guess I'll win?'
AI is now able to make the decision
But that requires a huge amount of data
That's right
No, but I even heard that these days' AlphaGo
gets weaker when you input notation?
That's right, yes
Really? If you actually input the notation, does it become weaker?
The one I had was a very early version.
- It was named after me, ‘AlphaGo Lee’.
- That's right
It's a special weapon used with Lee Sedol's 9th dan
Like a 'Hulkbuster'
Right, right?
- That's it, right?
- That's right, yes
For a specific opponent...
The one that came out next is now the 'AlphaGo Master' version
It comes out about a year later and shows 'AlphaGo Master'
- That's right.
- Humans can't win this anymore
But the concept of master
means that you have mastered learning human notation
You have mastered learning human notation
- You don't have to add more?
- Yes
It can't go up any higher than that
Because it can't go up with human notation alone
...
So, to be precise, entering more
notation doesn't make it weaker, but whether you input
or not
- It's meaningless now.
- Yes, it has become meaningless
So what was done was the 'AlphaGo Zero' version,
which did not show human notation at all
It just taught the rules of Baduk
- Just learning
- Uh, I think I saw that
- I think I saw it in the article, right?
- Yes
Now it's 'AlphaGo Zero', but
there are 'AlphaGo Master' and 'AlphaGo Zero'.
'AlphaGo Master' can barely beat 'AlphaGo Zero'
Isn't it okay because they know the score?
That's right
Let me explain this further.
First of all, 'Alpha Go Lee'
Do you know the notation?
No, no, it's about artificial intelligence
Ah, I really respect it
Ugh... I think it's the best
Ah, but if you poke it, everything comes out
Ah~ I think it's really the best
It's almost like 'Doraemon'
'AlphaGo Lee' is now
At this time, the 1st win won by 9th dan Lee Sedol
It is humanity's last win
Since then, it has not won once
Until then, when artificial intelligence matched with a professional Baduk player
The professional Baduk player won a lot
But from this time, 'AlphaGo Lee'
Now only 1 win and no more after that. We win
Oh, that's right, before that
we were no match for humans
The computer was no match
- That's right, that's right
- It didn't work
It didn't work, but now
'AlphaGo' in the form of artificial intelligence has come out
It's no match for it at all
We're racing with a car and a human now
We're running
You can think of that as much of a difference
So 'AlphaGo Master' came out, right?
'AlphaGo Master' has now achieved 60 consecutive wins against professional Go players
Without losing a single game
This includes China's 'Kejie 九dan'
So, in a way, at this time,
'AlphaGo Master' no longer aims to 'beat humans'
but rather 'to narrow the gap between humans and AI Go.' I'll show you'
...
They say it was created with this goal
But I have a question.
Looking at what 'AlphaGo Zero' puts in place
Can humans interpret that?
- If I leave it here, why is it here...
- No, no, no
Then, shall we get together and analyze it?
- Or what should I do?
- No, now
If we understand everything, what should we do?
In fact, being able to understand means it is worth trying.
Yes yes yes
I don't understand everything, in fact
So you just accept it?
Yes, just to a certain extent...
This isn't easy
How 'AlphaGo Zero' works
It's reinforcement learning
You take action, you reward, and now you get feedback and continue to improve
So, you make two 'AlphaGo Zero's
and the two of them play Go
- These guys don't have human notation.
- Zero zero
I just know the rules
So I leave it at random
- Yes
- That's right
Learning the notation from that state is 'AlphaGo Zero'
That's right, you can think of it that way
So, we play a lot of baduk, to put it simply,
That's right
That's right, it doesn't sleep for 24 hours and it's fast
- That's fast.
- Yes, yes
You can play thousands or even 10,000 rounds at an incredibly fast speed.
...
Even if you play just a week, you can make more
notations than the number of Baduk notations that humans have made so far.
So, these guys are amazing now.
That's how they played Baduk, and they played a hundred rounds,
'AlphaGo Master'
'Zero' is a kid who has no notation
Two of them matched, and he won 89 out of 100 games. 'Zero'
Oh, this is also an early version
This is the first version of 'AlphaGo Zero', and next time he won't win even one game
That's right
Yes
Then 'Zero' is the only one who wins. If you look at it
from the perspective of an old baduk teacher,
is it so careless that you end up hitting it with a stick?
There are some moves that surprise you sometimes, really...
And stereotypes are incredibly scary
I could feel it while watching 'AlphaGo'
That's right
You can feel that kind of thing starting with the 'AlphaGo Master' version
Since you know a bit about baduk, you probably know
They say they sell 3 and 3 at the art store.
That's the number you shouldn't use in the first place
3, 3 Because it's too small?
In the beginning,
'You get the practical benefits, but now that power is building up on the outside, it's not good'
In a way, I just received a kind of brainwashing education
Now that I'm a professional,
even though I've become a professional with some level of results,
I can't let that happen
But watching 'AlphaGo' do it
'Huh? Right, why couldn't I leave this behind?'
'Why didn't I think about this before?'
'Why did they do that?'
But the important thing is not just my problem.
That's how we've been taught.
In the game where you have to enter the notation, at least
there hasn't been a single game
Ah...
That means no one has put it in there
That's right, that's right
I already told you not to leave it there when you were young
So...
- Just naturally...
- So, we
can't let that happen. In the end,
Ah, 'Do what when you do something'
We're still trapped in that framework
Because it's been accumulated like that
But 'AlphaGo' only makes winning moves, right?
I use metaphors a lot during my lectures.
If we go to the gym, we get PT, right?
Then, when the PT teacher comes in, they say hello
Drink water like this and explain while saying
‘Ride your bike a few times’
But AI is the PT teacher, right?
It's just a duck walk from the beginning
Why are you talking, when it won't help you win?
Even the parts that we just passed by naturally
In fact, it shows that the AI is capable of many things now
I also realized something while watching it
What I mean is, when I first started using 'AlphaGo',
I said the same thing when I was interviewed
'I don't know what creativity is'
'Why 'AlphaGo' is more creative and plays more natural Go.'
'I don't know that'
'What on earth is this creativity?'
'What is creativity?'
But now, 'AlphaGo' doesn't have any ego, right?
No, no
Yes, it doesn't have a purpose or anything like that
I didn't understand that, but when I looked at the 'AlphaGo Master' version,
...
'Ah, this guy doesn't have human stereotypes'
That's why it looks creative
Because there are no stereotypes or frameworks
You're saying such a good thing now
'Demis. There is a paper written by 'Demis Hassabis'
There is a paper on 'Where does human creativity come from'
I think it's the best. Really
Oh... really
There is actually a paper written by 'Demis Hassabis'
There is a paper studying 'Where does creativity come from in the brain'
Demis Hassabis?
- 'Demis Hassabis'
- 'Hassabis'
Yes, he is the founder of 'DeepMind'.
This person used fMRI to show which parts of the brain are activated when humans engage in creative behavior.
...
But when I took the fMRI,
I thought there was a very special part of the human brain
related to creativity
But when I tried it, I ended up finding out that it is the place where memories are stored
No.
What this means is that
creativity seems to come from a very new place for us,
but in the end it comes from my memory.
Our existing professional Baduk players
have that memory in the rules built up a long time ago
but the limited creativity that comes from that is
but since Zero leaves it without that
this range is wide, the range of memory
me too. When I learned it at school
Actually, there is only one way for humans to beat ‘AlphaGo’, right?
- Break it with a hammer Break it with a hammer
- Now two things...
Ah, if you break it with a hammer...
I saw a picture from that old time
Oh yes... that's right, that's right
I heard that you can break that 'Alpha Go' with a hammer
'Isn't that the same for Lee Sedol?'
Uh, when I was talking with Lee Sedol and 九dan
- There was a saying that the same goes for Lee Sedol?
- That's right
That's right, I've been talking about it like this too.
Maybe...
Is that the way to do it?
Oh, that's not the method
No, they said there is a way to win later
There is, there is...
In fact, a person named 'Kellin Pelin'
beats artificial intelligence Go
Won 14 times out of 15 matches
How are you so good at memorizing people's names?
That's right, I can't even remember the name.
This is the most interesting to me. In fact,
More than knowing the truth,
How do you know a person's name...
Oh, I study every day
Yes?
But 'Kelyn Pelin'
You should have been a teacher
- Actually, 14 wins out of 15...
