In the 80th anniversary episode, I will welcome

one of my favorite guests for the third time,

and that is Tomáš Mikolov. Welcome here.

here. Hi. I

Hi. I

'll start with a question that probably gets asked

a lot, but I'm not watching the scene

myself like I normally do. I always

wait until I see you to ask. Eh,

when it comes to artificial intelligence, are we still

relatively where we were in terms of what it

can do, or has there been some major

breakthrough in the year since we met?

met?

I wouldn't even say that there was a

turning point, when I think about it,

actually some GPT chat, which

we've had here for quite a few years, people have

gotten used to it. Of course, now the models

have different, better capabilities and it's

starting to be promoted more in different

directions, but I think that now

I hope that the public is not as afraid of it

as when it started, when there were

those calls that the

planet would explode, the world would end, and

so on. I remember it very well,

because I said back then that I would

remind people of it when it didn't happen, but

yeah, yeah. Well, for some, maybe. But

anyway, I think that like many

people, I've gotten used to it, that it's actually a

technology that has its

limits. It's not like artificial intelligence in that

it's smarter than a human

in every way, but it can

simply read piles of text or read,

process large amounts of text, other

images, it can be very useful. Well

, for me, I don't

see anything as a

breakthrough in the last year. Maybe what emerged,

when I think about it like this, is that

a year ago, I don't think there was Deep Seek from

China, and then there are a lot of other

Chinese companies that are now releasing

models like on an assembly line.

They are doing quite well because they are technically

well-off, they have a lot of money, and they have

government support to support the ecosystem. So I think that

if anything has changed significantly in the past year

on the world map of AI, I

think that China has already broken America's monopoly

. Right now it looks like it will

be a duopoly, there will be a lot of

competition. Actually, Donald Trump

announced the Stargate here, that it will be a

competition with China, which they consider their

main rival, who will simply be the

first to create AGI, the true artificial

intelligence, and just in that year,

from our European perspective,

nothing has changed. We are always last. Yeah.

Yeah.

Yes. So, is it, as you say,

more or less like America, is it

like a capacity

or some kind of a certain tool as a

function? In what ways are they actually better

or at least comparable?

Do you mean China?

Yes. Yes. The sun. I think it's a

combination of many factors. The fact that they

managed to reach that level is a

combination of people, er, talents, universities,

which actually produce a kind of

foundation on which to

build. But then there's the

industry, uh, state contracts,

state support, they had that too. And in the end, thanks to

that, they simply have companies there, I don't know, I mean

Alibaba, Tencent, Baidu and I don't know what

else. And they

invested a lot of money in the EA area here, and that

's why they have what they have there. We do

n't have companies like that here in Europe, and then

no one invests in them, and then it

looks sad here, yeah. But for me,

if I had to point out

any factors, it's simply that they

consciously aimed to

play a role in artificial intelligence, not only in artificial

intelligence, but in technologies,

internet technologies in general.

We can trace this back to about

15 years ago, when China started banning

companies like Google and Facebook, while

at the same time supporting its own

industry to build

clones of these companies.

So even then it was kind of the beginning of

them creating that position so they

could then create something like Deep Seek,

yeah. So Europe

started to claim that at that time, sometimes around

2012-13, projects like

like

the Blue Brain project or whatever it was

called by Henry Markam appeared for a billion euros. Maybe we

discussed it last time, I don't know now,

but it was a total fail. And

the Chinese model of competition with

America, it seemed to me that it was

planned in the long term, and

yet the communist party doesn't have to

fight for power there, they don't have elections, so they

could have called it the

Chinese capitalist party last time, and someone

really liked it in the comments.

He wrote: "I'm giving a subscription based on this."

this."

Well, they call themselves

communists, but realistically,

it's just a name, yeah, because they

behave more capitalist than

America. That's what anyone in China will tell you

, just what they grew up with there, yeah. Well,

but we as Europe didn't do it and

so we simply don't have any deep pockets here

. Since then, we have also founded

founded

Bottle Cap AI with a few people, where we are also trying to

do something new, because it is

not enough to just talk about these things

. Someone has to try something too.

So there are some attempts here and

we will definitely try, but

it's like our starting

position is quite difficult because of where we

start in Europe. We would like to

create something interesting here, but for now I

think that while the

Chinese are simply supporting Chinese companies, and the

Americans are supporting American companies,

Europe is still kind of stepping over the edge and,

more like, coming up with ways to

regulate something, how to ban it, how to be afraid of

something, how everything is simply

terrifying, and if we just put our heads in

the sand, maybe it will pass and things will

go back to where they were in the 1980s.

Well, they won't. Well. Hmm. Well, eh, that's what I

wanted to ask, what are you actually

doing now in the Battle Cup?

