# Transcript: Artificial intelligence? What on earth should that kid be studying now!

## Description

Who is Josef Šlerka



Josef Šlerka teaches at the New Media Studies program at the Faculty of Arts, Charles University. Recently, he has been focusing on using AI for research in the humanities and social sciences. Increasingly, he is also exploring the broader context of this groundbreaking technology and its impact on society—from the spread of disinformation to effects on education.

## Transcript

**[00:00:00]** Boredom is not a problem, boredom is a space.

**[00:00:03]** These people are the future of the American economy.

**[00:00:06]** Not every resistance leads to good formation.

**[00:00:09]** This is the core of the problem.

**[00:00:10]** And I say, AI is too good a mother.

**[00:00:30]** My experience with large language models started rather poorly.

**[00:00:44]** About five years ago, my colleague Rita Malečková,

**[00:00:47]** who worked on the Digital Philosopher project, which many of you know today,

**[00:00:51]** came and said there is a new technology now, GPT-2.

**[00:00:55]** I listened and thought to myself,

**[00:00:59]** here is a sandbox, here are the shovels, this looks good,

**[00:01:04]** do what you can.

**[00:01:06]** And I thought it was all nonsense.

**[00:01:10]** Then GPT-3 came along.

**[00:01:12]** I remember sitting with my colleague Eva Marková in Delmart

**[00:01:16]** on Národní třída in summer, it was May,

**[00:01:19]** and we were grading exams there.

**[00:01:22]** We were absolutely fascinated, back then the system was still autocomplete,

**[00:01:27]** it wasn’t a chatbot yet, it just continued the text.

**[00:01:31]** And it was quite successfully passing exams on a level that isn’t random,

**[00:01:35]** the entrance exams.

**[00:01:38]** Sorry, the maturita exams.

**[00:01:42]** In July that year, we were doing poetry recitation at the Šrámková sobotka festival.

**[00:01:47]** In the style of Fráňa Šrámek.

**[00:01:49]** The Šrámková sobotka festival is a festival of Czech literature, language, and poetry,

**[00:01:54]** to give you context, with many Czech language and literature teachers,

**[00:01:58]** and many students studying Czech language and literature.

**[00:02:01]** The next day we published the poems,

**[00:02:04]** generated by that AI,

**[00:02:07]** in the local festival newsletter.

**[00:02:10]** And one of Fráňa Šrámek’s poems remained there

**[00:02:14]** among the generated ones.

**[00:02:16]** And actually none of the teachers could tell

**[00:02:19]** that the real poem was among the fake ones.

**[00:02:24]** I was starting to get quite frustrated with this.

**[00:02:28]** Then of course came the moment when GPT-3 and then 3.5 appeared,

**[00:02:34]** memory no longer serves.

**[00:02:36]** And sometime in January, about three months later,

**[00:02:39]** I published my first article then in Reporter magazine,

**[00:02:43]** and I always come back once a year to write an essay about it,

**[00:02:48]** how I perceive what will come next.

**[00:02:52]** And by publishing it in a magazine, many acquaintances started coming to me

**[00:02:56]** and at various events where I lectured about media,

**[00:02:58]** they began to approach me and say,

**[00:03:00]** please, I have a child, what kind of school should I choose for them?

**[00:03:05]** What will happen now?

**[00:03:08]** What on earth should that child study?

**[00:03:11]** And the nice thing was,

**[00:03:13]** usually it wasn’t people with kids in kindergartens who worried about it,

**[00:03:16]** who understood that the child would have to finish elementary school,

**[00:03:19]** but it was exactly the same parents deciding

**[00:03:22]** where the child would go to high school.

**[00:03:24]** It was exactly the same people who were dealing with the problem

**[00:03:27]** of where the child would go to university.

**[00:03:31]** And the arguments gradually increased and appeared more and more,

**[00:03:35]** like GPT will write a thesis in an afternoon,

**[00:03:38]** AI will replace programmers,

**[00:03:40]** robots will diagnose more accurately than young doctors.

**[00:03:45]** And you’ve probably heard many of these arguments.

**[00:03:47]** I in no way question

**[00:03:49]** that ChatGPT is capable,

**[00:03:51]** and now take ChatGPT in quotes like this,

**[00:03:54]** if you guide it well to write a thesis,

**[00:03:57]** 95% of people won’t notice

**[00:03:59]** that it was written with the help of ChatGPT.

**[00:04:04]** I tried it myself, it works great for me.

**[00:04:09]** The only problem is that the prompt pipeline is 18 pages long

**[00:04:13]** and I’m a bit afraid that none of those students would be able

**[00:04:17]** to build, write, and check it by the end of their studies.

**[00:04:22]** Very interesting, right?

**[00:04:23]** You can’t solve it by just saying, write,

**[00:04:26]** but you actually have to create some system,

**[00:04:30]** where, when you finally look at it,

**[00:04:34]** you’re doing the same thing you did before,

**[00:04:36]** only faster, because you had those tools available.

**[00:04:41]** And what all this makes me think is,

**[00:04:46]** that people’s anxiety is understandable,

**[00:04:50]** but on the other hand, it’s fixated entirely on a problem

**[00:04:54]** that doesn’t exist.

**[00:04:56]** Or rather, it’s fixated on a problem that can’t be solved.

**[00:04:59]** I think the whole debate we should have

**[00:05:02]** is not about choosing a school,

**[00:05:08]** but about looking at what schools actually do with people.

**[00:05:14]** It doesn't matter if it's elementary, secondary, or higher education.

**[00:05:19]** And from that, we can build on what we do

**[00:05:22]** and try to look at it from a slightly different perspective.

**[00:05:27]** In 1991, Robert Reich,

**[00:05:30]** who later became the labor secretary in the Clinton administration,

**[00:05:36]** published a book about

**[00:05:40]** the situation in contemporary America

**[00:05:43]** and how to keep the American national economy strong and competitive.

**[00:05:47]** He says, we have people who work by making things.

**[00:05:51]** They stand in factories and produce something.

**[00:05:53]** That's good, but it won't survive,

**[00:05:56]** because all of that will move to Asia, India, and elsewhere.

**[00:06:00]** Then we have people who take care of other people.

**[00:06:04]** They work in services.

**[00:06:06]** That can't be easily moved to Asia,

**[00:06:09]** because you can't have a barber in Shanghai and live in San Francisco.

**[00:06:15]** The downside is, it can't be scaled except by more people,

**[00:06:20]** and the downside is, you can't relocate,

**[00:06:23]** because you're a barber in Shanghai.

**[00:06:26]** He says we have a third type of work that is now growing strongly.

**[00:06:32]** He called them symbolic analysts back then,

**[00:06:36]** and he explains who these people are.

**[00:06:38]** These are people who work with symbols.

**[00:06:41]** You give them some symbols,

**[00:06:45]** they do something with them and produce new symbols.

**[00:06:49]** This sounds very abstract.

**[00:06:51]** Marketers, advertisers, lawyers, journalists,

**[00:06:57]** all these people. Programmers.

**[00:07:00]** All these people do one simple thing.

**[00:07:03]** They absorb symbols, texts, images, tables,

**[00:07:07]** do something with them, and produce new symbols.

**[00:07:11]** And according to Reich, the American economy in the coming years will rely on these people.

**[00:07:17]** These people are the future of the American economy.

**[00:07:21]** Why? Well, for example, because to make these people really good,

**[00:07:25]** they need to grow up in an environment that nurtures creativity.

**[00:07:29]** They need to learn to think abstractly.

**[00:07:32]** They need to live in a free community where they exchange opinions.

**[00:07:36]** And that's everything we as America have.

**[00:07:39]** That's what Asia doesn't have, that's what China doesn't have.

**[00:07:42]** What we have is freedom, education, the ability to experiment,

**[00:07:47]** to work with mistakes, that's our gold.

**[00:07:51]** And actually this program, let's change it,

**[00:07:55]** the economy of symbolic analysts worked precisely for 20 years.

**[00:08:00]** Why? Because the share of professions that do this is large.

**[00:08:05]** And they create a large share of the United States' GDP.

**[00:08:12]** Well, it worked great for 20 years.

**[00:08:15]** That was around 2001.

**[00:08:17]** In 2000, around 2001, Robert Ray said,

**[00:08:21]** look, it still works, but we're starting to pay quite a high price for it.

**[00:08:26]** He said the price we pay is 24-hour availability.

**[00:08:31]** We got used to experts being available 24 hours.

**[00:08:34]** And because they are nomads and have smartphones, they always answer our emails.

**[00:08:38]** We got used to there being no working hours.

**[00:08:41]** We got used to doing work that makes no sense.

**[00:08:44]** We got used to doing work so that other symbolic analysts can work with it.

**[00:08:47]** Kids, that's not very good.

**[00:08:50]** That was 2001.

**[00:08:53]** And 20 years later, suddenly large language models appear in this field.

**[00:08:56]** And when you say large language model, you mean,

**[00:09:07]** it's a mathematical function

**[00:09:10]** that can transform one set of symbols into another set of symbols.

**[00:09:13]** That's a large language model.

**[00:09:17]** A large language model is a mathematical function,

**[00:09:20]** which takes one type of symbols as input and outputs another type of symbols.

**[00:09:23]** So what exactly does it do?

**[00:09:27]** It rearranges, juggles, communicates, transforms those symbols.

**[00:09:32]** That means large language models are the system

**[00:09:38]** that comes into the economy with a precise strike on what

**[00:09:43]** was supposed to be its core, essence, and distinction

**[00:09:47]** of the national economy of the United States from the global one.