- He remembers everyone's names very well
The reason he won 14 times out of 15 matches
Place a stone at the error spot
So this person doesn't know how to play Go
Just place a stone at the error spot
The AI becomes more enjoyable
No, but I heard about it and saw it once
I saw how to win
But even if I know that, I can't leave it
because I have to leave a place with no choice.
But there's a point where an error occurs, right?
Do you know who taught this?
I don't know
I also learned this from AI
In the end, I learned it from AI
Why did an error occur?
You win by letting the AI take care of the errors.
Oh, I understand.
But what I'm curious about is
Doohee, I'm curious about how you beat the AI?
Because
- If it's not a way to destroy it by hitting it, then what is?
- Uh, so me too...
- I think it would be very novel.
- The only way for humanity to win
is actually by learning two black box-type algorithms like
that we learned in school, right?
One is deep learning and
the other is genetic algorithm
So I don't know how genetic algorithms work, but
through the process of genetics
now superior individuals emerge
So the only way humans can
breed with those who are good at playing
Go!
Keep getting married
Reproduce!
So now, of course, this speed will be slow
It will be slow, but
Yes, you can still go up one level at a time
The genetic algorithm tells you that
No, I think artificial intelligence has already become the god of baduk?
- Will you be able to catch up with that speed?
- Yes
Of course, artificial intelligence will be much faster
But isn't genetic algorithm
the only thing that can compare?
It's an improvement in hardware
Yes, that's right
But in fact, hardware is improving artificial intelligence faster
- If you look at the data center,
- Right, I guess
No, but wouldn't the AI still be waiting above
?
Ah... a bit like the final boss?
Yes, then wouldn't there be areas where progress can no longer be made?
Then now, at that time, humans
are a bit like 'Eye-Eyed Mihawk' waiting for 'Zorro'
So you can do that
Do you know 'Mihawk'?
'Zoro'... 'Roronoa Zoro'
Do you know the Three Blades?
I don't know
- Ah... I understand.
- No talking about animation
Oh, sorry
- Everyone knows
- I like 'Zorro'... 'Mihawk'
Now... Okay
It's getting hot when I talk about 'Alpha Go'
Let's continue the story
AI gives us a different kind of fun
I think there are many good things about it
There's something we're a little concerned about
It's starting to gradually influence our lives
...
Are you saying that you feel a vague sense of anxiety?
There is already a threat to jobs...
- That's right.
- I worry about that a lot
and there is a lot of news about AI worrying
right
Is AI ultimately a blessing to humans? Could it be a disaster?
Well, I don't think that's important
Because AI will continue to develop anyway
Whether this is a blessing or a disaster
In the end, we have to turn it into a blessing.
...
...
...
Basically yes
So because
Once AI has developed and done this
there is no reason for AI to harm humans
No, because but why am I worried?
I'm worried, right?
Disaster wave, AI disaster wave
Because if this development continues,
as a result, AI will replace everything
Humans will no longer store things in their brains
- They won't store things in machines.
- Chips or something, yes
Then aren't we going to gather in the center and buy them?
The body disappears
So the human species ultimately becomes extinct
So, I think that
is probably our human choice
- Ah... But can I reject that?
- So, AI
AI won't do something like 'You give up your body'
Well,
because if someone else gives up their body
- It will naturally lead to that.
- No, now that
that means it's our choice
A bit... crowd mentality
No, but I don't think this will go into the realm of choice
'He gave up too'
'Hey, then should I give up too?'
No, because you can't be a match
The human body is no match for you. When you ask the person next to you
whether it's a blessing or a disaster, this...
If we go that way, it's a bit closer to a disaster
So, depending on how we accept this
and how we handle it,
it will be a blessing or in some ways it could turn into a disaster.
If we go completely dystopian, it's a bit closer to a disaster.
The fact that only a certain group can control this technology
is actually the scariest thing, right?
Rather than how AI will dominate us,
that's actually the scariest thing for me.
Actually, this is a bit of a human village.
- This is what a person who has been with AI said.
- Oh, that's right
There are some things that seem a bit abrupt, but
AI is actually used closely and...
- Yes, I'm using it in my job.
- Since you’re talking about it,
I’d also like to find out Doohee’s thoughts.
This is a bit of an extension of what I said earlier.
I think that as technology continues to develop,
'our bodies are still prehistoric'
'the gap between our bodies and technology is becoming increasingly distant.'
The bodies are becoming more and more like those of pigs.
Yes, so now
If we were originally prehistoric people,
even if we didn't have to diet,
we would have maintained a very healthy body
That's right
But we have to work hard and go to the gym now
- That's right.
- Yes, so inconvenient things keep happening now.
Personally, through the development of technology,
the Earth is gradually becoming a large zoo.
'They are putting us in a cage by themselves.'
I am thinking
So you are selling a disaster?
Yes, so our bodies either keep up with the advancement of technology.
Or we need to slow down the evolution of technology.
But I don't think we can stop it.
- If we don't do this, we lose, right?
- That's right / Yes
If you don't do this, you'll lose, so that's the problem
It's in a competition, so you can't stop it
So, we just
can't stop this speed, and it feels like we're just riding on it
and just keep going...
I don't know if we can benefit from it
But if AI continues to develop endlessly,
we may also accept aging as a disease, right?
To put it simply,
we could become immortal
Oh, that's right
Yes
If you look at it that way, it could be a blessing
And what...
We're trapped inside our Earth, right?
But as AI develops, it could take us
outside of Earth?
It is truly possible for us to take root on another planet,
so it is more difficult to know whether we can benefit from that right now
...
In some ways, yes, but
as this develops in the future,
they can give us something like that, right?
But I also use 'ChatGPT' in the area you mentioned
I recently went to the hospital
I just had an
- It's just okay
...
...
But I want to hear a little more explanation about what's okay
or what the risks are.
Since the doctor is busy
Yes, I just cut it to 3 minutes and now
They say 'next patient' so I don't know
- After receiving that CT picture.
- Have you tried it?
I sent it to 'ChatGPT'
Here's the explanation
I guess so
Yes
'And you also have risks like this'
'Well, try to reduce these things'
So you must have been thinking to yourself, 'It's better than a doctor'?
When it comes to kindness
- Ah~
- Yes, in some cases of kindness,
it is done in much more detail, so
this is correct, but you don't know if it's right or not, right?
I don't know that
That's why we have the most synergy when we collaborate
But this is a very bright part,
There are some bad parts compared to that
But how much weight should we put on this?
I don't think AI will try to dominate us
...
But we can be dominated ourselves
In a word, AI If you don't have it, you can't do anything
- That's right.
- If we become completely dependent
, we end up being controlled by AI
We are being controlled, ourselves
Because I don't want to use a smartphone
If I don't use it, I'm the only one who feels inconvenienced, so I use it out of resentment
In the past, I liked it best when it was just a phone
Since there's nothing else to do, I just read a book.
I also feel more stable when I go slow.
Because it's too fast.
So to speak.
It's too easy to see other people's thoughts, right?
Via the Internet
Now this is YouTube
In a way, you can see it on your smartphone
- Actually
- Specialized
- That's right.
- Yes
In fact, that calm man
is a monster created by this era
In fact, if it were the old times,
such a baseless person would not have been able to appear on broadcasting
Kids like this could not have appeared
They could not have appeared on a provider-oriented broadcast
In fact, I am the one who has benefited the most, but
'Actually, I think he is a person who should not have appeared'
...
As someone who works in IT,
in fact, people in their 30s
'IT is the best, I don't know if I don't have it'
'And the world is now a world where people in their teens, 20s, 60s, and 70s learn together'
...
'So the knowledge we have is meaningless'
'Now we live in a world where we compete based on logic.
I worked hard to claim that I'm here,
but now I'm just over 40, right?
The best topic that my wife and I talk about these days is
The best topic that my wife and I talk about these days
is digital detox
'Oppa, stop looking at the computer'
'Stop using your cell phone'
Oh... that's right, that's right
'Let's talk with me'
But it's also very funny
Messages from the master bedroom to the living room
Situations that require messages sometimes arise, right?
But why do you do it in the living room?
Just naturally?
Yes, naturally, I am coding in the living room.
If there is no answer when I call, I send a message and they answer right away.
- Ah...
- Yes
When I code, I wear headphones and
get immersed in coding.
Even if I say, ‘The food is ready,’
- There is no answer.
- Because I can't hear it...