Yeah, yeah. So our big ambition

is to create an environment here for

interesting AI tools, AI

news and basically something like that,

well basically yes, something like a European

open we would ideally like to

create something similar and Open is not

just a GPT chat either, they had a whole range of

projects there and they will definitely be working on others

, yeah. So there's something along these

lines here. Well, we started by wanting to

significantly improve the efficiency of training language

models. That's what

the technology of all kinds of

chatbots is built on right now. A few months ago, for

example, we showed that compared to what is

currently commonly used, we are able to make

training about twice as effective,

which I know sounds abstract,

but they are currently investing a

lot of money in it. For example, this year,

when Facebook trained the Lama 4 model, it

spent about a billion dollars on that training alone

, as I heard. So

as you can imagine, if you

divide that by two and get the same

good result, that's a

lot of money, but it's definitely not the only

thing we want to do, because

training the model, thanks to that we will

create a model and so, to be honest, the model is

already available, because then again, if

you're just in China, it

's not that hard to get a billion dollars for a model. In

America it's even easier. And then you

actually have a lot of companies like that,

startups that no one really knows about, that

have models like

Openia, yeah, specifically from

China, I think there are

four companies there now. I think there was a

Ken model, then there was a Kimi model and then

two more names that I've already

forgotten and there will probably

be three or four more by the end of the year, yeah, so

one thing is to make that model, the other thing

is to make something out of it that's useful for people

and so on, so we're also

trying to make it into an application. Of course,

we can't talk about it here yet

until we release it, or what it will

look like, but we have

some directions in place where we could do

something that people could try out and that would

help them in some ways, so

that they can get to something useful. It

's just like me when I look at

various stories from America, of course

you just read it, so one of

the big challenges is, in

terms of energy, power,

water and these things actually for the various

data centers or as I

understand it. I've read a few reports that

in places where there are

giant data centers, there's actually a

certain level of very difficult sustainability,

if that efficiency were to remain,

people in the area are already feeling that they have

problems with water, they don't have enough energy, and

so on. So this is one of the

main challenges that is necessary

for the

phenomenon of these models to continue?

continue?

I don't know, I would probably be more

more

on the side of the free market, yeah,

you can actually imagine that

every time you see some big

news, if we

suddenly produce some super airplanes here or if we have more efficient

efficient

agriculture, or if we produce

computer chips, or if we have data

centers for artificial intelligence, it will

all actually grow in size

until it makes economic

sense. So if someone like

Google is now training artificial

intelligence models to actually

replace, say, the Google search engine in

complex queries, for them it's a

business worth

hundreds of billions of dollars, it's simply an

important thing. If they can build a

data center for $10 billion or

$20 billion right now, it doesn't really matter to them

. So they simply

push the size to the maximum, as much as they

can, when it still makes economic

sense, and the fact that it then drains

electricity and water from the wider area,

that's just life. They

just don't get along with it, it's just capitalism. I

believe that I agree with that, I believe that I agree with

that, right? It's more like a question of

whether there will be something here that, no matter how

you look at it from the left or the right,

will be a certain source of pressure, where let's

just say on the one hand, of

course, the race will be legitimate, because if

we don't do it,

someone else will do it,

and on the other hand, there will simply be an

increasing pressure on the

population, which if it were to affect

not only, as we were

talking about, monetary inequality,

but imbalance, but it could

also start to affect the specific ones, such as

simply living

resources, which people take for

granted, such as electricity,

so if this can also be a source of resources, I'm not saying

if that's right, but

if this can become a

subject of friction, then it

probably can happen, and if there

are simply significant economic

interests here, so that

certain resources are consumed, then there will definitely be

some conflicts. We knew that

even in the case of, I don't know,

Bitcoin, for example, someone will start mining it,

and thanks to that, electricity consumption will increase,

and thanks to that, the whole

price of electricity will rise, and so on. There

are just chain reactions like this. I do

n't think there's a simple

solution here, if I wanted to say something like a take-

off take-off statement, so if we

can make the training of

those language models more efficient, then not only will we

reduce the price, but we'll also

save electricity, but that's kind of like

it's more like

naive. I would say that if I wanted to

advertise here like this, but let

's be realistic. By making something more efficient

, it will start to be

used more, yeah, so the free market

and so on. So, like the idea

that money will be saved here, I don't

think it's entirely

relevant. Capitalism is more

about looking at where you can make the

most money, and, well

, you just ca

n't get very far by just saving. I think

that Europe is thinking a lot in this

way right now, as we see it with

ecology and everything. These are the

eternal lessons that we will simply save

electricity here, we will impose some

extra taxes here and we will create,

we will produce less greenhouse gases and thanks to

that we will save the earth and then we will find out that

it does not play any role in the global total

. And I have this

idea that in artificial intelligence, even

if we simply made

the operation of the AI module a hundred times cheaper,

their size would simply increase, a

hundred times more parameters would be put into it, and we would be where we were

. Clearly. Those models

will then be more accurate, better, and more useful. They will be

used where they ca

n't be used now, perhaps at a lower price, but I would probably be

very naive if I

said that we would save the entire planet and

suddenly we would all be sunny,

cheerful and artificial intelligence would be free,

well, it just won't be like that. Yeah, here's

something that when I talk about it, I'm

more likely to see it from certain

journalists who follow and it's

more about America than I think that in

Europe we're on the other side of that, where there are

regulations first and then

we actually make a product,

so maybe that was in line with that and

again in general, when we talk about, let's

say, artificial intelligence, right, they're

just throwing it in like or

machine learning that's [ __ ] smarter, they're just throwing

a lot of money into it

, right?

Eh, I suppose those people will

just want to see a real

product that will be better than the

next one. Do you think there is a risk of an AI

bubble that will more or less separate the

wheat from the chaff, where the projects that make

them more efficient will simply remain,

but a lot of money will go down the drain at the same time?

Yeah, but Silicon Vol, for example, is

built on that and it works. I think that

we have the exact opposite

problem as in Europe, that if I

put it simply, if in America they simply

calculate that if they try out, say, 100

projects, where each of them has, say, a 1%

chance of succeeding, but if they succeed, they will

earn a thousand times more money than

was invested in it, then it will

mathematically pay off. Yes, but if we in

Europe say that we won't

invest in anything that has less

than a 90% chance of working, then

all those opportunities will pass us by. Yeah,

sure, the AI Bubble is here, it was here,

and it will be here, and there are a lot of projects

that are completely stupidly conceived and

just burn money. I could just

talk about my field of artificial

intelligence, where in

the last five or six years

several startups have appeared in America, raising a

lot of money, billions of dollars,

hundreds of millions of dollars in those

years, only to burn through it

in two years without producing

anything useful. It just happens.

