A positive example: Hype Cycle following Gartner Inc.

I have now spent fifty years doing digitalization. Among other things, I witnessed many hypes.

In this article, I will write about the hypes I remember. Mind you, this is not a scientific work. It is more something to bring a smile to your lips.
Let me report.

First: what exactly is a hype?

It is quite easy:
There is a trigger. It can be a technological advance, a single event or a mass hysteria…

Then the gold-rush will begin. And, more often than not, there will be a bitter ending… 
It often ends with:
A waste of effort.

Here are a few indices for hypes:

  • The number of newly founded enterprises on hype.
  • The amount of investment capital in the hype environment.
  • The number of presentations given by politicians, analysts and other wise-guys on the topic (mostly they do not understand anything about it).
  • And others (all you have to do is think a little, then you will come up with enough).

Currently, everybody is talking digitalization. It is a mega hype that consists of many sub-hypes.

Hypes around Digitalization

In the 1970ies, I worked at Siemens in “industrial informatics” (that is how I called it). It was really engineering stuff. We programmed in the classical way. For instance operating systems, modules for remote data processing and computer networks, databases and transaction monitors.

All this is not very exciting. And it is also rather hype-free. Simultaneously, hypes like small computers that could play games, were invented. That was decidedly more exciting.

AI

The first hype I experienced consciously from the outset was the AI hype. Everyone who considered himself modern did a bit of Artificial Intelligence. For me, in the industrial computer science, the topic was out of reach.

In those days, the magic word was expert systems. It was particularly popular in such sectors as medicine. The workshops on AI were full of digital dreamers.

On top of the expert system, there were also nice exotics – some friends of mine, for instance, worked on a system that was supposed to hear from the noise of a chopper if all was fine inside. Well, if I am correct, I have to write that is what they tried – the project never became reality. So this is another one of those hype fates.

Those were the days when two programming languages divided the AI community into two: For the one group, Lisp was the only true AI language, for the other group, Prolog.

As far as Prolog was concerned, I soon saw that the hardware was not yet good enough for the costly back-tracking. Regardless, much was done with Prolog, especially in Japan. As far as I know, the projects were mostly academic – and not much was actually achieved. There was also a hobby version. It was called turbo Prolog (in analogy to Turbo-Pascal). At the time, one of the InterFace enterprises was the InterFace Computer GmbH. They probably had the world’s best Prolog. And the development of IF Prolog actually ruined the enterprise. To this day, I know nothing about Lisp.

There was no definition that indicated how artificial intelligence differs from normal intelligence or from algorithmically determined software. As you all know, there is another AI hype today. And whenever I ask one of the speakers how he would define artificial intelligence, I seldom get a good answer. Later in this article, I will give you my personal definition – but I am not sure that this is any better than what I usually hear from the experts.

In my life, I knew many hypes. Big ones and small ones. There were so many of them that I definitely cannot remember them all. Here are some of them (those that I remember):

Very early, there were the hybrid calculators. The synergy between analogous and digital concepts was supposed to open up new dimensions. This hype ended before it even seriously started. Then I remember the Ontologies. They were sponsored like no other informatics topic. Risc processors were supposed to make servers faster. Object-oriented programming was a hype and it had a few smaller children, such as object-oriented databases. Some of it disappeared, some became the generally accepted standard.

There was a time when everything was about what colour your office was. To be sure, today this is all self-evident, but my HIT/CLOU customers from the finance sector were really enthusiastic about the colour “RED” (for the red numbers on their balance sheets …).

The wish to have colour also promoted the client-server hype. It was based on colourful Windows PCs and, as I see it, it was to a huge extent due to people wanting to play Solitaire on their computers. The “organizer“ was hype until the smart phone replaced it as a combination of organizer and mobile phone. And on all these devices, you can also play solitaire.

Then there was the hype of rich clients (basically, today every smart phone is a rich client).

At the same time, the internet came. And later the WWW2. So we had two hypes simultaneously. The difference lay in the fact that some people had discovered that the internet was not only made up of consumers. This had been the case from the outset, because without someone offering something, there is nothing to read.

For twenty years, I have now regularly served on the jury board of BayStartUp (that is the enterprise that organizes the Business-Plan-Competition in Bavaria – and it now also has other things on offer for people who want to found an enterprise). So there I see what hypes are currently en vogue. Here are some of them:

There were times when most of the young entrepreneurs wanted business models around APPs in order to later develop PORTALS.

Two years were mostly about 3D printers. In between, we also had the 3D spectacles and the accompanying infra structure. … 
(Incidentally, 3D printers are a good example for how a patent can encumber the development of something and how, as soon as the patent has run out, the technology will prevail). 
As to 3D spectacles: there was one under the Christmas Tree three years ago at our home. At the time, I found the possible content rather weak – and it probably did not improve with time.