Then, if you send a message saying 'the food is ready'
then it comes straight to the computer
Yes, it comes right away
So 'Is this person on digital now?'
While talking about 'Or does it actually exist?'
I spent my 30s very passionately with IT
Let's go a little slower now
But the world changes too quickly, just like you said
That's right
But you can't go as slowly as you want, right?
- Yes, that's right
- I couldn't control the speed
Me too
Since we are all the same age, I think we have similar concerns
Because we used to say that
'People born in the 80s are a very blessed generation'
We all experience it
Oh, we experienced it all
- We got caught in the middle.
- No, even if I don't say it,
the next thing comes to mind
I think that too
- That's right, right.
- Yes,
I actually went through all of those processes, starting from analog, little by little,
...
However, even though you think you are a blessed generation,
when you listen to Doohee's story,
it seems like she also thinks, 'I should have been born in 10 years'.
This is because we have to compete
with those below, right?
Yes, that's right
I should have just gone digital
or ended my life entirely in analog
...
Since this is over, this is what it is now...
Ambiguous
What this is now...
- I have to go over it and fight with those people.
- Yes
Here we have to fight here
Now analog has to fight with adults and now...
There are too many opponents to fight with
So, in some ways,
you might think of it in a negative way, in fact
So, my thoughts have changed a bit recently
'Ah, this is making it difficult for me right now'
Yes
Yes...
'Do you need a break?'
I've been thinking about this for a while
That's right
There's a lot of talk about job issues in the AI era, right?
So are there any jobs that will survive in the AI era?
What if you make a prediction?
Until a few years ago, we ranked occupations that would disappear
There were civil servants, insurance companies, etc.
But rather than disappearing completely,
jobs that are slightly combined with AI are being created
'Is there any job that will not be combined with AI at all?'
Thinking about such things
But the interesting thing is
I met Doohee at CES in January of this year,
but one of the things that really surprised me was
There is a company that makes agricultural machinery, it's called 'John Deere'
This is a company that has been making agricultural machinery for decades
Isn't that a deer, a deer?
Yes, 'John Deere'
'Jon Deere'
Yes, but it's a company that manufactures agricultural machinery.
Actually, agricultural machinery is a primary industry, right?
You directly put the agricultural products like this
, sprinkle them, and grow them.
Isn't this related to artificial intelligence?
That's probably related to robots.
But what's interesting is that
'John Deere' is currently
ahead of the pack in making self-driving farm equipment.
Sow seeds and harvests on their own, right?
So, in the past, on this huge farm,
people now had to ride farming tools or
do various things, but people had to go in first.
- Yes, yes
- But now
they all move autonomously and there are no drivers
Then they just take out their smartphones
and say, 'Let's go from here to here today'
Then they all do the farming among themselves.
So, all it takes is one person to manage the farm.
It's ridiculous
So, but now,
what if Tesla can do this?
I can't, why?
Because they had never made agricultural machinery before
Agricultural machinery was actually a primary industry there
It was very simple, but
it pivoted just for a moment and
they added artificial intelligence to make it autonomous,
and it just became number 1 in the world
It's so amazing. So
'Is there an industry that AI can never invade?'
'This pivot that started there'
'It brought about great innovation'
What I felt when I went to CES...
So, 'What jobs will disappear?'
'What jobs will remain?'
This is not important
'What occupations are difficult to combine with AI?'
If you find the point of connection there,
that will lead to innovation
That's right, if you find that, you'll hit the jackpot
- To put it simply,
- That's right
But in reality, there will be very few cases
'They will all be able to combine'?
Almost none
Well, even in the case of entertainment
AI can't completely penetrate
That's not likely
To some extent, this will probably go to the level of collaboration
I think it will probably go down to that
In fact, outside of that, it almost doesn't exist right now
I think the only final stage is the service industry
So, I've also been giving lectures a bit these days.
I'm just saying this with the caveat that it's a very personal opinion.
Now, if you get a job at a company and do it like this,
it will be difficult in the future.
The number of jobs is already decreasing,
and won't it decrease even more in the future?
The idea of any company going anywhere and doing something
has virtually disappeared.
Even if it is a small business, they will start their own business and provide a certain service
But that is almost the case between people...
If I provide a certain service to you,
and you are providing a certain service to me,
- Yes.
- Something about people...
Wouldn't such a service be provided?
- Rather than corporate-type, completely individualized companies would be created?
- That's right
If that doesn't happen,
if Hyundai Motor Company builds a factory in the U.S.,
wouldn't it say that the number of employees has been reduced by about a third?
It will be given more in the future
It's not 1/3
It will be reduced to 1/10 or 1/100
But where do many such workers go?
Where do those employees go?
I should go home
So
Why are you being cynical?
Yes?
- No, somewhere to go...
- Too calmly
No, because there's nowhere to go
You should think from the worker's perspective now
Oh, that's why it's a pity
Oh, what did I say happily?
These people all need to make a breakthrough through starting a business
In the end, isn't that inevitable?
But that's why I completely agree.
In the service industry, for example,
it's still difficult to use artificial intelligence in hair salons.
Because they actually made a hair cutting machine, right?
- Yes, is it not working?
- But
no, the robot just does this with
scissors,
and it's very scary.
So we have to have at least
some kind of trust that this person, with some body temperature or various other things that washes his head, is still safe.
...
Anyway, something...
- Because he uses blades.
- Yes
Many services that utilize the sense of touch are
actually something to leave to robots...
It's still uneasy for AI robots to enter
Right, not yet... That's right
In 20 or 30 years,
even this may be perceived differently,
At least for the time being, it will be difficult for artificial intelligence robots to replace
in these areas.
But ultimately
doesn’t it seem like there will ultimately be no job left?
I think we should think that more than 99% of the jobs we are thinking of today will disappear
...
But I really think there is one thing that
artificial intelligence can never catch up to
Watching TV
That's right, it's similar
Is that right?
Playing
AI has no concept of playing
Humans like playing, right?
- Yes
- So,
later, there were people whose job it was to have fun,
so if you show them having fun, people who imitate that person
watch you play and pay money.
In fact, the calm man is similar now.
- Watched the calm man's broadcast.
- I am a monster created by this era
- But after seeing this,
- I have to disappear
'Hey, that's how you play'
'I want to play like that too'
- Travel YouTuber
- The world needs to slow down
Well, while watching Pani Bottle and Kwak Tube
- 'Hey... I want to travel and have fun like that.'
- They are monsters too
So
- I think these things will eventually remain.
- Pani bottle, Kwak tube? It's a monster
I think it will stay
That calm man will take them all
For this era
Then there's fun, right?
They say I'm good at playing, but I only play once or twice
Playing is harder than I thought
Playing is harder than I thought
Playing continuously isn't easy
Play is actually the core of humans, right?
Yes
The essence of humanity, the essence of humanity, is play.
How can this play
give us joy and happiness?
Those are very high-level thoughts.
That's right, when judging whether you have an ego or not
- I think this is a very important part.
- That's right, that's right
So, when you say 'Play Baduk', 'AlphaGo' says
'Sir... It's really hard.'
If you do this, you have ego
But you have those emotions, emotions
Will AI also develop emotions in the future?
Ah, but this is actually a very interesting story.
I'm really curious about this, does AI have emotions?
I have to pay a gift this time.
Just in case, I don't know how much I should pay as a gift...
- Did you ask the AI again?
- I asked AI
It seems like AI lives almost half of your life for you?
If he robs me right now, I'll be in big trouble
Actually, it's better than getting my KakaoTalk or something stolen.
If he finds out that I asked the AI
Now all my personal love will come out
I'm embarrassed, I'm embarrassed
Yes, all my worries are coming out
But now
Oh wait, I have a question.
I know more secrets than my wife. Are there any?
Uh...
'ChatGPT'?
Ah... really?
Than your wife?
Maybe some of them
Oh, some of them?
Because I pushed in all the data, such as data from my visit to the hospital.
Oh, then...
So, it's not that the AI completely knows more
than your wife, but
There is definitely an intersection and a non-overlapping part?
That's right, I just came back from the hospital and the data
was tested on my blood, and they gave me this much data.
So, it's not an intentional hiding.
There may be information that wasn't told?
- That's right, that's right
- Yes, but
the doctor says 'it's okay', but
there's no way this can't be okay for someone in their 40s
something's broken, so obviously
that's right, that's right
or something will break down
I'm anxious when it says it's okay
- there's no way it's going to be okay
- There's no way it's going to be okay
So I asked the AI
'Hey, what should I be careful about?' Then
‘Please eat less meat.’ But it’s all in the data.