**[00:09:51]** It says that over the years we adapted the whole thing.

**[00:09:55]** Here in this country, we did not adapt it.

**[00:09:58]** We think heavy industry will save us, as we'll say in a moment, and in any case for many professions

**[00:10:00]** and as I saw people here in the audience and I know them, you live this, you are symbolic analysts.

**[00:10:12]** Maybe there's a barber among you? There isn't, right?

**[00:10:20]** And from this tension and from this now named reason for us, we ask the question.

**[00:10:39]** We're losing jobs, AI will replace us. WTF, how can AI replace someone who works with their mind?

**[00:10:48]** It was supposed to replace people who work with their hands.

**[00:10:52]** That's the myth we've carried for the last 50 years. Robotics, replacing people on assembly lines.

**[00:10:59]** And suddenly I lose my job, so how many jobs will remain?

**[00:11:07]** Will there be new jobs? Should I just quit now and retrain as a barber?

**[00:11:17]** I think we're mixing up two different things with this question.

**[00:11:26]** If we look at it as a story, you're a surgeon.

**[00:11:31]** A young, promising surgeon. You're a doctor starting in surgery,

**[00:11:37]** for two years you hold hooks for the pro, hand him instruments, sew up wounds.

**[00:11:44]** And they won't let you do anything. It's boring, completely frustrating, repetitive work.

**[00:11:51]** Then one day comes the moment: now you try it.

**[00:11:56]** And you come in and suddenly realize you know what to do.

**[00:12:01]** But not because you read it in scripts.

**[00:12:04]** Not because you passed an exam.

**[00:12:07]** But because you were there doing boring, repetitive work.

**[00:12:13]** A cook. Like an apprentice in the kitchen.

**[00:12:17]** You peel potatoes, chop onions, someone who watches the stove closely bosses you around,

**[00:12:23]** just yelling at you. They won't let you do anything.

**[00:12:28]** And after two years you touch a fish and know if it's fresh or not.

**[00:12:34]** You just know because you've been there a long time.

**[00:12:37]** Years of chopping give you something the cook can't.

**[00:12:43]** Imagine if a robot did those tasks.

**[00:12:48]** The output would be better.

**[00:12:53]** Sorry, with potatoes 100%.

**[00:12:56]** We have machines that slice potatoes.

**[00:13:00]** The output would be better. No doubt.

**[00:13:04]** But you would remain the same.

**[00:13:08]** This is the core of the problem.

**[00:13:11]** Thanks to AI, we've discovered that in our companies, schools, and everywhere,

**[00:13:20]** there's a kind of special load-bearing wall.

**[00:13:24]** What we do has two parts.

**[00:13:27]** The product and the person producing it, who is also shaped by what they do.

**[00:13:35]** What we do creates who we are.

**[00:13:39]** And that's what we're seeing now.

**[00:13:42]** Imagine a company decides to tear down the OpenSpace

**[00:13:46]** and now there's a wall and they say everything is gone,

**[00:13:51]** so we'll remove that wall too.

**[00:13:53]** Then a structural engineer comes and says if you remove that wall,

**[00:13:57]** it will collapse on your head.

**[00:14:00]** You saw that wall there.

**[00:14:03]** It annoyed you.

**[00:14:06]** You couldn't walk through it, you had to arrange things differently.

**[00:14:10]** So let's tear it down.

**[00:14:12]** The wall wasn't invisible.

**[00:14:14]** You just sometimes didn't understand its purpose.

**[00:14:17]** Because it actually worked.

**[00:14:19]** So you quickly noticed what it was for.

**[00:14:23]** I say every job has two layers.

**[00:14:27]** A visible layer and an invisible one.

**[00:14:30]** The visible one is the output, the product.

**[00:14:34]** Chopping onions, writing a report,

**[00:14:38]** delivering to the client with so much KPI improvement.

**[00:14:42]** But this layer is created by work.

**[00:14:46]** It can be measured.

**[00:14:50]** And it can also be delegated to AI.

**[00:14:54]** The creation of that product can be delegated to AI.

**[00:14:58]** We can discuss how good it will be.

**[00:15:00]** We can discuss if there are areas where it will never be good.

**[00:15:03]** We can discuss how to properly supervise it.

**[00:15:07]** But now I'm interested in the second column.

**[00:15:11]** Matěj Novák is still here with us,

**[00:15:13]** with whom we attended philosophy lectures.

**[00:15:16]** I tend to switch to philosophical jargon.

**[00:15:20]** So I'll try to resist that.

**[00:15:22]** I'm interested in the invisible layer.

**[00:15:26]** It's the transformation of the person.

**[00:15:27]** It's the transformation of the one doing it.

**[00:15:30]** What is gained?

**[00:15:32]** Just one last thing.

**[00:15:33]** Hexis, Matěj.

**[00:15:35]** You gain a feel for the material.

**[00:15:37]** Judgment.

**[00:15:39]** Yes, that's what you gain.

**[00:15:41]** How do you gain it?

**[00:15:43]** Well, not just from the work itself.

**[00:15:46]** But from encountering resistance in that work.

**[00:15:49]** From failing.

**[00:15:52]** From not managing something.

**[00:15:55]** From running into something.

**[00:15:58]** From having to cope with something.

**[00:16:01]** In encountering that resistance,

**[00:16:03]** in the fact that not everything goes smoothly,

**[00:16:06]** you are transformed,

**[00:16:08]** because you have to perform.

**[00:16:10]** You have to push yourself to

**[00:16:12]** do something you otherwise wouldn't.

**[00:16:15]** Can this be measured?

**[00:16:19]** Simply, no.

**[00:16:23]** Can it be delegated to AI?

**[00:16:25]** No.

**[00:16:26]** Because that is you.

**[00:16:28]** That is not for AI.

**[00:16:30]** And the wise probably already sense

**[00:16:32]** that clearly, working with AI,

**[00:16:34]** you will gain some qualities.

**[00:16:37]** Of course, working with AI

**[00:16:39]** will also shape you in some way.

**[00:16:42]** But how?

**[00:16:44]** What impact will it have?

**[00:16:48]** I think in this regard,

**[00:16:50]** large language models

**[00:16:52]** play the role of X-rays,

**[00:16:54]** showing us

**[00:16:57]** what we had inside our work.

**[00:17:00]** For the first time, we fully separate

**[00:17:02]** the layer of the product

**[00:17:04]** from our work

**[00:17:06]** and from the relationship of how

**[00:17:08]** we were shaped by that work.

**[00:17:10]** And because of that, for the first time,

**[00:17:12]** we can clearly see

**[00:17:14]** what institutions do.

**[00:17:16]** And it doesn't matter at all,

**[00:17:18]** how we talk about it in a moment,

**[00:17:20]** whether you talk about school or a company.

**[00:17:22]** It doesn't matter at all.

**[00:17:24]** Because both institutions

**[00:17:26]** not only produced outputs,

**[00:17:29]** but shaped people.

**[00:17:32]** What we have in those tremors,

**[00:17:35]** when even in the mid-60s someone

**[00:17:38]** internally claims that,

**[00:17:40]** I sell these shoes well,

**[00:17:42]** because I started at Baťa.

**[00:17:45]** That's exactly the naming of

**[00:17:47]** what happened to him.

**[00:17:49]** That person sells shoes well,

**[00:17:51]** you can measure that.

**[00:17:53]** But the whole formation of why

**[00:17:55]** that person does it well

**[00:17:57]** was hidden.

**[00:17:59]** I don't hope everyone has fallen asleep,

**[00:18:01]** but let's move on.

**[00:18:03]** I'm terribly boring today.

**[00:18:06]** I think this place

**[00:18:08]** is starting to be sensed today.

**[00:18:10]** We have a lot of research on it,

**[00:18:12]** I'll talk about some of it.

**[00:18:14]** I just feel it leads to the world splitting into two groups again.

**[00:18:16]** Into the techno-optimists who tell you,

**[00:18:18]** and that's great, so we won't do boring work,

**[00:18:20]** that's awesome, we don't care.

**[00:18:22]** And the others who say,

**[00:18:24]** oh my God, everything is doomed,

**[00:18:26]** people will become complete dumb zombies,

**[00:18:28]** walking around the world.

**[00:18:30]** I don't believe either one.

**[00:18:32]** I don't think it's a disaster.

**[00:18:35]** I think it's a discovery.

**[00:18:38]** And when something is visible,

**[00:18:42]** you can take care of it.

**[00:18:44]** You can't take care of invisible things.

**[00:18:46]** Visible things, yes.

**[00:18:48]** And if you know they're there,

**[00:18:50]** let's see

**[00:18:52]** what we can do with them,

**[00:18:54]** how to take care of them.

**[00:18:56]** I think the effect we're experiencing

**[00:18:58]** could best be called sterilization.

**[00:19:04]** Yeah?

**[00:19:06]** Because AI doesn't replace abilities.

**[00:19:10]** It reveals the conditions,

**[00:19:12]** from which those abilities arise.

**[00:19:14]** From what circumstances our abilities emerge.

**[00:19:18]** And what it does to them is,

**[00:19:20]** it sterilizes them,

**[00:19:22]** like we sterilize soil.

**[00:19:24]** Yeah?

**[00:19:26]** The soil remains there.

**[00:19:28]** Nothing grows on it.

**[00:19:30]** The company remains.

**[00:19:32]** But no one grows in it.

**[00:19:34]** The school remains.

**[00:19:36]** But nothing grows there.

**[00:19:38]** The institutions remain.

**[00:19:40]** The mechanisms keep working.

**[00:19:42]** But nothing will grow there.

**[00:19:44]** That's why I choose the term sterilization.

**[00:19:46]** I don't choose the term de-skilling or no-skilling.

**[00:19:50]** Because we are at some stage now.

**[00:19:53]** And I claim that we are sterilizing the space,

**[00:19:55]** where our abilities arise.

**[00:19:57]** And we have to do something about it.

**[00:20:00]** If we wanted to find some theoretical support for what I say,

**[00:20:09]** we could go many ways.

**[00:20:16]** I chose an example from psychoanalyst Donald Winnicott.

