The Fifth User Interface


Izumo Ooyashiro, Izumo, Japan
1.4/50 Summilux ASPH, LEICA M (Typ 240)


Before you know it, it is the end of the year. I have not been able to update my blog for half a year *1 Besides, I have been thinking about it quite a lot lately whenever people ask me about it in various settings. I think it would be a good idea not to keep them like a lees, so I would like to write a little about them in the hope that it will serve as a momentum for me to write them down.

-

I continue to speak and be interviewed several times each week, and there are a few things I have been asked repeatedly in recent weeks.

  1. Will LLM eliminate jobs?
  2. What jobs will AI eliminate in the future?
  3. What can Japan do while global big players like OpenAI & Microsoft, Google/DeepMind, etc., are going all out?
  4. What can you tell us about "Bukkon-Densai" - Material spirit with Cyber tech as mentioned in Pivot, etc.?
  5. What do you think school education should be like in the future? (After the March throwback to the Ministry of Education, Culture, Sports, Science and Technology)
  6. What should we think about academic achievement in the future?
  7. What is the kinetic theory of the valley of the wind?
  8. Is there a prototypical land in the Valley of the Winds?

This is a good place to start. There are also a variety of other questions that I wonder why they are asking me, such as the current economy, future interest rate hikes and the outlook for the economy, and current stock price levels.

The first two questions, work-related topics, are quite common, and are often asked as if they are the flip side of the education-related questions in the latter part of the survey.

-

It is true that the last topic about the Valley of the Winds has always been a stealth activity that we did for the first few years, and we have hardly disclosed our considerations, but it has been a long time since we disclosed the existence of this project on this blog, and our website has been up and running stealthily for a long time. Although we have continued to be adamant about becoming a national project, we have shared several introductory and partial talks with the Ministry of the Environment and the Ministry of Land, Infrastructure, Transport and Tourism due to strong demand.

kaz-ataka.hatenablog.com
aworthytomorrow.org
Kazuto Ataka, "Considering a Worthy Tomorrow" (Central Environmental Council April 1, 2022)
Kazuto Ataka, "Considering a Worthy Tomorrow" (Central Environmental Council, General Policy Subcommittee June 30, 2023)

In addition, Tsunehiro Uno, one of the founding members and a core member, has compiled a considerable amount of articles on the "Slow Internet" in a semi-portal-like manner. For a comprehensive perspective, please wait for the Kaze no Tani (Valley of the Wind) book, which is currently being written, but if you are eager to learn more, please have a look first as there are considerable information links in the following pages.

slowinternet.jp
aworthytomorrow.org

If you read them, I think you will see that the last question is meaningless. It is an effort to define solidly what it is in the first place, and the truth is that there is probably no land in the world that satisfies the essential conditions, at least in terms of thick requirements. *2

-

Let's get back to where we started, but I hope to begin with a recap of the historical implications of LLM *3, which I discussed at the Ministry of Education in March *4.

The recent evolution of LLM has been tremendous, perhaps to the point of being a bit of a surprise to those of us who have been in the industry for some time. It's hard to realize unless you've been using the paid model of ChatGPT, but it's also a very good tool for analytical power by plugging in plug-ins, various external API'd functions, and purpose-specific AI features, As Google is making a concerted effort to launch Bard, it is in the process of becoming one of the new Internet gateways following Portal, Search, and Social.

If you ask most of them, they can produce a considerable level of output at a considerable speed, without any rest at all. Much of the hallucination problem that was raised a year ago has already been resolved as a result of OpenAI's alignment work *5. Google's Bard, based on searches, is naturally quite accurate, and ChatGPT does a fair job of writing condolence sentences with intricate backgrounds. It can also translate with the speaker's point of view in mind. There is no doubt that we have created an "unusually" well-trained and intelligent assistant *6.