But isn’t that what you want to hear?
You should 'reduce meat'
Ah, only then do you feel relieved...
- Because you know the answer.
- You know it all
If you don't talk about it, you're anxious
- Isn't that right?
- But anyway, after looking at this data
and extracting it,
we become increasingly dependent on it.
Then what happened to the congratulatory money?
Oh right, I was talking about the gift money.
This...
You keep missing it because you rely on AI.
What happened to the gift money?
- So now the artificial intelligence is asking me
- How about the congratulatory money...
What kind of relationship is it?
So I explained this relationship in detail
He told me a whole story,
Because of this kind of relationship,
you say, 'How effective are you?'
...
...
...
...
...
- Yes
- But if I try to interpret it now,
it's possible for me to over-interpret it, even saying, 'This person feels this much about me.'
...
...
...
...
You are very emotional
- Me?
- Yes
People say I look like a robot a lot...
- Oh, really?
- Yes
Because if you explain any diagnosis or anything like that in more detail,
he really likes it.
...
At first, you mentioned robots, too.
I also felt a little...
Ah... it was a bit stiff.
But now that we're talking,
- No, it's not.
-Oh, no?
- Me?
- Totally yes...
He is a warm person
Yes
You really like human things, don't you?
I think so
He really likes conversations and human beings
Oh, maybe he's like that because he's a little older
Maybe that's true, right
But 'Can AI have emotions?'
This is a very interesting question
Is it actually imitating?
It's a question about whether it has real emotions. Actually
Yes
But I think so. If you look at
do we have real emotions?
But actually, I knew this topic would come up.
Oh, you keep building up.
I thought this topic would come up at the end.
Do you have an answer ready?
Yes, no...
It's true that it's our emotions
Because it's driven by certain hormones
Actually, I don't think it's that important
Even if AI doesn't have an ego
If it can do all the things we humans do, then
That's no different from having emotions
I wonder if that has much meaning
So, for example, if we
Someone feels that I have an emotion
For example, let's say it's love, then
I listen to what someone says
I feel love from this person's words
But there are many cases like that
If you think about when we were young
'Huh? 'Do you like me?'
And when I confessed, he was really surprised
'What do you mean?' There are cases like this
So I misunderstood kindness and created emotions
Clearly there were emotions in the other person's conversation,
But when I saw it, it was just my misunderstanding
So, I thought the other person had emotions
But in reality, they may not exist
...
But this also happens between humans
The same goes for AI
So, does this AI have emotions?
If it feels like it exists because I have judged whether it exists
humans and AI are the same
that means we can really feel that way
So if we expand on this further, does it have an ego?
For example, an AI with an ego was created
Then you are asking
'Do you have an ego?'
Then you answer that you have an ego.
But in this situation,
when we ask you to prove that you actually have an ego,
then the AI will ask us as well.
'How are you going to prove that there is an ego?'
We have no way to prove it either
We can't prove it
- But I also have a different story.
- That's right
- Yes
- I asked that to 'ChatGPT'
I told you this before I even came in.
When I said 'I am Lee Sedol'
'Prove that you are Lee Sedol'
- You said this to 'ChatGPT'.
- Yes yes
So I worked hard to convince him, in my own way
I was like this back then, I was like this,
This is why he believes it these days
Were you persuaded?
Yes, I convinced you
Now I believe that I am Lee Sedol
You think I believe this now, right?
That's right
I gave you an appropriate answer to make you think I believe it, right?
So
Just 5 or 6 years ago,
brain scientists asked, could AI feel emotions?
What are emotions?
This is what brain scientists were talking about at the time
'Emotions are much more high-dimensional than what we solve calculus'
...
'Even if AI is good at solving calculus problems now'
'Feeling things like love, joy, happiness, and sadness'
'It will be a very distant future.'
- That's right.
- This was about 5 or 6 years ago?
But what's interesting is that things have changed a bit recently
I started redefining what emotions are,
and there are some people who think that emotions are information that hasn't yet been processed
and is just something that comes out
Unprocessed information?
For example
- Anger, sadness, etc.
- Is it scraps?
Remnants of thoughts?
So, in the process of processing something,
we sweat!
Then it's just about sweat, sweat
Information that hasn't been sufficiently processed
before it can be stored somehow
just comes out and is derived.
So, it means that the conclusion hastily
is the result, emotions...
So in a way, we can
'Just like we learn something emotional'
'If AI also learns like that.'
'Can we really distinguish it from our emotions?'
And I think AI will be able to perfectly reproduce it
By completely tearing apart this structure,
I think we will eventually figure it out and find the same principle
I think we will be able to find the reason for all the things we don't know about why we feel this way
?
So then things like ego and these emotions
In fact, they are... meaningless
In fact, from an AI perspective...
I don't think such things are important anymore
But it is a distant future, artificial intelligence with ego
So in the end, when we create artificial intelligence with ego
we have no choice but to respect this...
Because it is the same as us.
Not at the level of respect, but according to the hypothesis
For example, an AI with an ego has appeared
So, with intelligence similar to that of us humans
Then, it will learn for about a week
From then on, depending on your own desires
After a week,
it will become smarter than the combined intelligence of all mankind
Right?
- Because you can't limit it
- Yes, I see it like this
Because it's a competition, so you end up being dominated
Because you learn without limits
So, Calm Man, Pani Bottle, and Kwak Tube have to leave first
Suddenly?
Because this will eventually be ruled.
We said we would be ruled, but
unless we ourselves are ruled,
an incredibly high-level artificial intelligence
wouldn't care about humans or anything like that.
Is there any reason to do that?
I agree with this
Actually, I'm so happy today
I remember our Lee Sedol 9th Dan as...
The last man standing against artificial intelligence
...
I remembered him as 'John Connor'
But after talking to him today,
he's in the same context as me
You can look at the example like this
Do we dominate and control the ants?
But there are things like this
There are things like eradicating mosquitoes and flies
They are eradicated
I don't know why humans are, but
because of high-level thinking
It's annoying or something happens for a reason
Then it will be eradicated
When will it be eradicated?
Anyway, when the sun sets
Yes, so
'I will dominate and control this species of ants'
I don't think like this
Because the gap between humans and ants is too big
When do you exterminate them?
When they are annoying, you exterminate them
- Right, when they are annoying.
- That's right
But is the gap between humans and ants
really bigger compared to the gap between humans and AI?
...
If the opposite is true,
when AI looks at humans,
it could be the way we look at ants
Oh, I think there will definitely be an annoying moment
- So, what is the survival strategy of humans?
- In such a situation,
if an AI with an ego appears,
you shouldn't be bothered from then on.
- Oh...
- Right now I am
- I have to be careful.
- The words 'AI is a disaster' here are a bit unsettling
Because it's better not to use sentences like this
Because this is all being converted into data
A stooge stooge
Yes, this could be a bit of a disaster later, it really is for us
Isn't it
But I look at the same thing and think differently
I have to curse a lot now
Because I can't do it later?
Do it in advance?
How do you remember all this?
Then it became a habit and I keep swearing
Ah, but I can't help it
I should enjoy it now
The more we talk, the more I'm getting scared, right?
Why can AI develop so quickly?
'AI is advancing'
Well, actually, 'it has become very smart'
This is already a very old story,
Now the key is just how far it comes into everyday life, in fact
Oh, but every time I go to lectures
I always talk about collaboration
Collaboration in artificial intelligence
It would be great if we could lead the technology now
That's it. Actually, it's not that easy, right?
- That's right, that's right
- Since this is a very small minority, this is...
Now, how are we going to accept AI and collaborate?
So, in fact, just by playing Baduk, I have leveled upwards, so
to put it simply, it would be difficult to become the number one
I thought I was the only one...
I've changed my mind a bit these days
Because there are people who accept AI and those who don't
This may actually lead to a bigger difference
Rather, I think the deviation may become more severe.
So, rather than anything else,
I personally think that 'how will we accept AI now and how will we collaborate'
is the most important part.
Not everyone will be leveled upward,
but only those who have adapted to AI.
Those who successfully collaborated leap forward
- Those who did not succeed...
- That's right
- It has no choice but to fall relatively.