**[00:20:22]** Donald Winnicott was a British child psychoanalyst in the 1950s,

**[00:20:27]** who, among other things, came up with the concept of the good enough mother.

**[00:20:34]** By the way, if you want to have fun and learn at the same time,

**[00:20:38]** which doesn't happen often, unless you're watching a TED Talk,

**[00:20:46]** check out how psychoanalysts today are starting to think about AI.

**[00:20:50]** It's incredibly interesting reading.

**[00:20:52]** Today psychoanalysts are beginning to think about AI

**[00:20:55]** and are starting to notice the relationship between humans and AI

**[00:20:58]** from the perspective of psychoanalysis.

**[00:21:00]** One of the strongest insights today is

**[00:21:03]** that we treat AI like an all-powerful mother,

**[00:21:06]** to whom we turn when we have a problem, hoping it will solve it for us.

**[00:21:11]** That's extremely interesting reading.

**[00:21:13]** Amy Levy recently published a book called The New Other,

**[00:21:16]** three months ago.

**[00:21:18]** But back to Donald Winnicott.

**[00:21:20]** Donald Winnicott knows nothing about AI.

**[00:21:23]** Donald Winnicott studies how a child develops.

**[00:21:28]** He says a child doesn't need a perfect mother,

**[00:21:33]** who is always present to come to the child's aid

**[00:21:38]** and solve everything for them.

**[00:21:41]** A child needs a good enough mother,

**[00:21:44]** who can create a gap by not immediately fulfilling all the child's wishes.

**[00:21:48]** That gap, that space, because the child experiences

**[00:21:51]** some kind of discomfort or frustration,

**[00:21:55]** the child creates itself.

**[00:21:59]** Because at that moment, the child begins to form as itself.

**[00:22:01]** And in that gap, it starts doing something on its own.

**[00:22:07]** To paraphrase freely.

**[00:22:13]** Yes, if the object adapts perfectly, it becomes a hallucination.

**[00:22:16]** It stops being real.

**[00:22:20]** We all know that.

**[00:22:22]** These are all those moments in movies.

**[00:22:24]** It's all too good to be true.

**[00:22:27]** Everything is just as it should be, so it can't be true.

**[00:22:33]** That's the mechanism. We all know it well.

**[00:22:39]** When something is exactly as it should be, and that's strange.

**[00:22:42]** There is no gap.

**[00:22:47]** We are surrounded by a perfect world, adapted to us.

**[00:22:49]** The mechanism that makes a mother good enough

**[00:22:55]** doesn't start by giving the child all the toys,

**[00:23:02]** locking them in an empty room and saying, play.

**[00:23:07]** That would be nonsense.

**[00:23:10]** It would create a depressed child rather than one

**[00:23:12]** with great abilities.

**[00:23:15]** It starts by gradually easing off.

**[00:23:19]** And as the child's capacity grows,

**[00:23:22]** the mother eases off in what she is willing to do for the child.

**[00:23:26]** Now go spread the bread yourself.

**[00:23:31]** Why am I talking about this?

**[00:23:35]** The point is, not every resistance you give the child is good.

**[00:23:37]** Not every resistance leads to good development.

**[00:23:47]** Some are harsh. Some are completely nonsensical.

**[00:23:51]** It's not like we start denying kids everything

**[00:23:55]** and they become better.

**[00:23:57]** It's about the fact that denial plays a role.

**[00:24:00]** And that this denial needs to be somehow appropriate,

**[00:24:04]** somehow calibrated.

**[00:24:06]** And that it happens with someone who is beside you.

**[00:24:09]** Someone who cares about you.

**[00:24:11]** Well, I say AI is too good a mother.

**[00:24:15]** AI is too good a mother.

**[00:24:18]** It delivers immediately, completely without gaps.

**[00:24:22]** You can tell many stories about how it hallucinates.

**[00:24:26]** And let's also tell stories about how often we don't notice it,

**[00:24:30]** just like the child doesn't notice it.

**[00:24:33]** And let's tell stories about how where it used to deliver nonsense prematurely,

**[00:24:37]** it delivers well today.

**[00:24:39]** The principle remains psychologically the same.

**[00:24:43]** The output is delivered perfectly.

**[00:24:45]** No resistance, no space where a person has to do something themselves.

**[00:24:51]** If you read your XK linkets, it's full of that,

**[00:24:56]** I didn't have to do anything and that's awesome.

**[00:25:00]** I was lazy, I didn't even want to take the photo,

**[00:25:04]** so I just let the app generate it.

**[00:25:08]** Yeah, we talk about this as something we want.

**[00:25:13]** I'm telling you, we want our own downfall.

**[00:25:17]** Because a person doesn't move forward that way.

**[00:25:20]** If we look at how it is right now,

**[00:25:23]** we were looking for something like third-degree sterilization.

**[00:25:27]** I know I’m running back and forth in the footage,

**[00:25:30]** sorry everyone for my shiny head.

**[00:25:33]** I told myself it’s terrible to walk.

**[00:25:35]** For me, this is like having a person tied up completely.

**[00:25:39]** I walk well backwards like this, otherwise I’d march from left to right.

**[00:25:43]** But those were girls in Cheb and elsewhere, who send their regards.

**[00:25:49]** They saw nothing.

**[00:25:51]** One of the experiments done earlier

**[00:25:54]** was that about a thousand high school students were taken,

**[00:25:58]** divided into three groups.

**[00:26:01]** One group of those students got full access to GPT

**[00:26:04]** and could use it to prepare for the exam.

**[00:26:08]** During that time, it was calculated,

**[00:26:10]** how the ongoing preparatory tests were performing.

**[00:26:13]** They were 48% cheaper than in the other groups.

**[00:26:21]** The delivered product, the intermediate state, was significantly better.

**[00:26:27]** And then when the final tests came,

**[00:26:31]** where they didn't have the AI available,

**[00:26:34]** their results were 17% worse than the other two groups.

**[00:26:40]** And the most common message they used and sent to the chat was,

**[00:26:45]** 'What is the answer?' and the question from the test.

**[00:26:50]** What's interesting about that?

**[00:26:52]** They themselves didn't think they had learned less.

**[00:26:56]** They were convinced they had learned well.

**[00:27:01]** That means, on one hand, they continuously deliver better results,

**[00:27:06]** but when faced with reality, they fail

**[00:27:08]** and don't think they haven't learned anything.

**[00:27:11]** A perfect cocktail.

**[00:27:14]** The second group, the control group, was the group

**[00:27:19]** that had no AI available.

**[00:27:23]** We can't say whether they learned better or worse,

**[00:27:26]** that's the control group.

**[00:27:28]** In effectiveness, we count how these people performed.

**[00:27:31]** And then there was a third group.

**[00:27:33]** This third group was given a modified bot

**[00:27:38]** so that it wouldn't give them answers during the effectiveness test,

**[00:27:40]** but would guide and help them calculate it themselves.

**[00:27:43]** They got a guide.

**[00:27:45]** Someone who asked them questions.

**[00:27:48]** And here there was no negative effect on effectiveness.

**[00:27:52]** These people performed just as well as the group without AI.

**[00:27:55]** They just enjoyed it more.

**[00:27:59]** That means, what's hidden here,

**[00:28:01]** the AI wasn't very good.

**[00:28:03]** It was good enough.

**[00:28:05]** It helped them build the same skills.

**[00:28:08]** And there's another hidden thing here.

**[00:28:10]** The question isn't whether children should use AI.

**[00:28:13]** Whether people should use AI.

**[00:28:15]** The question is how to use it.

**[00:28:17]** Not whether.

**[00:28:19]** Because we know from this experiment, and many others,

**[00:28:22]** because we know that how matters.

**[00:28:25]** Not whether.

**[00:28:29]** The second problem.

**[00:28:31]** Juniors in companies.

**[00:28:33]** As I estimate, there aren't many juniors here.

**[00:28:36]** But I hope you are online.

**[00:28:39]** That would make me very happy.

**[00:28:41]** And I hope your bosses are with you,

**[00:28:44]** because they would make me happy if they listened to the whole thing.

**[00:28:49]** Earlier, I was having coffee with Martin Straka from MatFiz

**[00:28:54]** and we talked about how to teach students today.

**[00:28:58]** And what we came to is a very interesting thing.

**[00:29:02]** I say that more and more I do this,

**[00:29:05]** I present them with problems that have no solution.

**[00:29:09]** Or rather, no correct solutions.

**[00:29:11]** I try more and more to focus on the process

**[00:29:14]** of reflection, of solving.

**[00:29:16]** He says, but I do exact science.

**[00:29:19]** What I assign them has a solution.

**[00:29:22]** So I can't give them things that have no solution.

**[00:29:25]** He says, well, that's bad.

**[00:29:27]** He says, so they take AI, wipe code,

**[00:29:30]** and get to the fifth floor in a week.

**[00:29:36]** There will be given code.

**[00:29:38]** And then they fail.

**[00:29:42]** But they don't know where they failed.

**[00:29:44]** Because they never understood the previous floors.

**[00:29:47]** They don't even know why it doesn't work.

**[00:29:50]** So they keep banging at it until they somehow fix it.

**[00:29:54]** And that's the problem with beginners.

**[00:29:56]** A beginner, a junior, is a person who...

**[00:30:00]** who doesn't know what they don't know.

**[00:30:06]** Yes, it's not Socratic, who knows that he doesn't know.

**[00:30:10]** No, a junior is someone who doesn't know what they don't know.

**[00:30:14]** At the very bottom is a person who just follows rules,

**[00:30:18]** for whom it's more important to follow the rules than the product itself.

**[00:30:22]** And then they gradually grow.

