Aleksander Dmitriev, IBM: ‘It’s time to transfer from bureaucracy to some weird mess, and it's alarming’

Realnoe Vremya newspaper publishes a cut speech version of a speaker of the open lecture 'Business in the Era of Disruptive Innovations'. The representative of the IBM client center Aleksander Dmitriev spoke about information overload, when a person loses information about cognitive systems, which in a few years will help people to make decisions.

Information overload

I am a representative of the client center of IBM Company. I flew from Moscow. Thank you for the opportunity to speak here. I liked the previous lecturer [the speech of Evgeny Chereshnev from Kaspersky Lab via skype — editor's note], he is like good Lenin. He said that he wishes people spoke and things or systems understood him. In fact, today I will try to tell you about it very quickly. <...>

Somewhere in the 1960s, there were two philosophers: Tamotsu Shibutani, who was involved in the blood feud and some other stuff, and there was Stanislaw Lem. And they talked about one and the same thing. <...>

Well, Stanislaw Lem wrote… 50 years ago! He wrote a great book Summa Technologiae, and for whom it is boring — he outlined all these things in the stories and books about Ijon Tichy. For me, in my time it was a table Bible. He [Lem] says that large electronic machines faced a very difficult problem — an information fence. They can't find necessary information in tonnes, megatonnes of these databases and they do what the scientists and developers are trying to understand, what maybe someone already have done, or where to go next. He exactly identified this problem 50 years ago.

Not a bit elementary Watson

A couple of words about the IBM company. We realized this and decided that we will fight against that. We began to develop the Watson system, about which people talk a lot, but, actually, it is a huge range of products. We began to do following: creating a system where you can upload knowledge from various fields.

The most famous, popular entertainment was a cookbook. On the basis of a huge number of recipes you can tell Watson: 'Dear Watson, I want to make a new recipe from such-and-such things'. It will think, dig in its database, think what goes best with what, how to cook, and will give some very very good delicious recipe.

But. I have a last year calendar with an exceptionally cool thing, which indicates some problems of such systems. In one of such recipes it says: take the Halal-kosher salt, a kosher product and a kilo of pork.

Sane people understand that there are products that are not combined due to historical, cultural, and religious traditions. But a cognitive system, focusing on cooking, has not absorbed database of religion and culture. Therefore, it may make such mistakes, unforgivable for normal people.

Therefore, the moment when we can speak with the system will not come soon. Although such projects already exist. <…>

What restricts the action of this invisible controller? It does not exist but it is in our heads. The specified system of rules, like in chess: there can be a million of rules and positions, but everyone knows how to play

Cognitive systems through an example of queues

Here I would like to say a few words about the new era. It is not that there are disruptive technologies. It is quite different. We all say what's going on. I often attend different conferences — few people say why. Why the companies appeared like what will speak after me, dear Mikhail Fisher from Uber company. This is the company that really rocked. It has radically changed the approach… It has dramatically changed the approach to the large system that allows you to take a taxi.<…>

Who monitors the queue length in the supermarket? Is there a person who says: 'No, don't wait here, this line is longer. Wait here, this line is shorter?' No. A person does it himself, if he's smart enough to see. Moreover, we also think: 'Oh, there's a woman with a huge pile and these ones, but here there are only two bachelors with a jar of vodka and a piece of bread. I will wait here. The number of people is the same, but here it is faster, right?'

This is a complex theory and a part of the self-organizing system. What restricts the action of this invisible controller? It does not exist but it is in our heads. The specified system of rules, like in chess: there can be a million of rules and positions, but everyone knows how to play.

Now it's time to transfer from the rigid structures, limited in space, in time, bureaucracy, hierarchy (I'm CEO, I have an assistant, representatives, here we sit, we have an office) — to transfer to some weird mess, which no one can understand

Available information kills bureaucracy

What restrains it? Let's suppose that we have a large number of columns, pillars and I only see several queues. But the supermarket is huge and there are a lot of them. I will choose that is nearby. Information, how far it goes and how fast (speed), or I see that where people go.

By the way, like in the airport — I will use your example. They opened a desk. A person begins [to think] — a social trap – to go there or stay here, I've been waiting here for a long time already.

When it is a lack of time, the speed of information dissemination is the most important factor that affects the behavior of self-organizing systems.

What happened? A simple thing. The first point: an opportunity appeared to give the information to every person and to do it quickly. Uber is just one example, powerful, beautiful. But there are a huge number of things that we don't notice. <…>

Now it's time to transfer from rigid structures, limited in space, in time, bureaucracy, hierarchy (I'm CEO, I have an assistant, representatives, here we sit, we have an office) — to transfer to some weird mess, which no one can understand.

And it's alarming. Hence the word 'disruptive' (technology) carries some negative connotation because, in general, they do useful things, but you can use these things in different ways. <…>

The companies began to understand and customers began to realize that information is very important. Whatever you do, if we seek information on the Internet, we will face this fence. We can't define whether it's true or not

Trap for a dilettante

The companies began to understand and the customers began to realize that information is very important. Whatever you do, if we look for information on the Internet, we will face this fence. We can't define whether it's true or not.

Is it true that iPhone 7 recommend us to drill a hole with a drill? I'm not an expert. It might be mistaken. <...>

A person who wants to get information — he will find it everywhere. And now the company IBM is creating systems that… They are very complex large projects, on the one hand, but there are simple tasks that we solve quite successfully.

Basic direction in which the major IT companies are moving, operating in the market of analytics — it is an attempt to learn how to cope with a vast array of littered data (seriously, littered data), to find useful information and to allow a person to make decisions in cases where he is not an expert

Spherical horses in a vacuum

We fill [Watson] with knowledge in a certain area. For example, oncology. There has been a successful pilot project in Russia, in collaboration with the Ministry of Health. When in one narrow section (because there are a lot of types of cancer and each has its own specifics) the knowledge base is made for the doctors, a history of a personified disease for a specific client, the data is checked and cleaned (because in 25%, as practice has shown, the medical history contain mistakes when copying, rewriting and so on).

And when the data is cleaned, with markings: 'Please note, here a tumor is shown two, but here it is four — this can't be. Doctor, please, look, check, reschedule the tests,' a doctor has a compilation of information from various sources (the most recent studies, statistics of treatment). Then a person makes a decision based on this that established, selected information. <…>

Thus, to conclude this 20-minute lecture: basic direction, in which the major IT companies are moving, operating in the market of analytics — it is an attempt to learn how to cope with a vast array of littered data (seriously, littered data), to find useful information and to allow a person to make decisions in cases when he is not an expert. It is quite trivial things: choice of a product, of some services, where to go, what to do, what to do in a situation (including in emergencies).

I hope that in five years all of these things, just like 10 years ago if I told my wife that she would order food and they will bring them home. She'd say, 'What?'. But now you can order products home online, and not to carry pounds of potatoes. No, she did not do that, I did, but not always.

Therefore, I think that in five years what I am telling about, like about spherical horses in vacuum, it will be absolutely everyday thing for everybody sitting in this room.

By Yulia Krasnikova. Photo: Maksim Platonov