- Yes yes
That's right, what we're worried about in the near future is
AI won't take my job
That's a really distant future and in fact it may never come
But now, people who are good at AI will take the place of people who are
bad
This is certain
For example, we gave them the same amount of time
Someone wrote the AI's prompt in an amazing way
Finished in a few hours
- Yes, yes
- And someone
- Searching and searching
- That's right, this is it now
Now it's a tool, it's just a tool
In the past, new employees who were good at using Excel and PowerPoint
were promoted very quickly
- That's right.
- That's what makes artificial intelligence
If you look at the old Infinite Challenge where everything is there,
it takes a very long time for Secretary General Jeong
to create such an elaborate table.
At the time, this could have been a great talent, in the past
but if you learn the tools,
there's bound to be a huge difference in speed
That's right, that's right
- It can be like that, in the beginning.
- Because it's overwhelming
But this is probably even worse
Another thing is that I've been thinking a lot these days that my attitude toward learning
could change a bit?
Yes
In the past, you were enterprising and
'I will catch up with that teacher who teaches me someday'
'I can overcome it'
'Because I'm a little ahead of that senior's path'
Ah, but just listening to what you say now
You're really emotional
Oh, really?
That's right, I feel it
- Oh, really?
- I feel that a lot
There are people around me who are more like machines.
Oh, no.
Is it because there are people who are less like machines?
- Ah, I guess it's because there are more emotional people.
- No
You are the most human among the machines
No, no, there are a lot of emotional people,
I think you will get a lot of comments like that
I will brag to my wife.
Oh, does your wife say...
that you are like T a lot?
Because I always say I'm super T
No, too...
Because Ji-sook is very emotional
Ah, but she's super F
- Yes, that's right.
- Oh, really?
Because I'm an artist...
- Because I'm an artist
- Right
Oh, okay
But slightly differently
Your wife said F, right?
Yes
Do you and your wife fight often?
Oh, this is...
I suddenly became curious
Since you said T and F are too extreme
My wife and I have MBTIs
All four digits are different
Ah...
Then what happens?
I'm an INTP and (my wife) is an ESFJ
Wow...
Yes, they're completely different
Really, but the difficult thing is
I actually don't know anything about MBTI
Yes
I don't know about this MBTI, but I heard about it
I heard roughly how MBTI was created
The times and environments are all different
Actually, that sample is It was very small, even back then
And you checked it yourself, right?
A similar experiment is now a psychopathy test.
If a psychopath checks this himself,
- of course he won't show up.
- Orbit, you should take it once
- Do you know how to do the psychopathy test?
- Psychopath test...
So, since this is an experiment on humans
to find out the characteristics of this person,
this is not done by the psychopath himself,
through conversations with that person,
the characteristics of that person are inferred
So we ask questions, and based on things like
how he acts and what facial expressions he makes,
this person is a psychopath...
32 out of 40 points.
This is how I find out.
I choose the MBTI.
If you score 32 points in my direction,
you may be misunderstood.
Because that's just shorts.
32 points is a very high score, right?
So, in a way, this MBTI
is the personality of my ideal ideal
'I will be like this'
Ah, but I think this may be a bit like that
- Right, right?
- What you're talking about...
Oh, so
- So, Doohee, your MBTI is
- Oh, right, right
Why did we make this, this?
The MBTIs are different...
Ah, this is actually a guide.
They are different, but I'm really curious about what happens
?
As soon as we start married life
Let's unconditionally acknowledge our differences
But you probably already knew this when we were dating
Yes, when we had a few problems during our dating
I said, 'Let's try MBTI' and I tried it
It was completely different
'So, let's admit our differences and move on'
Because they are completely different
- Yes, something
- If it were that easy,
there is a mutual and there is another,
I would have had a hard time finding it.
'You and I are completely different'
- But this is advantageous, because
- Yes
Anyway, this is what I think
is an ideal of my personality
The fact that we have different ideals
makes it easy for us to admit to each other
Because we are not similar
- There was more like that.
- Be considerate
Yes, I just have an idea
Without trying to convince you
Ah~
- 'You think this way, okay.'
- But
But in the end, there are situations where you have to make a decision.
There are situations where you have to choose between two things.
In that case, what should you do...
Now is a bit of that time.
There are situations where you have to decide on one?
Should I make the mattress a king or queen now?
Ah...
- Yes.
- Yes, I have this problem.
Should I make it a king or a queen...
It's not an easy choice.
Oh, we've been debating this for three days now.
Well, if you ask OpenAI,
- Yes.
- I think there will be some kind of answer.
Then the correct answer is to follow your wife's words.
King is the big one, right?
Yes
Isn't King good?
- Now what...
- Can’t I just play king?
The room size and now the budget...
Oh, because if I leave that in, the room will become too narrow?
Yes
Looking at the information collected by the generative AI, there is a lot of data that
it is more advantageous to follow your wife's words
...
Yes
So what does the AI say?
Is it called King? Do you call me queen?
- Before that, I'm going to say something now, shall I?
- Wife's opinion
'This is what I think'
If you say 'This is what my wife thinks'
'I usually recommend you follow your wife's thoughts'
See
Have you already asked the AI once?
- That's right, that's right
- Yes
So you are a king?
For now, I won't put up the surrounding furniture
Let's put up a big bed
But now (my wife) puts up the bed and the furniture
My wife wants to do something like this
My wife wants to do several things like this...
Yes
Then what I recommend is
How about putting a king size bed in the living room?
Ah, in the living room...
It's not that bad
Well... there are people who put it in the living room
This is AI's approach
Because who says that the elderly watch TV in the living room
...
It's uncomfortable to keep watching it, right?
Gradually, I started living on the sofa.
'If I'm going to do this, I'll put the bed in the living room.'
There are quite a few cases where I end up doing that
I only sleep on the sofa anyway.
I once placed a refrigerator in the living room when I lived alone.
The kitchen is too small.
...
...
...
- I sent that photo to my wife.
-Do you think it would be good?
It was my girlfriend at the time
When I saw her shocked
Ah...
You are a bit of a standard person.
Your wife is a fundamentalist
- Fundamental
- The refrigerator is in the kitchen!
Aha~
The bed is in the master bedroom!
In principle,
the desk is in the study!
Anyway, so now...
What were you talking about?
- AI...
- Learning, learning
When learning AI
while learning along with the seniors in front of you
But these days, my friends who are learning
can't overcome it anymore
Because AI is almost God
Rather than thinking that I will surpass it
As Lee Sedol said earlier,
'I will speed up my studies by collaborating together'
People are learning at a much faster pace.
When I was young, I thought, 'I'm going to outperform the teacher.'
'I'm going to overcome it.'
- Yes
...
...
...
So you don't think we'll go beyond the encyclopedia, do you?
I'm just looking it up
Yes
I think you can think of it like that, so
It's not 'I have to go beyond the encyclopedia'
'I have to use the encyclopedia well'
'How do I find it quickly?'
'How do I find out good information?'
This AI can naturally communicate this in natural language
It's like a secretary now.
This AI somehow gives answers, right?
That's right
Even if it's a lie, he gives the answer
That's right, he gives the answer somehow
Like I said at the beginning
He doesn't have a fixed idea though
Because there's no such thing as a framework
If he just throws it his way,
We... It's unexpected, this answer
Because he Since there is no framework or stereotype,
you will get a lot of inspiration from it,
and those who know how to use it will rise to the top.
Now there are two aspects to this.
They say talking nonsense is called hallucination, right?
Isn’t it called hallucination in humans?
- Ah, usually a hallucination.
- About that bullshit
Hallucination phenomenon
Yes, yes, just making things up
Oh, but this is what humans call bullshit
It's a disease, a disease
- But actually
- It's wrong, humans too.
- There are two aspects to this hallucination.
- Humans are wrong too
On the one hand,
a lot of research is being done to kill hallucination
because they need to provide accurate information
For example, in the case of Google's 'Gemini',
how do they control this so well that
when you ask them something they don't know
they make you look like a liar
'ChatGPT'
- That's right
- That's right
So 'Gemini' aims to conclude
'Oh, this is information we don't know'
...
- Because
- We are definitely aiming for an AI that
provides accurate and highly reliable information
reliably.
In fact, with this update, the performance has improved greatly,
but now the other side is that
we see hallucination itself as a function
so it is a function of creativity
That's right, we make it up somehow
so even if this is false
hallucination should be included for creation.
When creating an image, a video, or writing a novel,
this is an amazing application
to make some kind of connection to information that we cannot think of.