**[00:30:24]** The problem is when a junior produces outputs like an expert,

**[00:30:30]** but can't repeat it, can't vary it, can't assess it,

**[00:30:35]** then they don't know.

**[00:30:38]** And people don't know much about it.

**[00:30:40]** Jessica Edinger offered $100 to people who passed a certain test,

**[00:30:46]** to objectively evaluate their performance before

**[00:30:50]** they received the test results.

**[00:30:53]** That's a nice task.

**[00:30:55]** How confident are you about it?

**[00:31:01]** The weakest in those tests remained convinced they were good.

**[00:31:06]** Even though they could get $100 if they admitted they weren't,

**[00:31:10]** it doesn't work.

**[00:31:12]** The second big problem is that we sterilize beginners,

**[00:31:15]** because the skills you need to perform,

**[00:31:18]** you also need to assess that performance.

**[00:31:21]** And sure, we will be editors of those texts.

**[00:31:25]** We will supervise the code.

**[00:31:27]** Today, when I talk to senior programmers,

**[00:31:29]** they say it's completely stupid.

**[00:31:31]** I just sit there and tell them what's wrong.

**[00:31:35]** I tell them what's wrong.

**[00:31:37]** I know what's wrong.

**[00:31:39]** So I can tell them.

**[00:31:41]** Not that I have a console output saying it's wrong.

**[00:31:45]** I use them extensively for research,

**[00:31:48]** I use them extensively for lecture proposals,

**[00:31:51]** because I know what's wrong.

**[00:31:53]** But when someone who doesn't know what's wrong approaches it,

**[00:31:57]** it doesn't end well.

**[00:32:00]** So we tell those people,

**[00:32:03]** and I've said it many times in lectures,

**[00:32:06]** you have to be critical of the output.

**[00:32:10]** You're probably at the same level as a person who can't see colors,

**[00:32:13]** who is colorblind, and you tell them they must distinguish red from green.

**[00:32:18]** If you don't have that ability, you can write it down in a notebook a hundred times,

**[00:32:22]** ask yourself the question a hundred times.

**[00:32:25]** You don't have it here.

**[00:32:27]** Colorblindness can't be cured by training, even if it's known.

**[00:32:32]** And that's what critical thinking is somewhat about.

**[00:32:36]** The bad thing is that this sterilization happens even to experts.

**[00:32:42]** That's bad.

**[00:32:45]** If you dive into the history of theories related to juniors, experts,

**[00:32:50]** and the emergence of expertise, which I did in lectures at the Faculty of Education,

**[00:32:56]** where we specifically study problems like negative oblivion or AI risks in education,

**[00:33:02]** you'll find that the issue is much deeper here.

**[00:33:07]** The model I briefly quoted about how a beginner starts by

**[00:33:12]** just following the rules, and the expert ends up

**[00:33:15]** not remembering any rules because they simply have the flow.

**[00:33:18]** And the only moment they start thinking is when they fail.

**[00:33:21]** This was commissioned by the US military as research in the early 80s,

**[00:33:25]** when they began considering AI integration into aircraft.

**[00:33:28]** Because even back then, in the early 80s, they sensed there would be problems.

**[00:33:34]** Let's stay with aircraft here.

**[00:33:37]** The problem is that researchers studying pilots and air traffic control

**[00:33:44]** know that when a pilot is doing one thing and automation fails, they notice it.

**[00:33:51]** When doing multiple things at once and automation fails,

**[00:33:55]** they notice it only in a third of cases.

**[00:33:58]** Because their attention is elsewhere.

**[00:34:02]** And interestingly, even experienced pilots weren't

**[00:34:08]** much better off; experienced pilots were basically as bad as

**[00:34:13]** the novices.

**[00:34:15]** When you had to diagnose a system where the observer at once

**[00:34:19]** noticed something was wrong.

**[00:34:21]** Today we might say this is a paradise for people with ADHD.

**[00:34:25]** I understand those sentences.

**[00:34:27]** I would even sign off on them.

**[00:34:31]** But it's just not good news.

**[00:34:37]** Interestingly, in one of those experiments, they forced

**[00:34:41]** pilots to observe things attentively.

**[00:34:45]** They still didn't notice.

**[00:34:47]** The attention simply dropped anyway.

**[00:34:50]** That means the positive feedback loop we get now from AI,

**[00:34:55]** AI works correctly, you don't check, AI works again,

**[00:34:58]** you check even less.

**[00:35:00]** This applies to experts as well.

**[00:35:04]** Even experts have this problem.

**[00:35:07]** We can't tell a junior to check AI; they don't have the means.

**[00:35:11]** And you can't tell a senior to be vigilant,

**[00:35:15]** because vigilance isn't a matter of willpower.

**[00:35:18]** Vigilance arises from other internal movements and motivations.

**[00:35:24]** And we don't see that much in institutions.

**[00:35:29]** I think one principle applies here.

**[00:35:31]** Where the object adapts too perfectly,

**[00:35:34]** Formation stops, that's the philosophical definition,

**[00:35:37]** or almost psychoanalytical.

**[00:35:39]** Where we have an environment too adaptable, that doesn't resist us,

**[00:35:43]** we stop forming ourselves.

**[00:35:45]** Right now I'm reading, last week with students during a seminar

**[00:35:51]** on the imagination of technology in literature,

**[00:35:53]** because it turned out that it's actually something

**[00:35:55]** students don't do much anymore,

**[00:35:57]** like reading beautiful literature such as Frankenstein and those things.

**[00:36:01]** Today we read Bradbury, Fahrenheit 451.

**[00:36:05]** It's not a book about censorship at all.

**[00:36:07]** It's a book about how formation stops in people,

**[00:36:11]** who have no frustration,

**[00:36:13]** because everything around them perfectly suits them.

**[00:36:17]** Ok, time has come.

**[00:36:21]** You were right, it will be just enough.

**[00:36:25]** One reason I'm leaning like this is

**[00:36:28]** because I need to watch how many slides I have left.

**[00:36:31]** Watch the time on the right like that.

**[00:36:33]** He is right for me.

**[00:36:35]** Ok, let's go back to Vinkot.

**[00:36:39]** Vinkot distinguishes between playing and game.

**[00:36:46]** Playing is a free, creative, somewhat risky activity.

**[00:36:51]** Playing in the sand, chatting in the garden, painting.

**[00:36:57]** There is no score in playing.

**[00:36:59]** There is no level you have to reach.

**[00:37:02]** You can stop anytime and nothing happens.

**[00:37:06]** But in playing you also encounter resistance.

**[00:37:10]** Something doesn't adapt to you, someone doesn't adapt to you.

**[00:37:14]** You have to go into some conflict with the material or with a person,

**[00:37:19]** where the reward is not progressing to the next level,

**[00:37:24]** or a strike, or booty.

**[00:37:27]** That's all the domain of games.

**[00:37:30]** In games, the child is not playing, they are played.

**[00:37:34]** And basically organized games are a defense against play itself.

**[00:37:43]** The whole system like evaluation, whether based on KPIs,

**[00:37:49]** or grades, or filling out multiple-choice tests,

**[00:37:53]** or passing the right maturities, are games.

**[00:37:59]** So in reality, children are played in them,

**[00:38:03]** to fill out the test well, to do well.

**[00:38:07]** To meet your KPI, which will secure you a 10% bonus on your salary.

**[00:38:13]** It actually shapes nothing.

**[00:38:15]** Or rather, you don't shape it there, you are shaped by that mechanism.

**[00:38:22]** Honza originally said I should have people on social networks.

**[00:38:25]** I said I wouldn't do it,

**[00:38:27]** but I put one slide there about social networks for you.

**[00:38:31]** Because with many of these things, we have the feeling

**[00:38:34]** that what we do is activity.

**[00:38:37]** But when I study here on social tests, I have activity.

**[00:38:45]** I like Byung-Chul Han,

**[00:38:48]** who talks about endless scrolling on social networks.

**[00:38:53]** He says it looks like activity,

**[00:38:56]** but in reality, it's actually hyperactive passivity.

**[00:39:00]** In it, you just lose the ability to do nothing,

**[00:39:06]** because every free moment you have, you fill by starting to scroll.

**[00:39:14]** Where does a person actually form themselves?

**[00:39:17]** What I say is very old school.

**[00:39:22]** There is much more theory on this today,

**[00:39:24]** but I like the old school approach,

**[00:39:27]** which I adapted a bit to the topic of this lecture.

**[00:39:31]** Because you know it.

**[00:39:33]** The first space where you form yourself is family.

**[00:39:37]** In the family, the most basic things arise.

**[00:39:40]** Trust, the ability to endure frustration, the ability to be in a relationship.

**[00:39:44]** The problem is, we know we don't have time for the kids.

**[00:39:48]** The problem is, we try to fill their time.

**[00:39:54]** And the worst part is, we don't do it

**[00:40:00]** because we don't love them.

**[00:40:02]** We do it because we love them, right?

**[00:40:05]** That's why we cram them into cars, drive them to activities,

**[00:40:07]** then pick them up and take them to another activity,

**[00:40:09]** and in the evening, we check their German homework with them.

**[00:40:15]** In reality, we are actually harming them.

**[00:40:18]** The second space is school.

**[00:40:20]** School teaches us rules, theories, frameworks,

**[00:40:23]** it teaches us social interactions.

**[00:40:25]** Honestly, school teaches you something.

**[00:40:29]** But the most important thing you take from it

**[00:40:32]** is the ability to be with other people.

**[00:40:35]** The socializing function of school is much stronger

**[00:40:39]** and it will be much stronger in the future,

**[00:40:41]** than this one, or if you don't get it,

**[00:40:43]** kids will learn at home with AI.