But if we were to say, based on this,

that it's good, then we're in a bubble,

we can't determine exactly how big

the AI market will be. Maybe it will be three times bigger,

maybe it will be three times smaller. We should

prepare for the worse case scenario and we should

say to ourselves, there might be a bubble here,

so we better stop investing in it,

stop working on it. Which in turn will

cause a cycle where

everything starts to get worse, we put

less money into it, people will be more skeptical about it

, it starts to fail and

in the end we do nothing. So that's

the European style. And then again, if we

look at the American one, which is

almost naively optimistic, which is kind

of boosting the bubble

because something here looks like

it could possibly work, then we'll

pour money into it and everyone there will

just have billions and they'll do

complete nonsense and a social network for dogs

or, I don't know, for cats and all that

nonsense, and sometimes something will succeed

and that will pay for all the nonsense around,

yeah. So if I were to look at it

completely disinterestedly, I think I can do it, the

American style, even though I don't

like their exaggeration, their

selling of what they don't have yet, but it's

simply been significantly more successful economically over the

last 30, 40 years,

while Europe, with its

skepticism and mostly looking for bubbles

and mostly such pessimism, will

definitely not succeed and it won't

work out for them either, and there's a bubble there and this is

bad and it won't work out there. Yeah. So

Europe won't win it all. So

maybe if we were somewhere in the middle,

it would be healthier, and if there is

someone ahead of us in the race, we should

look at what we should

improve on and what we should be inspired by from the person who has

overtaken us.

Certainly, certainly. I try to do it in a

way like a devil's

advocate or

actually go into those questions that I

think are

eh with progress just like, how to say it,

they are like inevitable. at least until we

sit here as, when we

sit here simply as two people and not as

some other, let's say some other

form. I'm still wondering, eh, how, eh, it's

probably not something that will be a question of the

next year or two. Mostly you

say that like eh we are too

optimistic about what can happen like in a

year, but not optimistic enough about what

can happen like in 10 years, that the

progress like how many times is between that, as

far as just some like the job

market, eh how actually expects that

this year like artificial intelligence can

affect just in the future you know, when you

listen sometimes

like very optimistic conversations with

people they just got siliconized, so they just

say like

in a few years like it doesn't exist, that if you

order something from Amazon, that any person would even

touch it just like a

human. All the people will be gone this week.

gone this week.

So you actually say to yourself, oh, what a

challenge, and ultimately, in one book

I liked that we will never get rid of

problems, but in the best case scenario, we will have more

interesting and better problems

in the future. H.

So what a challenge it will be

when artificial intelligence simply takes over a

number of jobs and a

certain number of people simply won't be

needed anymore.

Well, as in recent years, my perspective on it has probably

changed a bit. I

remember this question, it's

basically what journalists always asked,

yeah, maybe 10 years ago I remember being very

very

optimistic. I simply looked at it

from the perspective of progress, like

we have artificial intelligence, machine

learning, thanks to which we can

automate a lot of things. Actually,

computers as such can do this. I can

make a table in Excel, then

the computer will add up a million numbers for me in

one second and I don't have to do it

manually, I save time. It's actually good.

We have, we have a more efficient society, then

people can drink coffee, record

podcasts, do stupid things and still have

enough money to support themselves. Yeah.

So, eh, so on the one hand, I'm

kind of optimistic about the progress. On the

other hand, what has changed

or expanded my mood, maybe not

even completely changed, but what has

expanded for me, as I

think about it, is that these technologies will

affect different parts of the world in different

ways, yes, I spent a

lot of time in America, where I

worked here in those technology

corporations, and I also spent many

years here in the Czech Republic, and I see as such a

mental gap, that

people in the Czech Republic, who often haven't left

home, can't really imagine how it

works in that world somewhere in

those big corporations, where there

's just clouds of money for everything and the

goals are completely different. Here it's

like, how would I put it,

more fearful, yeah, in America it's like

terribly predatory and eh like

very, very different. But what I'm getting at

here in this discussion is that when it comes to the

labor market, if we

don't fight with those corporations and

fight with the rest of the world to

maintain the added value that

artificial intelligence will

generate for us, then we will actually

expand those scissors in the meantime, where the

profit from that work is actually going now, I would

say it simply,

if we could, for example, automate

30% of jobs in Europe using

tools that will be produced by America and

China. We will just be consumers. So, it will

help us a little economically,

but it will help our competitors

tenfold, because we

will actually outsource our work like this

. Similar to how Europe moved

factories to China and then of course that will be a

problem. In the future, it's a bit like we

're not producing anything here, so we'll move

other jobs, maybe I don't know,

lawyers, tax advisors or teachers and

what not, we'll basically move everything

to America. So for a while

we'll think how good it is that

we saved that money, but in

the end we'll just keep falling.

I said here last time, when we

were, uh, that statistic, uh, which is already being

repeated more and more in Europe, that

we have lost, I think, a third of our GDP in the last 15

years compared to America,

that's a huge number, yeah, that's

not it, that's not a little, yeah, just a third,

that's 30%, I think 2008 9

somewhere there. Well, Europe's GDP was

roughly at the level of America's.