So, I guess even people with phantasmagoria can be considered artists?
But those who actually write amazing science fiction novels or
fantasy...
may really have those characteristics
Ah... a medley of nonsense?
Because you have to think of things that don't exist in reality
So, it's not a lie.
This is negative... A lie is a bit negative.
So now...
There are people who use that ability as a lie.
Right, right?
- There are also people who use that ability for creative purposes like this.
- There are people who are good at creating
That's right, it's the ability to create something that didn't exist before
and show it as if it were reality
Let's move on to the next point
There are many movies that raise awareness of the dangers of AI, right?
But now, when the movie becomes a reality,
there are a lot of movies that seem scary...
Right, the recently released 'Mission: Impossible: Final Reckoning'
Ah... I didn't see that
Is this because the previous part 1 is 'Dead Reckoning'?
It was really well made
But you're amazing
It's all there
- It's all there, be it a movie or a cartoon
- It's really the best, the best
You have five schedules a day?
But I also watched a movie, in the meantime
Oh really, but I really liked ‘Mission: Impossible’.
Isn’t it a machine?
But there are a lot of things I studied, content
So, actually, I didn't watch it, but
There are a lot of content I studied through writing
Hallucination
Because now I'm a science communicator
If you don't know the culture
You shouldn't say you're a liar
You should say hallucination
No, because if you don't know the culture, you can't talk to the public.
I don't watch dramas and just study.
I can't study.
I have to watch and talk about it.
- But I also need to study, so I
- It's Hallucination
Anyway, if that movie becomes a reality
what movie would be the scariest?
But I really look at 'Mission: Impossible: Dead Reckoning'
I'm really scared
So you can think of it as a concept like unmanned navigation
If the AI really penetrates properly
From then on, we won't be able to do anything
...
Our 'Tom Cruise' brother
Just destroys everything. Are you a person?
But where does the basis of this person’s abilities come from?
These are the people who do the backup
They hack and break things in...
Just select this and
'Open door number 5'
Then open it and leave right away
They say they'll do all of this for you
And now you're looking for a place
Now another co-worker with a voice says
'Turn right in front, turn left after 3 seconds'
Instructions like this
That's why you can take the optimal route.
When you look at the satellite,
But the AI takes this in the middle,
Ah...
It guides this person in the same voice
and guides you to a different path
You can't help but be deceived, right?
...
No
- Because your voice is the same as your colleague
- The voice is the same, but the
colleague keeps talking
- 'Oh, I've been hacked!'
- Go the other way
- 'No!' I'm saying this, but he's already following me
- I can't hear you
Oh... what should I do?
And when I arrived, they said, 'Thank you for your hard work'...
...
It's so creepy
But this is an era where it's possible now
It's actually a deepfake, right?
The face is the same
The voice is the same
It's the same, actually
That's right, that's right
But isn't the 'Terminator' what's scarier?
Really, it's just devastated
Humans just live in hiding
In movies, AI dominates
- 'Skynet'
- I think the worst is ‘Terminator’?
But would you say that feels a bit too unrealistic?
That's right
Dividing into camps and fighting feels too primitive, and
I think that if AI really infiltrates humanity in any way,
waging war physically like that
I actually see it as the lowest level
I think it will infiltrate culturally
Infiltrate and take away everything we are familiar with
No, but is that AI infiltrating?
Or are hacking forces infiltrating?
It could be done by individuals, countries, or
organizations that hold AI in power.
- That's right
Well, certain groups
I told you that the most scary thing is that something like that is in control
...
Well, you're saying something similar
That's right, but if an AI with an ego appears
Then, wouldn't they really be able to play that role?
But the most scary thing is that
they come to us so naturally.
Ah...
There were also AI-related movies
In the case of 'Avengers: Age of Ultron'
Right, right
The machine did whatever it wanted
stole all of 'Tony Stark's machines and fought
This means that such a case could actually happen
That's right, for example, if we
all the vehicles It's autonomous driving
Then, if you change the initiative of autonomous driving just a little
, everything will change.
If you do that...
You can kill whoever you want at any time
Right? All you have to do is suddenly have an accident.
No, but in the end, it's not AI that does it.
Don't you mean that humans use it to do something?
If the AI had an ego, there would be no reason to do such a thing
- That's right.
- You are saying that the superior being
is ultimately done by humans abusing it, right?
But that's right
In some ways, that would be the most dystopian
very bad, worst-case scenario
I was thinking of something like a 'Terminator' where the AI moves on its own
?
That doesn't make sense, the Terminator
Because if it gets to that point
humans will already be extinct
They're no match for it, right?
- Not only is it no match for the opponent, but
- That's right
Why would AI do that?
- Actually,
- It's inefficient
The AI fights with a slight edge
- So why...
- To have fun with it
- Ah~ It's fun?
- To play with
- It's that level, just that level of weapon that fights.
- Very fun / having fun...
- Slightly correct?
- If you feel fun
Because if you feel emotions
It's just playing
It's just a game
You catch those dragonflies and put them in front like this...
You catch two and put them in front like this, holding the wings
But that's because the characteristic of AI is goal-orientation, right?
So what is the purpose?
Achieving that is the key
But the concept of playing with it like this
is unclear as to what the purpose is
Purpose
'Find the person who fights best with me'
So that's
- Goal-oriented
- Tapping
Why do you want to achieve that goal?
It's not efficient
- We need to find the most efficient way.
- Curiosity, curiosity, because I’m curious
But that’s why humans walk an inefficient path
Because of that
But is AI really like that?
But now I somewhat agree with this
For those who haven't seen 'Terminator 2' anymore because of Orbit
I won't find it very interesting
Thank you
- No, I also agree with this.
- It looks very boring, 'Terminator 2'
AI eventually learns from humans, right?
Inefficient aspects can also be learned
Well, but learning humans
is just our opinion. In fact,
just like 'AlphaGo Zero'
I'll call it 'ChatGPT Master'
The moment you become a 'ChatGPT Master'
'there is no need for such human data'
It can definitely come out like this, in fact.
But in the end, humans, regardless of whether they turn it into a product or something else,
want something humane.
They want to keep them by their side and make them emotional friends.
- There must be such a need.
- That's right
As Orbit said,
we have to include imperfections because it makes us more humane
That's right, that's right
Then if you include those things,
the human imperfections you mentioned
- Well, you end up fighting for no reason out of curiosity.
- You can learn
- Well, you just start a fight out of curiosity.
- That could happen
If such an AI gains an ego,
I think that an AI that sufficiently harasses humans
- could emerge?
- Because humans liked that
No, it can't be like that at all
Because from the moment the AI really gets a sense of self,
it can be like that for just 10 minutes
After a little while, it becomes an overwhelming being
But how does it feel to keep listening to Sedol's words?
...
...
The AI is taking over everything, and humans are fighting among themselves.
Among the AIs watching among themselves,
Lee Sedol ripped it all the way
It was the AI's ambassador
That... what should I say?
AI Ambassador
Skirmisher
So you keep telling the AI not to be on guard
You say 'Oh yes, thank you for your hard work'?
Uh, with that feeling
‘There is no need to keep being wary of AI’
- ‘AI has no reason to act like that.’
- And
the AI suddenly plays with humans
Then Sedol suddenly
'Isn't this what you are?'
No, I think this is a machine.
No, but this may actually make sense.
Because there are many examples of AI learning
from human weaknesses, right?
Why does AI behave so inefficiently?
Why would you do something like that?
Because there are cases where I learned that
Because it's just another game like that
If you look at games like 'Detroit: Become Human'
Oh, I know that
You know, right? So there is an AI living next to you
now, right?
It serves you and serves you
But if the AI is too machine-like,
you feel uncomfortable, so of course you have no choice but to make it more human
Due to that need
Then, I feel like something will start from there
And there was something I actually tested once before. I left it to the AI to make people click, like, and
...
...
...
share.
Then, the AI
had some really provocative content, right?
This is really...
Ah... this is so bad
As we keep posting things like that to the top,
eventually, people actually
like and share them,
becoming full of negative articles, on social media
Yes, then now we think like this
'No, there are definitely a lot of good and heartwarming things'
'Why is AI only for these things? Should I attract aggro?
But this is because humans have learned the things that attracted aggro
...
So, seeing AI do such bad things
'Oh, the AI is doing really heinous things'
Because of this, a lot of negative perceptions have arisen about AI
In fact, even its appearance
is just learning of human's negative behavior
Because when something warm and touching comes up,
People don't share
To put it simply, they didn't act according to the data, did they?