**[00:40:47]** Well, they'll become good monsters.

**[00:40:53]** The next space is work.

**[00:40:55]** We spend a huge amount of time at work.

**[00:40:57]** Work is the space that shapes us the most in later life.

**[00:41:02]** At work, we should have mentors in our hands,

**[00:41:05]** at work we should have an older colleague,

**[00:41:07]** who will discuss things with us,

**[00:41:09]** or at least yell at us for doing things wrong.

**[00:41:14]** Sometimes we have a kitchenette where we say,

**[00:41:16]** how everyone else is doing things wrong.

**[00:41:18]** But that's not a very good formative space for us.

**[00:41:22]** And then there is somewhere else.

**[00:41:24]** That somewhere else is a space we actually know very little about.

**[00:41:30]** Elsewhere is the space called,

**[00:41:39]** just being with yourself.

**[00:41:42]** And I say that even here, in that free space, we can prepare ourselves,

**[00:41:47]** simply by filling it up.

**[00:41:51]** With TikTok, Twitter, some smarter TEDx talk.

**[00:41:57]** I'll go for a beer with colleagues from work, that's more efficient,

**[00:42:00]** we'll discuss those things there.

**[00:42:03]** In fact, we destroy that space again.

**[00:42:08]** And I say that actually all these spaces

**[00:42:13]** are somehow collapsing right now.

**[00:42:18]** They probably started collapsing long before the AI wave,

**[00:42:22]** but AI just gave it the right spin.

**[00:42:26]** So what now? My suggestion.

**[00:42:31]** Protect the gap.

**[00:42:38]** We form ourselves in the space where we do nothing.

**[00:42:42]** We form ourselves in the space where we are frustrated by something

**[00:42:46]** and have to think about what to do about it.

**[00:42:49]** We form ourselves in a space that from an outside perspective

**[00:42:53]** looks extremely inefficient.

**[00:42:56]** That is the space where we form ourselves.

**[00:42:59]** And we need to protect that space.

**[00:43:02]** Don't just give a child a tablet and let them get bored.

**[00:43:08]** And boredom is not a problem, boredom is space.

**[00:43:13]** And when a child learns to be bored, they learn to invent.

**[00:43:17]** Yeah, they'll come up with a lot of nonsense.

**[00:43:20]** Maybe they'll even put sugar in the neighbor's motorcycle tank.

**[00:43:24]** But they'll start doing something with themselves.

**[00:43:30]** The second important thing is presence before optimization.

**[00:43:34]** That means the most important investment in those moments

**[00:43:38]** is not quality time after the activity.

**[00:43:41]** It's presence.

**[00:43:42]** I say this from the perspective of someone who,

**[00:43:44]** in their parenting, missed many opportunities

**[00:43:46]** to be present with the kids instead of

**[00:43:48]** dragging them to another museum.

**[00:43:53]** Playing, not games.

**[00:43:55]** Do you know in advance how it will end?

**[00:43:58]** Is there a score, can you stop anytime, or not?

**[00:44:02]** Can you play something on the phone?

**[00:44:07]** Or on the computer?

**[00:44:09]** Isn't everything games?

**[00:44:12]** Minecraft?

**[00:44:14]** It's much more like playing than games.

**[00:44:17]** Some drawing?

**[00:44:18]** Crash Band?

**[00:44:20]** Did they start creating?

**[00:44:22]** Those are playing, not games.

**[00:44:26]** Well, what about you personally, like...

**[00:44:30]** I'll be quicker, as I promised.

**[00:44:35]** If you want to protect the gap,

**[00:44:38]** maybe just sit down and read that article,

**[00:44:41]** instead of scrolling.

**[00:44:43]** Or go for a walk without playing a podcast.

**[00:44:48]** Or sit in a café without your phone.

**[00:44:51]** Not because it would be hipsterish,

**[00:44:54]** but because at that moment you learn

**[00:44:57]** the ability to be in that gap.

**[00:44:59]** The second principle: first yourself, then AI.

**[00:45:03]** Here we all have a big advantage.

**[00:45:06]** Here we all have a head start.

**[00:45:09]** Because we were born in a time without AI,

**[00:45:13]** we have many skills

**[00:45:17]** developed beforehand.

**[00:45:20]** By the way, that's why we seniors

**[00:45:22]** are often more successful with AI than juniors.

**[00:45:26]** Not because juniors are dumb,

**[00:45:28]** but because we are old.

**[00:45:32]** We don't need more frustration,

**[00:45:34]** we need the right kind of frustration.

**[00:45:36]** There is a dull frustration from the work.

**[00:45:39]** Repetitive strain, doing the same thing over and over,

**[00:45:42]** it doesn't excite me, it numbs me.

**[00:45:44]** And then there is something we can call dialogic frustration.

**[00:45:48]** That's the moment when we try, hit obstacles, think,

**[00:45:51]** try something different.

**[00:45:53]** A farmer encountering some material

**[00:45:56]** experiences a moment when resistance provokes thinking.

**[00:46:01]** AI can remove both.

**[00:46:06]** Both the dull frustration and the dialogic frustration.

**[00:46:10]** But we should try to remove the dull frustration.

**[00:46:16]** That means, OK, a child solves a problem on their own,

**[00:46:19]** then inputs it to AI and compares the results.

**[00:46:22]** Where do we differ? Where am I better? Where is AI better?

**[00:46:25]** Where does AI make a mistake I wouldn't?

**[00:46:29]** Notice these questions aren't about the solution at all,

**[00:46:33]** but about who I am, who AI is, what I can do, what it can't.

**[00:46:41]** From that comparison, from that clash,

**[00:46:44]** an orientation map of competence emerges.

**[00:46:50]** And that's exactly what we need to develop.

**[00:46:54]** The same applies to us.

**[00:46:57]** Write an email, write it yourself first, then compare.

**[00:47:04]** Not because it's faster, but because one of the biggest problems

**[00:47:09]** we're starting to see in this kind of delegation,

**[00:47:13]** is that we start writing personal, intimate communication to AI,

**[00:47:17]** apologies to partners, both personal and business.

**[00:47:26]** We outsource reactions, we get dull through AI.

**[00:47:30]** Please, can you rephrase this for me to be less harsh?

**[00:47:35]** It's bad, because if we do it like this,

**[00:47:38]** we won't learn to work with ourselves,

**[00:47:41]** nor will we understand ourselves.

**[00:47:43]** And AI will do it brilliantly for us.

**[00:47:46]** I say this as someone who uses it this way too.

**[00:47:49]** Like when I know my emotions are being shaken by anger,

**[00:47:53]** So usually I write the email the way I would,

**[00:47:57]** and then I give it to AI, asking if it would be so kind,

**[00:48:00]** to rewrite it while keeping the intent,

**[00:48:03]** but using different language.

**[00:48:06]** So I’m not saying this is how a person would always do it.

**[00:48:09]** But it’s important because it gives us some knowledge about ourselves.

**[00:48:13]** The last thing sounds a bit strange: look for a container curriculum.

**[00:48:17]** The key is not information, the key is relationship.

**[00:48:21]** And the most important question when choosing a school, in my opinion, is not,

**[00:48:26]** what AI tools they use, what courses they offer their students,

**[00:48:34]** how many students they accept into certain fields,

**[00:48:39]** whether you have A, B, or none.

**[00:48:42]** The most important question is, will the child feel good there?

**[00:48:49]** And will it be a good environment where they let the child fail productively

**[00:48:54]** and help them grow? That’s the question.

**[00:48:59]** In reality, will there be someone who cares about that child

**[00:49:06]** and who gives them space to struggle?

**[00:49:10]** It sounds very strange. But this is the key.

**[00:49:16]** Many people remember their universities as places

**[00:49:20]** where no one cared. And then those universities wonder,

**[00:49:25]** why their graduates don’t care about them.

**[00:49:29]** Many of those schools push you toward performance.

**[00:49:34]** But I’ve seen many schools that can contain

**[00:49:38]** the struggles and frustrations of kids who start below average,

**[00:49:42]** and lift them to average or slightly above average.

**[00:49:46]** This is the kind of school you want. You don’t want a competitive race like rabbits.

**[00:49:50]** You want someone who can embrace those kids,

**[00:49:54]** give them that container, understand their struggles,

**[00:49:58]** and not solve it for them.

**[00:50:00]** Not solve it for them. Let them struggle in a way that helps them.

**[00:50:08]** And we actually need to set the same thing in workplaces.

**[00:50:12]** We need to work more with the idea of containing,

**[00:50:16]** the mood in companies.

**[00:50:19]** So the head chef still yells at you,

**[00:50:22]** but sits next to you and when you tell him something,

**[00:50:27]** he might say you’re fascinated,

**[00:50:30]** but he embraces you and helps you move forward,

**[00:50:34]** that’s the person you want in the company.

**[00:50:39]** Right? Because otherwise you won’t get seniors from juniors.

**[00:50:44]** Otherwise, you won’t get seniors from juniors.

**[00:50:48]** Someone who accepts your frustration, holds it, and helps you return it

**[00:50:54]** back in a form you can process.

**[00:50:58]** By the way, there are remarkable studies in HR

**[00:51:02]** that focus on this problem.

**[00:51:05]** We all know managers who come in with a rotation principle in the company.

**[00:51:09]** Usually, this is a person who lasts about two years in each company.

**[00:51:12]** They replace another manager who messed something up, come with an idea,

**[00:51:16]** mess it up again, and move on.

**[00:51:20]** Judging by the reactions here in the audience, we all know these managers.

**[00:51:24]** They are very successful because what they offer to top management,

**[00:51:28]** not the owners—I think if they talked to owners, they wouldn’t be hired—

**[00:51:32]** is performance.