We are now a third lower than in the past 15 years. Now, let's

say that in the next 15

years, our intelligence will be another third lower. And

where will Europe's relevance be in

the world, yeah? So I think that artificial

intelligence will definitely help us as humanity

, but it will create

imbalances somewhere, which can then lead to the fact that,

if I see it as perhaps too

pessimistic, Europe's

population will age, we will have a lot of

pensioners here, no one will work, we will

outsource our work abroad. What are

we left with here? We'll have

the regulations here, we'll have the officials here,

and we'll talk about how

great we were 150 years ago and we'll remember the

glorious times of Europe during

colonialism. But what will Europe actually

have, what will the world want? Now we

have an automobile market here, it will

probably take our money, for example, and we'll see

how it turns out. He's definitely trying now.

trying now.

What else do we have left here?

Technological news. There are a minimum of these here

and it is not a given. It's not

that the people here were stupider, but

the system just isn't

set up well for that. So I'm simply

concerned that this development, this

trend where Europe is collapsing, that we have

n't stopped it. And just to say, just to

admit that this happened here. That

took Europe years. I think it

was the Dragho report a year ago. It

took so long that

all the people who knew the economic situation

had it right in front of their noses for a long time, and only a year ago did they

say, well,

we're really losing economically.

Competitiveness is a big topic,

we have lost it. But the European

solution is to take the people who

made those mistakes last time and let them

propose the solution. They will propose solutions

in exactly the same style as the

solutions were proposed to them when they

caused the problems. So actually it's an

increase in bureaucracy, an increase in

regulations, an increase in redistribution. We'll

raise taxes, then pour those taxes somewhere

and hope that it somehow starts

working on its own. And that's socialism again,

where the official says where the

money is going and the investor who

invests his money somewhere and then has a

stake in whether it works out or

not says it. Well,

I think it's quite good as he said,

that we really had a

huge technological and

wealth advantage, let's say, in the 20th

century or maybe even

before that. Hmm.

And actually, we managed to draw from it for a very long time

. It's actually

basically about the fumes of how

well we were doing.

Hmm. But sooner or later,

the one who is determined to be in conflict with reality must come, when the one

who is actually standing still, and even those people

who started from a place much

further away than him, actually catch up with him

catch up with him

and are already in full sprint.

Whereas you have to say, yes,

those are the trends, that we are

not doing badly here yet, we are doing well in Europe

, but the trends are simply

downward, yes. So that's why I think that

if we talk about those problems, when we

simply have a GDP per

capita lower than, I don't know, Vietnam

or whoever, well, it will be too late, yeah.

So we need to talk about it again, but it's still

early. And again, for me, eh, and these

are political things, maybe it needs to be

done, but for me, it would probably be good to kind

of let go of the

over-regulated socialism that we have here in

Europe, the idea that we will have

huge taxes, huge redistribution and

some wise official will simply

invest my money somewhere wisely and do it

well, if it simply hasn't worked

here in the long term, then it probably

won't work much. and simply try to

reform the way Europe

works. It's just not just our

problem. I think that one thing is

that the Czech Republic is lagging behind even

within Europe, we don't have

any great results.

Just look at our region.

Now people are starting to talk about how

Poland is overtaking us, in salaries and so on

and in terms of GDP growth, so I

admit that I'm not completely

surprised by that, because the way the

Czech Republic, for example, is approaching the

union and intelligence, it also seems to me

that it's like, somewhere, it's just a

gesture and they say yes, artificial

intelligence is important, it's our

priority, and here in general, science, research,

innovation, that's important,

we're putting money into it. that then, since I've been

moving around the entire area for five years, I

can say that it's simply not true. I

guess I said last time that

I simply led a group here for five years at the

university and apart from subsidizing my own

money, I never received any grants for it

and I can say that

I am simply the most cited in video in the

Czech Republic and I have

n't checked the statistics for a long time. I hope that's

still true, but I just want to

come here as if I were from one of those top teams where

I worked with people who already have

Nobel Prizes. Maybe a year ago, when we were

having fun, it wasn't like that yet. For

example, a former colleague from Google Brain,

Jevinton, won the Nobel Prize for artificial

intelligence, which is something that

Nobel Prizes are not normally awarded for, so they put it in the

physics category. Another person I know

who founded Deep Mind, Demiza Sabis,

also received the Nobel Prize in the colony, and I

think he received it for biology or something.

Yeah, so that's pretty interesting.

So, the people I

worked with in America are

doing like this, and I'm coming back here for five years

without grants, so it seemed to me that

things would probably be a

bit of a mess. But that's just the point, how come

when some

politician or official comes here and starts

talking about how we innovate

and how we invest and how little

intelligence is a priority, right? No, it's not like

that. Hmm. It feels like the truth to me. E,

what I find very,

So sometimes maybe it's

too strong for me to say never. Yeah, it

's like I think we have

good examples and worse examples here. People

often talk more about the worse ones because

it's just more likely to arouse

emotions, yeah, so then you get angry

when you see that something isn't working,

like in this case, so I wouldn't be so

harsh as to say that

nothing is working here. But as it

is definitely true. It's definitely true

that I think a lot of people here are

just playing for the initial effect, that

something is simply said, I throw a number here, I

throw a number there, and it immediately

looks good. But besides

European socialism, I

have another thought that I would like to

share, and that is that it

really annoys me. I've just been watching

various election campaigns about who to

vote for or not to vote for. I

think the

main thing that bothers me, for

example, in these various election

promises and campaigns, is that politicians

often play the role of

pretending that they are giving money to something

, we fought about that last time, but like

that politician, he won't go to the bank and

withdraw money from his account and then

give it to someone. He doesn't give his money. For

the politician to say, "I gave 10

billion here for this," he

first had to take the 10 billion from someone somewhere to

give it to someone else. But

given how the

authorities and bureaucracy work and everything,

out of the 10 billion, it will be five, yeah, so

the politician always has to

take the money from someone. He doesn't give his money,

he gives someone else's money to our

money, our taxes, yeah, I would

shorten it, simplify it. I measure

where the wishes are,

who is better, who is actually better for those

people, who gave them more. I am your

dalili and I gave you 50 billion and I will

give you 100 billion. Well, yes, but they

all give not their own money, they

give our money. So I

think that if we had it

more widely here, it wouldn't be a bad

idea for people to simply not take it as if

someone promises mountains and valleys, that it's

actually good. Well

, yes, that's not it, I think it's

like a realization that it's more like a

redistributive role than like any kind of like.

like.