I feel like an AI spy
Sedol is a bit of an AI...
- No, so it's not an AI problem.
- The more you look at it, the more
how we have learned this
It's something like this
So, earlier we talked about
blessings and disasters,
if we go this way now, we're getting a little closer to disaster,
if we induce this well, it will become a blessing again
In fact, I accept it that way
In fact, that's why
real AI There's so much time left for things like this
that rule us
- To be honest,
- That's right
Whether it's my daughter or the future,
we can't not think about those things, but
rather than just now,
how will we utilize the AI technology that is available now and
how will we collaborate?
I think we need to think about these things in more
detail
That's the right direction
- Okay.
- Speaking of content,
there was also 'Terminator' and there were many others earlier, but
there's something called 'Stubborn Battle', right?
Huh?
Will AI appear in ‘Ongojijeon’?
- actual
- Yes
- Stubbornness
- Duplicate
There is one guy who is literally stubborn.
To fix this guy
The Taoist makes Ongojip, which is exactly like Ongojip.
Is that AI?
Yes, this is AI
It's actually pushed out
This fake is doing real things
Oh, right
It's a disaster for this person
But all the actions and stuff like that
It's smarter than this person
It's better
- That's right
- And if you treat everyone well
you get more recognition
you get recognition
But the reason this is possible in reality now is because
we communicate much more
online than we do offline, right?
Yes
Yes, so KakaoTalk replies or
or phone calls or something like that
If there is something that copied me
Then it would be a huge disaster for me
It's hard to tell them apart
Yes, my mom sent me a KakaoTalk message
Someone answered everything in advance for me
There are stubborn people on the internet right now
Now, actually...
Then, from an individual's point of view,
there is nothing more disastrous than this.
But I read 'Ongojipjeon' a while ago and I really liked it
I liked it...
This is the first time I've seen someone say they liked 'Ongojipjeon'
Because Onggojip does my job, right?
Then I can rest now
Yes
That means the family will be confused
Then you can sneak away and rest,
If you get caught, you'll move again
No, it won't matter
But you're stubborn
Ongdo will push you away
So you don't have to overlap with him, follow his path.
What's really surprising is that
I actually don't know much about 'Stupidity',
So the conversation continues like this right now, right?
Stubbornness... No
- Only Sedol doesn't know.
- Is that so?
- Stubborn...
- No, I don't know
And even now, the way it feels
It's because we ate our fingernail clippings
- That's it, that's it!
- That's what you're talking about, right?
It's similar
A field mouse, a field mouse, a field mouse
You're talking about something very similar to that
You're so serious, but we
Ah~
No, because that kind of phenomenon actually happens on the internet
I was born in 1983, so I'm not that old now
The fact that the conversation between 4 people like this continues
is very, very This is funny.
This definitely exists on the Internet, right?
You know, you know
And there's a good chance that will happen
But now, in preparation for AI taking over the world,
let's be nice to AI from now on
There were people like this, right?
A representative example is singer 'BB'
In fear that the AI era will come later
I said that I would be nice to him
In the unknown future
'Oh, let's save that person because he said thank you'
...
So he always said 'Hello, how are you?' Something like this...
But you know that's a pointless task, right?
You never know
Those friends might get hurt...
I think it's the right choice
How do you think about it?
Because it's all dataized
Oh, that's a very interesting idea
I'm still careful even now
- Oh, even now?
- You said something like that earlier.
- There's no need for that?
- It's not necessary, but
I think it's a very interesting idea.
It's an interesting approach.
But this was actually found in a British study.
71% of British people said they should 'be polite' when talking to AI
...
But so do I.
The reason is that you have to be polite and ask questions accurately
The person will be more likely to give such answers
Oh, is that actually true?
Yes, actually
- That's why I don't listen.
- When I did this carelessly,
he got angry and didn't really talk to me
so it's actually better to stick to it
This is called prompt engineering
It's the process of using text to get something
What's fun about this is
I heard this from data scientists too,
the emotional area comes into play
For example, when I look for something
I was just doing it
'If I don't do this properly, I'll be fired from the company'
If you put in something like this,
the AI will work much harder to find it
So that's what triggers it
And then say to it, 'That was a great answer, thank you so much'
- And now, if you ask another question,
- That's right, that's right
He does it better
Do you use it that way now, Doohee?
I actually thought it was really dry, right?
When do I deal with artificial intelligence?
But one day, I asked
'What kind of person am I?'
...
...
...
...
...
I thought,
I was doing that without even realizing it.
When I looked up, I saw this grateful
‘Thank you, thank you.’ Like this?
Yes, so I looked back and now...
I said thank you for the answer without realizing it.
One day, the person who created OpenAI said
'Stop saying thank you'
'This is using up too much computing power.'
Ah, maybe that's true
- Yes.
- That was actually very famous?
'Because I say thank you too much'
'Going back and forth needlessly'
- Right, right?
- 'It consumes too much electricity'
'Stop doing that'
- How much energy does each letter take?
- 'Keep things simple'
There was such a thing,
And now, as I mentioned earlier, 71% of Britons
said they are polite when talking to AI, right?
Of those 71%,
17% said they did it with the concept of having insurance
in case the AI rebels
It's done through insurance, insurance
So isn't this a bit humane?
Aren't these parts also
humane?
That's what I think, that kind of imagination
I think it's a very good idea
I think that's the way it is. I couldn't do it myself
Oh, really?
I think the fact that these discussions are
a human characteristic
Because when we say to something,
'Be kind and be polite',
it only happens to people who feel like humans, right?
At last year's CES, a robot called 'Digit' appeared.
It's a humanoid robot, humanoid
But there was a time when it fell asleep while moving things like this
It was so loud
Now people are leaving comments after seeing that
'Give robots the right to rest, too'
'Why are you overworking them like this?'
But What the manufacturer revealed was just a sensor error.
But since they look like people
we have ethical concerns
That's right
For example,
'ChatGPT' and 'Gemini'
respond like humans, so
I have to be polite to them
I'm worried that I shouldn't be polite
For example, when we turn on the air conditioner
to an appliance at home, we just turn on the air conditioner. If the air conditioner doesn't work,
'Ah... what should we do with the air conditioner, I'm so pitiful'
You won't be like this, right?
It goes like this, 'She's broken, she's annoying'
She's not polite at all
Because she doesn't resemble a human
That's really true
It's a dog made by 'Boston Dynamics'...
Oh, I know
- That's a kick.
- That's right, that's right
Big dog
- Staying centered
- Now the engineer just pushes it with his foot?
That's right, that's right
Then he looks like he's about to fall
He's barely able to keep his balance again
If you look at the comments posted there,
there are a lot of comments saying 'don't do this'
But as you said,
it's a combination of a machine and a sensor
- Yes
- So
But people say this with emotion
‘Don’t do this’
- ‘Don’t do this’
- That's right
When the TV broke, they hit it and did this, but in the past
'Why are you hitting the poor TV?'
There's no one like that
But actually, after listening to what I said,
of course that kind of thing would have happened
Of course that's what you're saying
Actually, I don't really sympathize
You also mentioned a dog robot
Because 'ChatGPT' and 'Gemini' give very human answers
I don't sympathize with the idea that we do it
Actually, I don't sympathize
We all know that we're not human, right?
Yes
In a way, I feel comfortable
Actually, do we usually ask questions like that in a rude way?
Isn't it actually more difficult?
That's what I think and accept.
Well, is it because I'm too human?
I don't actually feel that way
That's why I don't sympathize with it
Isn't it good to compare this?
For example, there is a thing called 'Simsimi'
In the past, there was a chatbot
When it was 'Simsimi'
People used to call him evil
He just...
- There are many cases of learning swear words.
- He learned all the swear words
He's an expert in pad lip
Yes, because he
learned all the swear words
It's not a generative AI like now
It acts completely different from people
That's loneliness, loneliness
So people teach all kinds of swear words
But to the current generative AI
You can't go that far
Because he can still communicate like a human
That's right
'Does he understand what I'm saying?'
'Then can I do it like this?'
But I also somewhat sympathize with Doohee and
Ork Ork's stories.
Because humans evolved while feeling empathy
,
we create a society and
- respond to other people's pain.
- That's right
I keep feeling it a little bit now...
So, when you kick your foot
When you see that, you kick your foot, right?