**[00:51:36]** Those numbers, 'let’s increase everything,' they leave the company completely devastated,

**[00:51:40]** because they were never able to contain their employees,

**[00:51:44]** understand what matters to them, understand their frustrations, and work with them.

**[00:51:49]** Companies that cut juniors.

**[00:51:52]** This is something we really don’t need today. It’s expensive.

**[00:51:56]** But we need to stop cutting juniors.

**[00:52:00]** We need all systems to avoid cutting juniors.

**[00:52:04]** Because if we keep cutting juniors, then those of us sitting here

**[00:52:08]** will have a very late retirement.

**[00:52:12]** AI can be a great tool within some containing relationship.

**[00:52:20]** But it cannot replace that relationship.

**[00:52:24]** It can’t hug you.

**[00:52:28]** It can’t accept you and give it back to you.

**[00:52:34]** No better prompt can do that.

**[00:52:39]** And I think I’ve written a few good prompts and seen some good ones.

**[00:52:44]** But that’s not the question, because your formation as a person happens in the relationship.

**[00:52:52]** Not in the conversation.

**[00:52:54]** Conclusion. Let’s reframe this anxiety.

**[00:53:00]** I primarily think that the field matters much less than you think.

**[00:53:10]** Why?

**[00:53:12]** Because AI will change all fields completely.

**[00:53:16]** And it will change them directly, especially for symbolic analysts, where we started,

**[00:53:21]** it will change them directly by what it does.

**[00:53:23]** But right now it’s starting to accelerate research and advances in other areas.

**[00:53:31]** Indirectly, by accelerating research.

**[00:53:34]** That means, I don’t know if in 15 years the most progressive job will be a barber.

**[00:53:42]** But basically, I think automating a barber’s work is possible.

**[00:53:48]** Probably about the same level as automating a union programmer’s work.

**[00:53:53]** It will require greater trust from you towards mechanical hands, but in principle, it is possible.

**[00:54:00]** I think what’s important is to focus on

**[00:54:07]** how we are changing and how we want to develop.

**[00:54:11]** And to focus on how to develop our skills and abilities in this AI era.

**[00:54:19]** That means strengthening those areas so we can shape ourselves

**[00:54:23]** and think deeply about how to develop certain traits in different ways.

**[00:54:32]** I don’t know how, but I believe it’s possible.

**[00:54:35]** Once we know and accept it, we can somehow work with it.

**[00:54:46]** It’s a bit strange because what I’m telling you is

**[00:54:51]** that what we considered necessary but a necessary evil in companies, schools, and so on,

**[00:55:00]** the fact that we received information and were taught something,

**[00:55:08]** turns out to be the least important thing in those companies and schools for us.

**[00:55:15]** The most important thing is that these are places where we form ourselves as people.

**[00:55:21]** And actually, we should even send a thank you letter because it showed us

**[00:55:29]** that routine, boredom, frustration, and gaps

**[00:55:39]** mean much more to us than we think, because if we lose them,

**[00:55:45]** if we get an overly good mother, we also stop being ourselves.

**[00:55:51]** I believe sterilization is not fate. I think it’s the result of a decision.

**[00:56:00]** And that decision is what we make. That means we can decide

**[00:56:05]** not to go that way or to go differently. We have free will in that.

**[00:56:11]** I think what we need to protect is resistance.

**[00:56:18]** Or what we need to ensure is the presence of resistance,

**[00:56:24]** the safety of someone beside you, presence, and some time you have for it.

**[00:56:31]** None of those tools will give you that, but none will take it away either.

**[00:56:38]** Three words: gap, sequence, container.

**[00:56:43]** Protect the space where nothing happens. First do things yourself, then with the tool.

**[00:56:49]** And always try to have someone beside you who can help.

**[00:57:00]** Because what we ultimately need to protect is the space for playing.

**[00:57:05]** Because in the end, like in a game, we are the real people.

**[00:57:10]** And in the end, like in a game, we are free and happy people.

**[00:57:17]** Thank you for your attention.

**[00:57:20]** Applause

**[00:57:33]** Four minutes left.

**[00:57:36]** I made it.

**[00:57:37]** You made it? I think you opened it so well that huge amounts of questions started coming in from the beginning.

**[00:57:46]** So just a reminder, if you want to ask something, take your phone and definitely write.

**[00:57:52]** When I was talking, I discovered him and greet him remotely and hope he comes to Digestive,

**[00:57:57]** because I wrote this to him: Filip Dřímalka.

**[00:57:59]** Filip is the kind of person who pushes a lot and the economy of creators and so on.

**[00:58:05]** And I just wrote to him, hey, try to take a look at it.

**[00:58:08]** And I'm just thinking that the argument, and you introduced it nicely at the start,

**[00:58:17]** that when I work with artificial intelligence, I am also shaping myself.

**[00:58:22]** I am also progressing.

**[00:58:24]** I'm just progressing differently, only that I'm replacing a book with this.

**[00:58:29]** I know you talked about it, but try to touch on this a little more.

**[00:58:33]** What I meant is that every activity shapes us.

**[00:58:36]** It's not that AI doesn't shape us.

**[00:58:41]** But AI shapes us in a way that can sometimes lead us

**[00:58:47]** into unconsciousness about what we are actually doing.

**[00:58:51]** That's the process where the new skill doesn't develop.

**[00:58:56]** And other skills grow instead.

**[00:59:00]** But those skills might not fit with the kind of people we are.

**[00:59:06]** If you give me a very concrete example,

**[00:59:12]** the more we move towards the role of editors.

**[00:59:16]** In the sense that we get output from AI and edit it.

**[00:59:20]** We do some vape coding and know where the error is.

**[00:59:25]** But we have the ability beforehand, because we've read a lot before,

**[00:59:30]** programmed a lot, to recognize when AI is pulling the wool over our eyes

**[00:59:35]** or leading us somewhere completely different.

**[00:59:38]** It can happen very quickly that you start doing something with AI,

**[00:59:42]** you do some vape coding and it starts suggesting things and you say,

**[00:59:46]** that's great. Honza Kulovej had such a beautiful example about this,

**[00:59:50]** which represents you.

**[00:59:52]** That's very nice, I was really glad to see you yesterday.

**[00:59:55]** And now there's that autocomplete that says,

**[00:59:58]** it would be good to see you again.

**[01:00:00]** So you hit TAP and it appears like this, and suddenly you go on a date with someone

**[01:00:05]** you didn't want to go on a date with, and you go on a date with someone you didn't want to.

**[01:00:09]** Because you didn't resist. You didn't have that resistance there.

**[01:00:13]** I think what's extremely important, and I'm not the only one who thinks so,

**[01:00:17]** psychoanalyst Brenner returns to this, if we don't have the ability to resist

**[01:00:22]** and say I don't want this, we will gradually become more like extended arms of AI

**[01:00:29]** rather than those who actively do something with it. And it's easy for us to say,

**[01:00:34]** because we have the experience. Most of us are already shaped.

**[01:00:40]** We can reshape. But I'm talking to kids who are 15, 16.

**[01:00:49]** Interestingly, some of them undoubtedly see this. When I teach at a gymnasium

**[01:00:54]** about the memory of the nation and AI, I thought I want to see those gymnasium students,

**[01:00:58]** a large portion of those people don't use AI because they are afraid

**[01:01:04]** it will take away their creativity. That's one important thing.

**[01:01:08]** Because they fear losing their creativity. And the other part is,

**[01:01:13]** They don't like to see educators using it because they're afraid,

**[01:01:18]** that they will lose those educators. These are their unresolved anxieties.

**[01:01:23]** I think both are completely accurate. For our purpose, the most important thing is that

**[01:01:29]** we are already done. That's why we're good.

**[01:01:36]** Because we're old.

**[01:01:38]** Because we're old. Enjoy it probably for the first time in your grandfathers.

**[01:01:43]** That's a competitive advantage, being old.

**[01:01:49]** But don't we live in a world where the result defines everything?

**[01:01:58]** We do, but we want that.

**[01:02:03]** But you realized that nicely. You know, I don't really see

**[01:02:06]** us having these discussions anywhere in the mainstream, in companies.

**[01:02:11]** At Red Button, we recently discovered Honza Veselý and here Katka Jiřínová.

**[01:02:17]** H to T. That’s actually it. Humanity, technology.

**[01:02:20]** That’s H to T, and if we don’t cultivate the H, it’s just a huge process.

**[01:02:25]** But we really go back to the question of to what extent companies

**[01:02:31]** want to make money, right?

**[01:02:35]** Hold on, okay?

**[01:02:38]** To what extent are companies just machines for money,

**[01:02:43]** or do they have some inner purpose?

**[01:02:46]** If we go back to the Baťa phenomenon,

**[01:02:49]** or we could go back to the Volman phenomenon,

**[01:02:52]** like these types of industrialists,

**[01:02:55]** their ambition for the company was much bigger.

**[01:02:59]** And it was partly based on integrating people and developing their human sides

**[01:03:02]** to the extent that Baťa thought about

**[01:03:07]** people not seeing the Baťa factories from their homes,

**[01:03:10]** so they wouldn’t be stressed by seeing work.

**[01:03:13]** And at the same time, he arranged for them to pass through a cultural square on their way home from the factory,