So I took 20 billion and I'll give you

10 billion, if they put it that way.

So it's like a general

problem with politics, right? Those things

must sound somehow, right, like they

really are. That doesn't

matter much. Probably the last thing

about artificial intelligence, which

interested me because I said in an

interview that I had

never used GPT chat or anything like

that before.

Hmm. And I don't think I want to say that

that's a bad thing. I think

it can be like a glass tool. And then we can

have a discussion about what I think

might be some of the pitfalls. Not for

society, for which it is of course

good, but for individuals, if it

actually starts with that, I don't know how to

replace the activity of the brain and creativity,

which I think is very

important for some kind of life. E, to

what extent do you actually use tools like Tydle

, or do you have your own,

like, um,

not like a system, but I do

n't use it explicitly so that

someone pays a subscription, but I

think that a lot of people use those tools

, they don't actually know, yeah,

so I don't know if you use

Google search engine, for example, it also sometimes

writes you an AI summary, this, that if you

write some more complex details there, it also

throws me into some language, so I can give you an

answer. And that was my vision, which

I actually presented

at Google years ago, when I was still

working there. So I'm just

not surprised at all by the potential I

saw there, that people won't actually have to

click on a page that someone

wrote and they'll get a personalized

answer, which by the way can be

significantly improved and I would expect

that in a few years it could be even

significantly better than now, when the summers

like from that AI won't be the same for

everyone, but it will be different for you and

different for someone else. So it will

still happen. Anyway, just by

reading the AI summaries, you are actually

using some kind of intelligence, and

if I were to put it more generally,

it would be just that you use the search engine

itself. There's also machine learning running there,

like the algorithms that select

the pages that are probably relevant to you

, yeah. So AI is

all around us, whether

someone explicitly pays $20 for a GPT chat

or reads the summaries

in an article somewhere, which itself

is generated by a chatbot and

you don't even know it, right? Well, it's a

bit of a wordplay, yeah, so even people

who think they don't use it

have it all around them, they just don't know it. I would

probably rather mean a certain

complexity, because as you say exactly, it can do it much better

, it can

answer weekly questions better, and actually,

where I would make a certain

division or whatever you call it, is

when a person uses it as a

tool, because here they can find an answer to it

. But there are actually

times when a person, let's say

, when you actually start with, you can, you can

say more like you're just

saying, oh, how would I put this into my

thoughts, I could have changed my mind like

before, but how many times the

work of the brain, in which a person is

simply engaged, when I don't know, he wants to figure

something out or he's reading a book and now he

comes up with some kind of theory in connection with it,

you know, that's like

creativity, yeah, that when you leave

your thinking to the computer, then

nature doesn't actually use it more or less, right?

So it's, it's similar. Maybe I've

said it before, it's similar to

when people just get used to driving a

car. For example, in America, it was sometimes

so exaggerated that people would drive to the

store when it was 200 meters away, because it never

occurred to them that people

could walk there, they were just used to it

, but then it leads to various

civilization diseases, because when a

person doesn't walk, their legs atrophy

and their overall health

starts to deteriorate, so if something

similar didn't happen to the brain. And I

don't know, well. It was said here

maybe 10 years ago that as people

use their mobile phones and maps in different ways, they

actually lose their sense of direction. [Music] I

[Music] I

don't know, it's hard to judge me,

because I saw it as often

different, such reserved opinions,

some who said it's all

fine, forget it, and other opinions

like we're rushing to hell and in two years a

person here without a mobile phone won't even be able to go

to the store and these things

and such extremist opinions are

heard much more often. that I have

never done any psychological studies

. I didn't really notice it on myself

, yeah, so, but again,

I'm not saying that the effect isn't there, yeah, but it's

hard to comment on it. And for me, it's

just that in these

intellectual questions, it often

seems to me that someone starts writing these studies

with a conclusion in advance, and when someone simply

wants to be against progress

and against new things, against

technology, they simply think, is there

some new thing here and how

could I write a story about it to disgust everyone?

disgust everyone?

There is a lot of noise.

Yeah. And by the way, this was in the

90s, when computers were just starting out here, and I

still remember how I was

playing computer games as a kid and

various television discussions and studies were coming out,

how it was just a

terrible disaster and how those kids would

just destroy it and stuff like that, yeah. So and

so, I think it will be the same with

my intelligence now. It's actually

that some people are trying to

claim that we'll lose creativity because of this

, but I'm not saying that it wo

n't be that way, but I don't really think so myself.

I can't prove it right now.