There was also a time when block-chain was totally fashionable. It might even take a top position as far as hypes are concerned – especially if you also consider the speculation money that was thrown into it.

One of the less noticed hypes is probably currently the use of sensors. After all, they will see to it that we will soon no longer talk about a human-machine interface, but about a world-machine communication instead. The humans will then only be part of the world (hopefully not a too disruptive part).

Many fellow hypers already warn against the dictatorship of machines. However, it will probably not be more damaging than the man-dominated governments.

Today, it is quite clear:
The current hype is again AI. In combination with big data. This is how we are warned against an atrocious world – and how we paralyse digitalization with data protection. This deprives us of many social chances. And we forget that AI and big data are only technological progress. Which always changed society. Mostly, life became easier.

To be sure, it was always important to tread carefully where technological progress was concerned. Humanity has not always succeeded in this.

This is why I also like Bertrand Russell:
» Every increase in technology will also bring an increase in wisdom, provided it does not decrease human happiness. «
I like to mention the motorized individual traffic. It destroys the planet and causes 1.3 million fatalities every year and several times that number in injured persons. They probably lacked wisdom when they invented it.

But then, who could have known this a hundred years ago? I believe technological progress will always cause change. And where there is change, there is always risk. However, if you consider the risk, you should not forget about the chances. Especially since mostly you only see the true disadvantages a posteriori. This is why I believe you should be cautious (wise) but not fearful (stupid) when it comes to hypes.

To me, the entire hype about data protection and data security looks like a good example of a very special kind of swarm stupidity – I adopt this term from Gunter Dueck.

From Philosophy to Technology.

The hardware that is used for digitalization still works with the same principles as in the 1970ies. The enormous progress we perceive as far as calculators, connections and storage (the holy trinity of IT) are concerned is a consequence of  Moore’s Laws  (Moore’sches Gesetz of 1965). This explosion in pattern recognition is a fantastic basis for many applications.

The principles his kind of AI follows are fairly old. Turing described it many decades ago – I think in the 1940ies.

It caused huge progress in the recognition of speech, which, basically, is nothing other than the interpretation and application of patterns. This is the central requirement for the new kind of software that is self-learning when it comes to translating languages and steering a car. Basically, it simply was not possible in old times by what we then had in terms of hardware.

For instance, Deepminds Alphazero-AI managed to beat the best Asian GO player, and after a short time it also managed to hammer Goldfish (the best chess game) . Mind you, the Chinese now woke up as far as GO is concerned. They now invest money and manpower in unbelievable quantities to promote AI. And they will soon take over the Americans with their old Watson and perhaps also Googles Deepmind.

? And just like the Chinese, the Bavarians now also woke up … and they want to show the Chinese what they can do – which is something to make the experts in digitalization smile. To be sure, the videos show a lot of space science. Eventually done by robots. For both, you need IT and AI. And that is not something you get for a few million euros. You will need to invest billions.

Of course, in Bavaria, they not only gave a nice performance. They also backed it with activity. So they installed a  ministry for digitalization and made a young lady named Judith Gerlach the State Minister for Digital Affairs in Bavaria on November, 12th, 2018.

Mrs. Judith Gerlach was born on November, 3rd, 1985 in Würzburg. She is married and Roman-Catholic and the mother of two children.

She is young, since she turned 33 years at the end of 2018. And it is certainly to her credit that she has two children. After all, I, too, learned a lot about digitalization from my children.

Her CV shows that she is a highly qualified lawyer. This makes me a little thoughtful, because it sounds more like data protection than like high technology. I wonder if she knows what exactly an operating system is. But perhaps her counsellors know all about it?
On the page you find if you follow the link above, the State Minister wrote:


“I see extremely good chances for Bavaria. We already put ourselves to the top of the queue in many sectors of digital development and we will continue to focus all our power in order to develop new technologies and use them in a socially responsible way. To this end, the new Ministry of Digitalization serves as motor, coordinator and think tank.“

Judith Gerlach, MdL 
State Minister


If she knew her business, she would never have written such a sentence. Or else, she would have asked someone to proofread it before publication, for instance someone sitting in the Bavarian ZD.B, which is also a foundation of the Free State and where people actually know something about it.

But then, being competent in the legal sector, she always has a good excuse. How is she supposed to know where Bavaria (or Germany and Europe) are digitally situated if compared to the world standard?

Perhaps it is a new (old) hype that we now use legal experts for the social design of our technological advancement (digitalization is nothing else)? That guarantees at least that the AGBs (and data security) will be perfect.

You will find information about the international position of Europe if you take a look at the Economic Partnership Agreement EU-Japan ( EPA, or JEFTA ) that was recently signed.
A single sentence states that Japan will lower the custom duty rates for European farming products and in return Germany will lower those for high tech products from Japan.