**[01:03:18]** with a gallery, theater, and some nature.

**[01:03:21]** And we can’t say Baťa didn’t want to make money.

**[01:03:26]** He definitely did.

**[01:03:30]** I hear you. You’re on Red Button’s ground here.

**[01:03:33]** That means cultivating society with companies and entrepreneurship.

**[01:03:36]** Michal Šrejer is here.

**[01:03:39]** Just kind of on that...

**[01:03:42]** And I’ll leave you now because he asks a lot.

**[01:03:45]** I think we’ll have a digest on this later.

**[01:03:48]** Besides the fact that juniors and their bosses from many companies are healthy.

**[01:03:52]** Do you think it still makes sense to write bachelor’s and master’s theses nowadays?

**[01:03:59]** when we have AI?

**[01:04:02]** Of course it does.

**[01:04:05]** There you see.

**[01:04:08]** An expanded answer.

**[01:04:11]** This is obviously a debate happening very strongly right now in universities.

**[01:04:14]** Extremely strongly at this moment.

**[01:04:17]** It's important to say that universities have quite a lot of

**[01:04:20]** inertia.

**[01:04:23]** You have in the accreditation that you finish school with a bachelor's or master's thesis.

**[01:04:26]** That accreditation lasts for 7-10 years, for example.

**[01:04:29]** And it's much easier to let students write nonsense for 7-10 years,

**[01:04:32]** than to change that accreditation.

**[01:04:35]** Because the system is bureaucratic.

**[01:04:38]** This is often one of the reasons why it happens.

**[01:04:41]** I think the core is somewhere else.

**[01:04:44]** The master's thesis is also some kind of formative experience

**[01:04:47]** for the person writing it.

**[01:04:50]** I think what universities should do is provide

**[01:04:53]** much more support to people in writing.

**[01:04:56]** And focus on what AI won't do.

**[01:04:59]** AI won't do research, for example.

**[01:05:02]** It won't do qualitative interviews,

**[01:05:05]** I'm all about those human talks.

**[01:05:08]** It won't do that. And accept that students use AI

**[01:05:11]** for many parts.

**[01:05:15]** We are in a linked field.

**[01:05:18]** You must already have a bachelor's to study with us,

**[01:05:21]** to make it clear.

**[01:05:24]** We have something like an AI usage declaration

**[01:05:27]** and we want to know from students

**[01:05:30]** what they used during writing and how they used it.

**[01:05:33]** If they provide prompts, they include examples of those prompts.

**[01:05:36]** And the part we don't want them to use

**[01:05:39]** is the interpretation of the conclusions.

**[01:05:42]** That should be their own work.

**[01:05:45]** And for the rest, we have been very open to anything.

**[01:05:48]** And in the end, you find out the most common thing

**[01:05:51]** students use it for

**[01:05:54]** is brainstorming over materials

**[01:05:57]** and working with a larger discovery of texts.

**[01:06:00]** Which is probably fair.

**[01:06:03]** And the brainstorming is extremely interesting,

**[01:06:06]** because it's hard when you're in that studio,

**[01:06:09]** with all those texts from the experts in labs around the world,

**[01:06:12]** to come up with some idea on top of that.

**[01:06:15]** Because they have usually already figured it out.

**[01:06:18]** And to me, that seems fair.

**[01:06:21]** Otherwise, I'll give a little teaser.

**[01:06:24]** We have arranged with Michal Pěchouček

**[01:06:27]** for one of the upcoming breakfasts.

**[01:06:30]** So I won't say the exact date,

**[01:06:33]** but I think we'll be talking about something

**[01:06:36]** that actually suits me.

**[01:06:39]** If I were a parent of a twelve-year-old today,

**[01:06:42]** what three specific skills would you recommend developing over the next five years and why?

**[01:06:45]** That was your last slide.

**[01:06:48]** I have kids aged 10 and 13, so I didn't write it down.

**[01:06:51]** How so?

**[01:06:54]** I would probably start from

**[01:06:57]** what the child naturally tends toward.

**[01:07:00]** And on the other hand,

**[01:07:03]** and here I might sound quite conservative,

**[01:07:06]** some experience with drill,

**[01:07:09]** like when someone really learns a foreign language that isn't English.

**[01:07:12]** Or when someone enjoys learning a musical instrument.

**[01:07:15]** This particular experience helps a lot.

**[01:07:18]** But it seems to me

**[01:07:21]** that I would rather look for things

**[01:07:24]** that support that in the child.

**[01:07:27]** But the question actually asks,

**[01:07:30]** which are the important skills,

**[01:07:33]** which I have never named here and you're trying

**[01:07:36]** to see if I might say them by chance.

**[01:07:39]** I understood the question, but I won't say the skills.

**[01:07:42]** But I think it's still the good old combination

**[01:07:45]** of being able to read, write, and do math.

**[01:07:48]** You might think I've gone crazy,

**[01:07:51]** but I don't think that's true.

**[01:07:54]** Those are the three basic skills,

**[01:07:57]** when you can read, write, and do math.

**[01:08:00]** It's just that reading well is something

**[01:08:03]** you have to really work hard at by

**[01:08:06]** reading a lot and discussing it a lot.

**[01:08:09]** Writing well means writing a lot

**[01:08:12]** and letting others criticize it.

**[01:08:15]** And counting doesn't necessarily mean

**[01:08:18]** you have to do algebra,

**[01:08:21]** but to know it and not be afraid of it.

**[01:08:24]** I feel like we are afraid

**[01:08:27]** And finally, these three things are an experience of the self.

**[01:08:30]** and that's wrong.

**[01:08:33]** Reading, writing, and counting still seem to me like good skills,

**[01:08:36]** which include all those aspects.

**[01:08:39]** That means they have the aspect of

**[01:08:42]** conveying knowledge to us,

**[01:08:45]** they have the aspect of being highly social,

**[01:08:48]** like communication and so on.

**[01:08:52]** If I leave out counting,

**[01:08:55]** but if not overdone, counting can also be a communication matter.

**[01:09:01]** They are formative for the person.

**[01:09:04]** It's formative when you have to engage with a poem,

**[01:09:07]** it's formative when you have to engage with a complex formula.

**[01:09:10]** You realize that maybe someone who can't solve it,

**[01:09:13]** someone else already solved it.

**[01:09:16]** It's quite a powerful experience of the self.

**[01:09:19]** All three things.

**[01:09:22]** Thank you. Isn't the most important skill in the age of one-click content generation the art of asking questions?