My opinion is that if we started like,

oh no, with the conclusion in advance, when we

want to drop something or

highlight it, then the companies that

sell it will do exactly the

opposite study, they will say it's really great and

people will come up with great ideas that they would

n't have thought of otherwise. But as if I were

somewhere in between, somewhere it will definitely take something away from us

, somewhere it will add something to us. I think

it will add more to us than it will take away,

because the fact that the model

can read me those 100 books and

make me some summaries that

will be useful to me, it seems like a

great added value,

for example, now that children are growing up in a different

world and with the fact that we are all simply going through

some development, that the brain

is kind of plastic up to a

certain age. For example, we

learn languages at a younger

age and then we lose that ability, so

for older people it is

basically impossible to go back to that

childhood and imagine what it would be like to

grow up now, with a mobile phone in their

pocket and they can

ask a very complex question on their mobile phone

and it will actually pull up something for them that they would

never have read on Wikipedia before,

because there is no such Wikipedia page,

because it is really some complicated

thing. And similarly, when Wikipedia first became

available, it

was a lot better than when people

had to travel to libraries by

train to a district town, because that

knowledge simply

wasn't available before the internet, yeah. So I think it will

advance the possibilities, what all the

new generation will be

able to do within a few years of their

lives, to get their bearings in

some interesting areas. I would

be optimistic about this, that

thanks to this, people will simply be able to

books to a library or read

encyclopedias, but maybe it will be people who

will chat here with chatbots, when they

are, say, 80 years old, but they will get

to some super interesting topics there and

will educate themselves here in a faster, more

efficient way. And maybe that's how it

will be, that's how

I optimistically think. Well, of course we won't

know for sure until decades later. Well,

sure. I think it's about the

measure, right, which on the other hand

hand

is nice and good, what they say like

arrest all change simply doesn't work

in the long term, the world will somehow

develop anyway, and actually a person who would

just stick to some of his

idealistic values and the world would

move on without him, will simply be the last one

standing at the station. I'm more interested in the

extent to which it can actually be used as a

tool, which

can actually have more

good effects.

The second is when he actually starts

using it as a person and now I don't mean

in some kind of Matrix or

these things,

but actually as if at the moment when a

person starts, how to say it, when he actually

becomes just a vessel that

actually only functions based on

some physical, as he said,

stimuli simply as connections or I'll

look here and actually, how to say it, he

only receives it, but doesn't consume it, does

n't consume some thing. And there I'm

actually wondering if, like, if we don't

become something

between a human and a machine by that time,

to what extent, and of course it does

n't happen from year to year, but over decades,

decades, hm,

hm,

the brain can actually

say, evolutionarily, I don't need this part as much anymore

, the part that was, I don't know,

responsible for

doing the harder thing more than once, even though

it's a harder thing, which is also a

great book about, actually, behavior, it's

called behavior, by a

neuroscientist who

talked about it quite interestingly. To what part of the

brain is this space?

More or less, it's not used much anymore

. Plus, it's like very

complex or very like energy-

intensive, right, the prefrontal cortex

and the tydlencty parts. To what extent

can we actually start withdrawing back into

that bigger shell, because we actually

need more like tydlencts to a certain extent

like automated eh like connections.

Exactly. I can find it simply

as a summary. Great. Well, for a person

who actually knows what they want to

use it for, I think that can

lead them to the position where they can do something that a

person couldn't do 10 years ago

. Hmm.

And actually, maybe the question is aimed at the more

ordinary person, right,

right,

who doesn't really know how to use it

as a tool, lets it use

itself and then it can actually just be

like, I don't know, like a

body, which is like floating somewhere and

here like simply receives the

information, here comes the hormone,

here receives the food and it's actually like

maybe it's like a stupid analogy, but

like actually, but actually like a vessel

vessel

that actually consumes, but

no longer digests the information, which

like walks towards it.

And I wonder if we're

not a bit like a civilization in that state,

but oh well, oh well. Sometimes it seems to me that,

uh, how the world's population has grown

is unbelievable, for example, in the

last 200 years, that I don't know,

tenfold, from hundreds of millions to now we have

billions of people, and

we can find it with the help of artificial

intelligence. Yeah, it doesn't seem to me that we're

10 times more efficient than

humanity, because when

you travel around the world, you start to

feel a bit like you're in the

Matrix, yeah, in that you just

fly somewhere for a few hours and get off

in a city and people are watching football on the

same things, drinking the same beer and

everyone is having fun with the same

things, but overall, as you said, the

container, that you just

basically just mechanically fit into

some kind of cycle that

some corporations scribbled on paper and then you get

involved and just like that they actually

forward those things back and forth and it goes from nowhere to

nowhere. So I

think we're a bit like with

globalization, when we look at

how much we've actually

lost as humanity, when it comes to

exploration, the exploration of different

philosophies, directions, technologies,

and so on. Just because we're all so used

to the fact that everyone knows Ronald and

everyone just knows Trump and everyone

talks about the same movie and listens to the

same singer. Just like I think,

if I were like the

Matrix here, a pessimist, I would say that

we are maybe a little bit there already. But maybe you

meant the question a little

differently. Did he understand it well, could there be

some kind of

brain atrophy, if I understand it correctly, that if

that if

people don't use their

brains because they aren't forced by

external conditions and if a

computer can do the thinking for them, then we'll

actually be a bit spoiled. Yeah,

if I think about it like this,

I think it brings me back

to what

society wants from these people

. I think that when we have the company set up

correctly, it's always about

pushing it as far as we

can go. And a properly

set economy should be about

how we look for new ways, new

opportunities, new possibilities, how to

simply make a company more efficient, how to

produce something. And if something can be

automated, it will be automated.

There won't be the person who will play the

role of the automaton. So here in

that case, I would

n't be afraid of it at all, because we could

deduce that people have long since become

complete zombies and are just

here in those corporations, they come to the

office, they do the work they were

assigned. So, so eh, I would see it

like this. If I see anywhere that it

could zombify us like this,

it's an area that we

didn't talk about today. maybe last time. And I think

education is quite important. Mmm.