The trend seems clear: Bavaria and Europe are on their way to becoming farming countries. Maybe along with being a tourism country. We are certainly nowhere near the top when it comes to high tech exports. This is not only true for international trade agreements, but also for the technological reality.

Just look at the problems we have with 5G and Huawei. I remember my first Siemens job. Communication was a hype. And at the time, nobody could do anything without Siemens in communication technology. Today, however, nothing can be done without (red) Chinese technology. Nobody will pay anything today for yesterday’s joy.

The aforementioned sentence by the State Minister sounds like mockery:
We (Bavaria) are already at the top of digital developments in many areas.
The opposite is true.

To me, it is a surprise that the powerful VDA (Verband der Automobilindustrie e.V., Behrenstraße 35, 10117 Berlin) did not sabotage the EPA/JEFTA agreement. After all, the Toyotas might now well enjoy a price drop of 10 %. Mind you, I do not think the Japanese will be stupid enough to let the German Idiota drivers benefit from what they saved in duties (excuse the primitive pun).

Is it possible that the VDA currently has other things to worry about, because by now our holy cow Motorized Individualized Traffic has become a sector where others are miles ahead in technology?

The trade agreement is very beneficial for Germany as the world’s number three meat exporter. Especially with respect to pork, we are far ahead of everyone else – and this will soon also have the label “animal welfare” authorized by the Federal Ministry of Food and Agriculture (BMEL). Maybe this will cause a new pork hype.
I think it is all rather surreal?

RMD
(Translated by EG)

P.S.
So now I still owe you my definition of AI: 
As I see it, an Alg-SW becomes an AI software 
if it realizes its functional added value through learning and practice. In other words, it must collect and analyse many patterns that can then be correlated and evaluated with results in order to build up a huge amount of knowledge from experience. 
This is probably a little like the human brain works.

P.S.1
Here is my first Jubilee Article  after 50 years of IT.

2 Kommentare zu “2019 – My Personal Jubilee: 50 Years of Digitalization “My Hypes“”

  1. Hans Bonfigt (Monday February 11th, 2019)

    Interessant ist m.E. auch die Rezeption von “Neuheiten”.

    “RISC” sollte Prozessoren schneller machen. Sie zitieren das Trend-o-meter aus dem Jahre 1989 schon richtig.

    Aber schon damals, als die Menschen, weil ohne Internet, wesentlich mehr zur Reflexion neigten, war die allgemeine Wahrnehmung so grottenfalsch, daß einem auch noch nach 30 Jahren das Herz stockt:

    Es war die Vorgängerarchitektur CISC, die “den Prozessor schneller machte”. Ende der neunziger Jahre war der Speicherzugriff _DER_ Flaschenhals bei der Programmausführung. Also wendete CISC zwei geniale Tricks an:
    1.
    Kompaktifizierung des Assemblercodes, insbesondere Reduktion der möglichen internen Zielregister. So konnte man einen xchg register/register in zwei Bytes packen. Weil es nur vier mögliche Register gab, brauchte man für deren Spezifikation hat nur 2 Bits.
    Je kompakter der Code, desto mehr Instruktionen pro Sekunde konnten aus dem Speicher gelesen werden.
    2.
    Mikroprogrammierung. Für typische Programmieraufgaben wie “Vergleiche eine Zeichenkette” gab es Spezialbefehle, die etwa fünf Instruktionen zusammenfaßten, welche dann, typischerweise als Schleife, prozessorintern abgearbeitet wurden, ohne daß ein weiterer Zugriff im Instruktionsdatenstrom erforderlich gewesen wäre.

    Es war also CISC (in Zusammenarbeit mit einer “prefetch queue”), welches “die Rechner schneller machte”, und zwar ganz ungemein.

    RISC dagegen war schiere Simplifizierung. Die Codegröße explodierte. Es gab genau KEINEN GRUND zu applaudieren.
    Außer vielleicht den Hardwareentwicklern, die unglaublich schnelle Speicherbausteine zur Serienreife brachten.
    Aus diesem Grund hat IBM beispielsweise nie RISC-Prozessoren gebaut – auch wenn es jahrelang draufstand.

    Der CISC / RISC – Hype ist ein Paradebeispiel für den erkenntnistheoretischen Ansatz des Till Eulenspiegel:

    “Wenn der Kaiser nackt ist, ist der Kaiser nackt”.

  2. Hans Bonfigt (Monday February 11th, 2019)

    Wah. Ende der achtziger und nicht der neunziger Jahre war der Speicherzugriff der Flaschenhals. Für den Fall, daß Chris Wood mitliest.

Kommentar verfassen

*