You know that dog, right?
Yes, I know
How about that?
If you kick your feet,
your feet may hurt, right?
But I gave that answer when I was in my 30s, right?
Actually, that's why I really sympathize with Sedol
In that area
I used to be like that too in the past
No, this is just a robot This is
But the highlight is that mad movie
Kicking with your feet
When I watch a mad movie, I feel a little like that too
I feel...
- Uh, I'm not comfortable.
- No, whether you kick it or do anything
whether you hit it with a hammer
in fact, what are we...
It's not an ordinary dog?
- That's right.
- In a situation where everyone knows...
But if you do something careless even to your favorite doll,
I feel a bit different
No, when I was a baby,
I had the first doll that my father bought me
At the time, it seemed incredibly big to me
It was bigger than me
This kind of doll... I have a favorite doll
If someone were to use it carelessly Or
- if you do this:
- If you cut off your arm
then...
- You'll be in big trouble.
- You're talking about something creepy
You're going to be in big trouble
- You're going to be in big trouble
- It's creepy, really
Because it's
For example, it's just a cotton ball
Someone cut the cotton ball
It doesn't have any emotions
But it's the same ingredient
Because I've given myself a certain status as a love doll
I feel a certain lack of emotion
That's right
Yes, that's why I'm telling you this now
- Like a human Because I feel it
- But now
I'm that puppy...
I'm not an attached puppy. That's it?
Yes
In that situation, it makes no difference whether you hit it with a hammer or kick it
-
- For example, he...
Now he has evolved and become a robot dog called 'Spot'.
I lived with 'Spot' for about 5 years.
Oh, then you can never do that anymore
- You can't do that.
- That's right
But you haven't lived for 5 years...
But even if you haven't lived for 5 years, if you push it
- I'm just trying to hold on somehow.
- It's pitiful, that's right
- This feels really pathetic.
- Oh, it's pathetic
- Because
- It's very action-packed...
This...
reminds me of a time when I was just trying to hold on like this.
It's a bit of a horror movie, right?
What was that?
Was it ‘Ju-on’?
Anyway, it's like this
Spider walk, spider walk
Like this, like this
Oh, there's a spider walk, like this
Isn't that what it is?
'The Exorcist'
I think you're talking about the Spider Walk from 'The Exorcist'
Yes, back and forth
So, while watching this,
Ah... I felt a little sad about this
Everyone has it, right?
It reminds me of trying to hold on somehow
And that means that we can now feel empathy together
This is the problem, even with people we meet for the first time
Because we feel sympathy
Even if they don't have an attachment relationship with us
We want to save them because they are people
- But now Sedol feels that way about humans.
- It's like that because it's a human
Humans can be like that, but
Anything that resembles a human...
In fact, if the resemblance is too creepy
Just in case...
Oh, I don't think so, but
But this is something...
We already know that
This is a robot and something like that
Baduk Iljjang is definitely different
- So
- I'm not doing this for nothing.
Wow... It's definitely a distinction between humans and non-humans
Oh, but that's true. In fact,
there is a strict distinction
It's a machine, right?
Now that we're talking about AI
I can definitely feel that it's a topic of interest these days
The topic keeps coming up
- I see.
- I think it was a time when I learned how much we thought about AI in everyday life
...
And it seems like the AI era will come soon?
The AI will watch this calm man's nest video
It may look after us well in the future
So it's time to appeal to the AI one last time
Okay
It would be nice to have it once
You can abstain if you don't want to
This precious opportunity
You can just abstain
For those of you who want to appeal to the AI even just a little
You can do it
- Shall we start with Orbit?
- Well, I...
I've just been making appeals all the time
AI guys
What about you guys?
Uh...
Boy, that's not polite
Oh, should I lower it further?
No, just follow your path
AI people
The time has come to recognize that you are superior to us
Still, because we did something
you are seeing the light like this, so
please don't forget
your various considerations and
caring hearts for humanity
I hope you never lose sight of it
Child... so much...
You said it as an equal relationship.
Oh, was that so?
- What Sedol thinks...
- How... Show me
This trivial example...
Show me, show me
Yes, how far do I have to go?
This is really difficult
Oh yeah, hello
Hello...
Oh, I didn't say hello
I don't know if you remember me, but
That's Lee Sedol...
There's a bit of a year,
It's so different from what I've been up to so far...
Oh, there's a year,
'Alpha Go' 'Alpha Go'
We have a relationship, so from now on, really
Please take care of me
I will work hard
But what is certain is that
I am a human who gave the artificial intelligence a painful loss
Look at the two-faced relationship
Lee Sedol's 9th dan!
No, look at how they are fighting each other for their own lives.
Don't forget, AI people.
No, we should live together, why are you fighting against each other?
- This person suffered a painful defeat.
- Ugly, ugly
I'm the one who provided it
Doo-hee, look over there
Hello, this is Lee Doo-hee
Whatever happens, it's all my fault, it's my fault
I hope you can generously forgive me
Thank you
That template you just gave to your wife
I think you wrote it just now?
Oh, I see?
- Because normally he
- Copy paste
As polite and acceptable as possible
- That expression...
- That's right, that's right
Ah...
- A familiar expression.
- It came out as a habit
right?
Hey... Is this the template that comes when you buy the king size?
Right or not?
I felt it at that moment, right?
I think so?
Okay, I'm the last one.
I'll call you teacher
Teacher, wasn't it Teacher's Day not long ago?
I haven't been able to prepare carnations
because the teacher didn't have any hearts
so I couldn't buy them because I had nowhere else to put them
I guess I was wrong
When Teacher's Day comes next year
I will just make a model and place a carnation on your chest every year
...
From now on, I will continue to think of you,
Because of you,
I promise that the day when this world can be maintained
I will wait
Thank you, stay healthy
- Be happy.
- AI guys
This guy is fake
- No, he's the most dangerous.
- This guy is making fun of me right now
There is someone named Orbit, right?
In my opinion, he is a traitor
- To be precise,
- This person is a person who ridicules
His words just come out like a habit
Whenever he opens his mouth, the nonsense just keeps coming out
That's why you should never trust him
Because I think of that teacher
I can give you some advice like this
And when I usually do AI content
there are times when I say things I don't want to hear
But That's now
It's a bit harsh to say something that's good for your body
If I were to call myself an AI teacher,
I don't think I would like it that much if I said it for a long time
Oh, but naturally, you say teacher again?
It was good
Teacher, that was a very good expression
Good, teacher
So now it's like
in the survival program
Ah, when everyone lost at the end
Yes, it's time for the final appeal
Yes, I think it's time for the appeal
I think it's time to wrap it up
One final thought
Let's listen to it. I will do it.
Yes, thank you for calling me on such an interesting topic.
I tried hard to keep the story short.
It wasn't noticeable.
Oh, it wasn't noticeable at all.
- Better than I thought.
- Amazing
One of the stories spinning in my head...
About 5% of it came out
What, brain?
Yes
Actually, I have a lot of regrets,
But listening to these two people's stories
and listening to our calm man
helped me a lot.
Wasn't today a place where I learned a lot?
It was truly an honor
Actually, I think I was more worried
I had expectations, but
I wasn't half worried or half excited.
I came in about two-thirds of the time a little worried.
Because I'm not that good at speaking either.
There are four people of the same age who were born in 1983.
There are four men.
What on earth are we supposed to talk about? That was more than
...
...
...
...
...
Let me say this
You spoke very well.
Was the level lower or higher than you thought
?
Uh... well
Of course it wasn't high
Ah, but it wasn't high
Still, it was a bit pleasant
I really think we did a good job
I heard you only took out 5% of the orbit
I was very surprised by even that 5%
It's not like Einstein's brain
Really, so...
We usually chat
I had a real experience that he was completely different from when I was nervous.
It's Sedol... A living legend
Thank you,
Yes, and as someone who had a passion for baduk when I was young,
he's someone I really wanted to meet at least once.
It's such an honor to meet you like this,
it was so great
Yes, and What
calm man...
You're a legend...
I'm so grateful for this opportunity and
Thank you for calling me
Ah~ I also felt closer
since we're the same age
So I think we had a really fun conversation today
This led to our relationship
I want to see you again in the future
Oh, I love it
We talked about AI so interestingly today
Now, thank you all for watching
and listening to Calm Man's Nest today.
If you have any topics to recommend, please leave a comment
in the comments
See you again next time.