And I think that if we allow ourselves to be

locked into those stereotypical ideas

that we should be like parrots who

memorize certain data, when even

50 years ago it made at least a little

sense because there was simply no

internet, no Wikipedia and no

language models at all, then a

person who simply knew more

about the world, because they simply learned it

in school, as if they were more educated,

had a little more insight and then it could be

expected that they would have better performance in their

working life. that it will actually be more

successful, it will create larger, more

valuable societies, so in

my opinion, this kind of education made

sense there. Of course, I would still

prefer something

more creative and more education focused on

leading somewhere, but even the

education we had here 100

years ago

had its drawbacks. But the world has moved on in the meantime.

And if I've

extended the education system now, yeah,

just a few decades ago it wasn't

common for people to study at university

until they were almost 30 years old, and today it's no longer an

exception, yeah, so we're

kind of extending the time that children

spend and adults actually spend

time in schools, but they're still learning the

same things there as they did decades ago. And

the world has only changed in the last

three years. And if

we teach these people now, when they have those adaptive

brains and when they

can still learn quite quickly, when

we teach them that the most

they can do is memorize a

poem, memorize a list of things, that when

I say a b, he will answer me de E and they will

simply know that here

is the correct definition for this, here for

this problem they will use this

formula here and they will remember everything, but

then they will come to the job market, where they

will now have global competitors, it is

simply no longer a local competition, the world is

globalized. and I'll just go to a

company and it will produce, for example, these

modern language models, like we do,

or it will produce some chips or it will

produce, I don't know, 3D printers or it

will produce cars, yes,

yes,

or some software, or some

consulting services, it doesn't matter.

These people will simply have to create something in their work

, some added value, and

when they go through an education system that will

force things into them for three decades that

are long outdated, and in the

job market, their company will

compete with, for example, a Chinese company or an

American company, where they

didn't play around with the gimmicks here, but there they simply

said, aha, we can do the job

10 times faster, cheaper and better if we

just use modern tools. We'll

use the internet, we'll use language

models, we'll use AI, it doesn't matter, if

it leads to a result and it's good, we'll

use it. If it doesn't arrive there, we

won't use it. Just like

normal thinking, capitalism, so

we, if we are built here on doing it

like the three

generations ago, because that's the

right way, we have to follow the ritual,

we won't change it, but then

we will lose in the global competition and I think

we are already there. what

I am describing. This is not some dystopian vision of

the future, but this is where

Czech industry actually is. Well,

no. Education is something I

want to pay more attention to. I have a

few guests lined up for the education department,

some of whom will talk about this and some of whom will talk about that.

Chile, eh, today I don't feel like

I know enough to answer that. I

could tell you one incident that

just comes to mind. I went to

see some school, some

event, it was a high school here, uh,

here in Prague. I've already forgotten the

name. Sorry, I'm not exactly a

genius at this either, but what I want to

say is an incident, and the kids there just

had an event on artificial intelligence

and they wanted to invite some experts to

comment on it. And I was

quite surprised at the time that the task

they were given by the teachers was to

detect cheaters with artificial

intelligence, that they did the homework,

but with the help of artificial intelligence, so it

's actually like they copied it, it's

actually like a crime and we have to

find them and punish them. I said that it's

just completely stupid

thinking, yeah, because we have

some technological innovation here, thanks to

which it turns out that the way

the children are given tasks here is probably

not entirely good. So instead of

thinking about how they assign these

tasks, so that it reflects the

current world, so that the tasks make

sense in 2025, they will keep

playing around with the idea that the person who uses the

calculator in math class

is a cheater, the person who uses the

GPT chat is a cheater, and the person who uses the

internet is a cheater. But when

those people graduate from school and go to

work and then they have an employer

who simply has a customer and pays them

$10,000 for wanting this and that

and then that employee

works there for six months or two days,

then which company will survive, which

employee will stay working there and

which one will have to go to the employment office,

yeah. So, I think it would be great

if we stopped pretending that

education and training are something lofty,

some kind of cloud, and that this is the right way to do

it. And if we looked at it more

as what the industry wants, what it really

needs, what it will want in 10 years and

not prepare those people for the world of 19,

1960, but for the world of 2040. Well, there

's a good point that you actually make, which is the

contradiction between a certain status quo

, simply the status quo, the way it

was, and what

we actually call the next fields, which

is actually a certain adaptation, an

adjustment to the fact that we

can never simply close time

or return to something that once

worked here and live from it, which is

actually the one, which is also perhaps related to

my point, because that's actually where

I was heading, with that power.

So I think that the tydlencts as

ideas, the resistance that I

have to it, is that I'm actually

trying very hard to explore what it is like to be

human, what it is like to be human in our

kind of like actually in some kind of

evolutionary like development. But then you think

think

that our anthropic or

anthropocentric approach actually

leads to this, and of course many

religions also look at it as if

humans were the foundation. And for me, the

idea that I sometimes actually argue

against myself from this

position where I said, yeah, wait, no,

maybe the progress isn't right, hm,

hm,

is that as a human being, there are 200,000

years out of like billions that we've been

here for, like,

a while. And I think that the

one thing that I'm slowly

letting go of and that is the basis of my

defiance is actually the emphasis on being a

person today. However, if we

look not at the universe from a

human perspective, but at a human from the

perspective of the universe, we are actually like a dot.

dot.

And that was actually the topic that I wanted to

bring up, that we kind of drifted

towards, how

you actually look at the role of a person from the

perspective of the universe.

if it's like one of those

like actually the most advanced,

like an emergent phenomenon, and it

doesn't really make sense to

assume that it definitely does

n't end with humans.