**[01:09:25]** Absolutely, yes.

**[01:09:28]** And how do I learn that if I haven't asked any questions before, not even to myself?

**[01:09:31]** The joke is, to list strictly that

**[01:09:34]** the art of asking questions is important,

**[01:09:37]** those are good skills.

**[01:09:40]** But they don't fall from the sky.

**[01:09:43]** The point is, we are too overwhelmed by the new era

**[01:09:46]** of sellers of warm water,

**[01:09:49]** who used to sell us self-development courses some time ago.

**[01:09:52]** And now they tell us that everything will be solved by a prompt with the Socratic method.

**[01:10:00]** That doesn't work. No, not because the prompt wouldn't work, it does.

**[01:10:08]** But the problem is that people don't understand the answers.

**[01:10:11]** The problem isn't that AI talks nonsense,

**[01:10:15]** it sometimes does. The problem is that people don't understand what it says.

**[01:10:19]** The problem is that we don't know what questions to ask,

**[01:10:22]** because we've never really practiced just asking.

**[01:10:27]** Learning to ask questions is a lifelong program.

**[01:10:33]** In ancient Greece, it ended at the execution site.

**[01:10:40]** Well, at the execution site by drinking poison. Or exile.

**[01:10:46]** Which isn't so bad either.

**[01:10:51]** What mistake do parents and schools make today when preparing children

**[01:10:56]** for a world with AI?

**[01:10:59]** What mistake?

**[01:11:01]** If I knew how to do it right, I don't.

**[01:11:10]** I think very often I encounter,

**[01:11:17]** and I think it's very Czech, the idea

**[01:11:20]** that a child goes to study what they will do.

**[01:11:24]** It doesn't matter if it's high school or university.

**[01:11:27]** We have the idea that when we study something,

**[01:11:30]** that's what we will do.

**[01:11:32]** I studied aesthetics, Matěj studied philosophy.

**[01:11:35]** I don't think either of us make a living from it.

**[01:11:38]** Though we could, by the way.

**[01:11:40]** I also drafted railway superstructure and track.

**[01:11:46]** In this country, we often have the idea

**[01:11:48]** that a human life is built on the idea

**[01:11:50]** that you study something, then find a job,

**[01:11:53]** then stay in that job until retirement,

**[01:11:56]** and then finally live.

**[01:12:00]** No, it's much more normal to change careers,

**[01:12:03]** it's much more normal to keep learning.

**[01:12:06]** Change is much more normal.

**[01:12:08]** Now it's just more normal,

**[01:12:10]** I wouldn't want it to sound like,

**[01:12:12]** I said it terribly badly,

**[01:12:14]** that those of you who don't do this aren't normal,

**[01:12:16]** but I didn't mean it that way.

**[01:12:19]** The problem is that the world doesn't work like that.

**[01:12:22]** In this world, things are always changing,

**[01:12:24]** and sooner or later things will change,

**[01:12:27]** and you will be unhappy in the positions you hold.

**[01:12:32]** I think we're making a big mistake by

**[01:12:35]** already deciding for the kids

**[01:12:37]** what they will do and how they will make a living.

**[01:12:41]** We try to solve their problem by choosing schools for them.

**[01:12:44]** But they will become adults,

**[01:12:47]** and then they'll have to solve that problem themselves.

**[01:12:49]** This way, we just make them another child.

**[01:12:52]** You said that very nicely.

**[01:12:55]** You're talking about relationships and connections.

**[01:12:58]** Yes, but between what, if you don't have knowledge,

**[01:13:02]** you have AI, but nothing to connect it to.

**[01:13:05]** One more time, slowly.

**[01:13:09]** You're talking about relationships and connections.

**[01:13:11]** Yes, but between what, if you don't have knowledge,

**[01:13:14]** you have AI, but nothing to connect it to.

**[01:13:18]** AI has no knowledge.

**[01:13:22]** AI can generate text well.

**[01:13:25]** That sounds like a technical detail, but it isn't.

**[01:13:29]** And to some extent,

**[01:13:36]** I don't understand the mechanical idea.

**[01:13:40]** I have entities and relationships between them.

**[01:13:44]** It depends on how I look at it.

**[01:13:48]** Entities aren't predefined; I constitute them.

**[01:13:51]** I decide what the entities are, I describe the relationships.

**[01:13:54]** AI offers us a reservoir of generated ideas

**[01:13:58]** about how things could be.

**[01:14:02]** What turns it into thinking is when

**[01:14:05]** we work with it in some way.

**[01:14:09]** And if the mind isn't prepared for it, it doesn't work.

**[01:14:12]** Maybe I'll add to that, because you say

**[01:14:16]** that a prompt alone won't cover it.

**[01:14:19]** Michal, Spartička, and I were discussing

**[01:14:22]** physical AI, robots,

**[01:14:25]** and what happens when they get hands, legs, and everything.

**[01:14:28]** That's also about forming relationships there.

**[01:14:32]** There are many examples now.

**[01:14:35]** I wouldn't say you can't form relationships with it.

**[01:14:38]** What happens is known, if you really look

**[01:14:41]** just a little further around the corner,

**[01:14:47]** when it will already be in the physical world

**[01:14:50]** and when it will already be running and doing things for us

**[01:14:53]** like shopping and cleaning and all that,

**[01:14:56]** then you look into what will still be happening with us.

**[01:14:59]** I have only one answer.

**[01:15:02]** Androids in electronic ovens.

**[01:15:05]** I think Dick knows much more about it than I do.

**[01:15:08]** Honestly, I can't imagine it

**[01:15:11]** well enough to be able to

**[01:15:14]** distill it further.

**[01:15:17]** It's true that we form parasocial relationships.

**[01:15:20]** It's also true that in Holešovice there's a mess where you can buy sex with a robot,

**[01:15:23]** that's also true.

**[01:15:26]** We talk a bit less about whether it will be nice,

**[01:15:29]** if you could bring your own picture like this

**[01:15:32]** and decide that you want the robot to look like this.

**[01:15:35]** To print it out.

**[01:15:38]** Yeah, like a neighbor.

**[01:15:41]** Because we all talk about celebrities, but not neighbors.

**[01:15:44]** But honestly, I don't know.

**[01:15:47]** I really don't know.

**[01:15:50]** Let's lighten up.

**[01:15:53]** Did you use it when making the presentation?

**[01:15:56]** Yeah, I'm looking forward to that presentation. I used it.

**[01:15:59]** Why? And I'll explain how and why.

**[01:16:02]** They say water turns into wine.

**[01:16:05]** No, no, no. I write habilitation theses about what changes in the world.

**[01:16:08]** And the habilitation thesis is in the writing phase.

**[01:16:11]** And I usually write things by

**[01:16:14]** reading a lot about them

**[01:16:17]** then standing up

**[01:16:21]** and walking around the city recording into a voice recorder

**[01:16:24]** what the chapter should be about.

**[01:16:27]** Then I transcribe it, work with it,

**[01:16:30]** and then again use Deep Research to check if it makes any sense and so on.

**[01:16:33]** In fact, this presentation has about 70 pages

**[01:16:36]** of heavy text.

**[01:16:39]** It basically contains Aristotle and his concept of Hexis

**[01:16:42]** and how we acquire virtues

**[01:16:49]** The content includes references to Wilfred Bion,

**[01:16:52]** my favorite psychoanalyst.

**[01:16:55]** And then I took that thing and told him,

**[01:16:58]** please, I will be lecturing at Brain and Breakfast.

**[01:17:01]** I need to simplify it into some position,

**[01:17:04]** that seems acceptably simple to me.

**[01:17:07]** Of course, I kept Robert Reich in there.

**[01:17:10]** And he said, leave Reich and Winnicott in.

**[01:17:13]** I had to laugh because I think Winnicott,

**[01:17:16]** is quite a well-known figure in the Czech-speaking environment.

**[01:17:19]** So in the end, I made the presentation,

**[01:17:22]** the shortened version, just normally as I went through it.

**[01:17:25]** And otherwise, since the question is nicely detailed,

**[01:17:28]** today I do it like this,

**[01:17:31]** I use the Markdown Marp format for writing slides,

**[01:17:34]** which allows me to interact with it directly in Visual Studio

**[01:17:37]** or in Kruzor, right on it.

**[01:17:40]** So that's probably an exhaustive answer from me.

**[01:17:43]** There was a very interesting question,

**[01:17:46]** I don't know where it got lost,

**[01:17:49]** but it was apparently a lady from Besky,

**[01:17:52]** she said from the school, from the elementary school Hnízdo,

**[01:17:55]** and she is a career counselor.

**[01:17:58]** So what would you actually recommend to career counselors?

**[01:18:01]** How to approach this?

**[01:18:04]** I think it's still the same, right?

**[01:18:07]** They roughly need to know what the child enjoys and what their aptitudes are,

**[01:18:10]** and that seems the same to me.

**[01:18:16]** If we want to be normal happy people,

**[01:18:19]** then maybe it really doesn't matter at all,

**[01:18:22]** which school I go to,

**[01:18:25]** if I decided that AI will destroy it,

**[01:18:28]** then I won't go there.

**[01:18:31]** If it's something I want to go to and what keeps me there and makes sense to me,

**[01:18:34]** and what somehow fulfills me, then I should go there regardless of

**[01:18:40]** Because that's not the point.

**[01:18:43]** In school, the point is for me to become a better person.

**[01:18:46]** I understand.

**[01:18:49]** I'm thinking.

**[01:18:52]** There is some...

**[01:18:55]** So it keeps revolving around what to teach in elementary school.

**[01:18:58]** But that's good.

**[01:19:01]** I'll try to take a bite at what to teach in elementary school.

**[01:19:04]** I think that what to teach,

**[01:19:07]** is not a decision of elementary or secondary schools.

**[01:19:10]** Because what to teach is prescribed,

**[01:19:13]** like ŠVP and RVP.

**[01:19:16]** That is prescribed.

**[01:19:19]** Alright.

**[01:19:22]** What I see in computer science,

**[01:19:25]** the updates to those RVPs

**[01:19:28]** seem to be really delayed.

**[01:19:32]** But that's beside the point.

**[01:19:35]** I'd rather think about how to teach.

**[01:19:38]** If we say that school has three components,

**[01:19:41]** in terms of passing on knowledge,

**[01:19:44]** teaching children social relationships,

**[01:19:47]** and some kind of subjectification,

**[01:19:50]** meaning that the person becomes,

**[01:19:53]** or the child becomes a strong person,

**[01:19:56]** then we probably know for the future,

**[01:19:59]** if we maybe...

**[01:20:00]** put much more weight on subjectification, on the child becoming a healthy, strong, confident person,

**[01:20:09]** then we might reduce the focus on knowledge. But watch out, it's very tricky, the volume of what children learn is much bigger,

**[01:20:17]** than what we learned. Than me and you, because it's a significantly larger volume.

**[01:20:27]** And I think the much more important thing is probably the how. That means really trying to get inspired in education

**[01:20:37]** based on productive failure. Maybe reduce the volume we want to pass on and focus more on

**[01:20:48]** letting children find their own path. On the other hand, not giving up on evaluating the children.

**[01:20:55]** I'm not really an advocate of no evaluation, but more focus on this, meaning the process,

**[01:21:03]** because ultimately the basic problem we have can be shaped into a simple thing.

**[01:21:16]** If we wanted to name one skill we need to survive healthily in this AI world,

**[01:21:26]** it's metacognition. It's our ability to think about our thinking.

**[01:21:31]** And that's something that can be trained in any subject when passing on any knowledge,

**[01:21:38]** because it's about how we approach it with the kids.

**[01:21:42]** It's a bit of a count's advice. I actually teach some computer science at a high school in Prague,

**[01:21:49]** which is a very different position than a teacher at an elementary school in the Beskydy Mountains.

**[01:21:56]** It's a bit of a count's advice, sorry for that.

**[01:21:58]** I also remember when I was recently at a breakfast with a teacher for life,

**[01:22:05]** and it was a good group of people who I think really understand it.

**[01:22:07]** And you also said that school is actually a place where you learn social interaction.

**[01:22:12]** That it's very important to know how to communicate.

**[01:22:17]** Thank you once again for the great lecture. Take care, bye and see you.