Hmm. Yeah, that's very philosophical and

quite complex. I don't know how to respond to this

without offending anyone,

but I don't think humans

are the most complex thing in the universe. Well, more

like from the ant's perspective, the most amazing thing

is another ant, of course, because

it's like that, we all see

only a very narrow slice of reality

and can't even imagine reality

as such, because we're not equipped for that

and we simply don't have the senses for it, we don't

have the brains for it. I just kind of have a

feeling that the things

around us are much more interesting and in a way

more complicated, in another way

simpler, than we can

imagine, because all we have to do is

try to empathize with other

organisms, which we

consider to be less advanced. For example, I

don't know, if you imagined you were a dog

and someone showed you how the square

root works, you'd never understand it either,

right? But such things exist, but

it can be similarly shifted, that there are things

like square roots for us as

dogs, for humans, which we are absolutely

unable to understand right now, but

one day someone, perhaps our

successor or some person who

will simply have an expanded brain will

get there through artificial intelligence, yeah. Hmm.

What would it be like, I know that,

as you say, because a dog can't understand

square roots, it will be difficult for a human to

even think about this,

because it's like you have it as

if he were asking what thoughts are that

you can't express in language. I

won't tell you that either. Well. Well, of course, no. He's asleep

by definition, because if I

told you, I'd deny it. Well,

I was more thinking if it's

going in a direction like, I don't know, into

quantum phenomena, let's say, into

various theories, like multiverses.

multiverses.

No, no, no, no. Not at all, not at all. It

seems to me that quantum

things often get people involved when they don't want to

offend anyone, yeah, just when

they want to seem interesting, but

at the same time there's not much substance to it, but I do

n't want to completely reject it, yeah, but it's

just that it's such a universal

answer. I've heard quantum phenomena

as the answer to artificial intelligence and the

universe and the nature of consciousness. By the way,

it's the answer to absolutely everything.

Universal answer. When you have a

complex topic, you want to seem like you're

intellectually advanced, so to speak, and

maybe quantum phenomena could solve it.

So, look, sorry that I'm such a

cynic after all these

decades of listening to this stuff

, but that's about it. For me, not for me

, more like the complex systems. Well, I think

that there are simply enough things that we are

not able to think about from the

perspective of our, uh, working

memory, yeah, that, uh, I don't know, those are those

psycho-psychological lessons that

we are able to keep in our

working memory what we are currently

thinking about, some like six plus or

minus three, I think, those like

some elements or

objects. And it's interesting that when you

look at different, like, brilliant people in

history, you find that they were, for example,

like, significantly above the average,

like the average population, because

their brains worked a little differently.

Maybe he sewed it, right? So it's

just that these people then have to memorize

some 20-digit numbers or calculate

some 10-power of some large number,

because they can simply retain that memory,

maybe I don't know, maybe it's not like they have to

learn how to hack

the calculations here, by the way, so it doesn't have to

be true. For example, if instead of just

those six things, you could hold on to maybe 20

or 30 people by having

visual thinking, by imagining those

things and then they wouldn't disappear as quickly

as they do for normal people, and you could

write them on paper in your head, then

that alone would move us

forward tremendously. And then we simply know

examples of people in history who

had such superhuman

abilities and then were able to achieve

results that no one

else could achieve. So that's an

example of the direction it

could go, the way super

intelligence goes, if we were able to

expand our thinking. Imagine, for example,

that when you're trying to recall

a memory, you have

some kind of hint in your head that makes it

accessible to you quickly, that you could see it

written in front of you, that you

could write it down with your eyes like this,

and many things would suddenly be

solvable and much

easier. And we could get

to that level where we would simply

solve square roots like those dogs, who

seem to have managed to move even

further in terms of evolution. Yeah, for example, if I were to say what the

square root is for us and those classic

things from the AGI communities and maybe

immortality, we might be able to solve

aging. Well, maybe, if we continue to

research and

study telomeres and renewal and

these things for hundreds of years, I think it will be a while, because

the human organism is like a

machine and it works for decades and then

ages, so I think there are

various justifications for why

evolution actually invented aging, that

it is actually how to prevent

population stagnation and how to kick-start that

progress, so that there is a possibility that

something new will simply arise, because you

need to change something, try

new paths, how to

get ahead. After all, one of the

main things that kills these organisms

are various infections. You need

the organism to also go through evolution,

for example, in terms of the immune

system, yeah. But what I was

saying is that this way we would be

able to simply get to the level where we

would be able to think about other

things. And that is unsolvable for us now,

like writing square roots, like

immortality or solving cancer,

just those things that bother a lot of people

right now because probably no one wants to be

sick, everyone would like to be healthy,

young, satisfied, happy, rich, and I think we

could get there

if we had, like, not

ten times, but just

twice the intelligence, yeah. that it would be

like we could already be, for example,

the difference between a dog and a human

is not as striking as we, as homo

sapiens sapiens,

tell ourselves, as we try to pretend. We are a

little smarter than some

animals, but there are also animals that

surpass us in some ways,

such as our ability to solve problems.

If we were to define intelligence so broadly

, for example, different

animals have better sense of smell. As for the

visual memory I mentioned just now

, there are

experiments with other monkeys, for example,

which are better at it than humans,

yeah. So we're definitely not

optimally intelligent enough to

not even imagine that things could be better.

As we know, there are organisms that

are capable of solving some things that

we simply have limits to. Hmm.

Yeah, I'm not the kind of

person who thinks I can

explain everything as quantum physics

or quantum, to make it more like metal, I wouldn't

even think I've

focused on how you

could actually think about this. This sounds

YouTube TranscriptPreparing your results…

YouTube Transcript:
Tomáš Mikolov: AI, GPT, DeepSeek, Člověk, Vědomí, Vesmír, Evoluce #80

Video Transcript

Paste YouTube URL

Get Our Chrome Extension

Get Instant Transcripts: Just Edit the Domain in Your Address Bar!