What amazes me is that LLMs are able to use language, the most important part of our human capacity, at a level that surpasses most of us. Words are half-magic, and I recommend listening to Yutaka Matsuo to understand what I mean (then Associate Professor at the University of Tokyo, now Professor at the University of Tokyo) speak at the first TEDxUTokyo event held at the University of Tokyo in the spring of 2012. He is now a man of the hour (the guru of Deep learning and AI in Japan), but this video is probably one of his lesser-known supreme talks. *7

youtu.be

From another perspective, this is the emergence of a completely new intellectual production UI*8. The UI of human intellectual production has changed from paper and ink, to black screen, to GUI*9, to multi-touch screen, each of which has created huge industries and wealth.


Kazuto Ataka, "Considering Future Human Resource Development," March 24, 2023
Ministry of Education, Culture, Sports, Science and Technology Expert panel on the future of curricula, learning guidance, and learning evaluation (3rd meeting)

Not to mention paper and ink. In the absence of these, there were only a limited number of places where letters could be left behind, and the transmission of human knowledge depended on people like Hieda-Arei, mentioned in my previous blog entry, who excelled in extreme memory. There is no doubt that the advent of paper led to the opening up of the brain by an order of magnitude, and that the impact of thoughts becoming objectively readable letters was tremendous. The world's first bureaucratic selection system based on competence, the Chinese mainland's system of national examinations, would not have been possible without the paper and ink invented in this region.

kaz-ataka.hatenablog.com

The advent of the word processor, office computer, and PC*10, which I call the black screen, created the fusion of computing and communication that NEC held up C&C*11, revolutionized white-collar working assumptions, and brought enormous wealth to companies like DEC, IBM, and NEC.

The GUI, which Apple brought to the general market and spread rapidly with Windows 95, gave birth to the browser, and the concept of multimedia, which until then had existed conceptually, took shape, giving birth to the Internet economy. Although the Internet itself is quite old, it is fair to say that up to this point it had a military and academic existence and nothing that could be called an economy. It was also the moment when text, images, music, and video all finally melded together through the power of the bit (digital) (see Being Digital, a historical work by Dr. Negroponte, the first director of the MIT Media lab).

en.wikipedia.org

The multi-touch screen, which originated from Sony CSL/University of Tokyo's Dr. Rekimoto's SmartSkin, gave birth to smart devices, especially smartphones, which have fundamentally changed the world. Having a smartphone is now close to a basic human right in many civilized societies, and if you look at the top seven companies in terms of corporate value today, there are five companies that are involved in businesses based on smartphones in one way or another *12.


Companies ranked by Market Cap - CompaniesMarketCap.com
(as of 12/16/2023 12:35 JST

And now we can communicate with computers using the most useful tool for thinking, the "language," which is a characteristic of human beings themselves. There is no reason why this should not produce enormous change and wealth. So the world is up in arms, and I think it is only half natural. It is the fifth UI revolution.

And it's a return to human origins.


ps. A few days after I wrote this, at a meeting of a certain company, a question came up as to where the current LLM falls in the hype cycle. I replied, "This change is a fundamental one similar to the advent of smartphones, and it is not something that can be handled in the hype cycle. The world can no longer return to the way it was when we were not able to communicate directly with computers using natural language. The smartphone should not have had a single period of disillusionment, and it should have just continued to build a market and ecosystem around it". I am sure that many of my readers will agree with me.

*1:I have meetings every day until late at night and am over capacity one after another, and I have received a lot of miscellaneous requests, so even the little time I have left is being used up to respond to them. Naturally, I have only attended a minimal number of year-end parties.

*2:Whenever I get this question, I also realize that we are a nation that really wants answers. They believe that the example is out there somewhere. It has not fallen into place that the key is not to answer the question, but to create the future and to think itself. So, in discussions with the government and public organizations, the idea immediately becomes like a backcasting discussion. That in itself is not wrong as a procedure to finally come up with a plan, but many people do not understand that drawing that image is really the most difficult and valuable part. And they want easy answers. The future is in our hands. It is time to stop looking elsewhere for answers.

*3:large language model

*4:note: not exhaustive

*5:let's just say we've reached a level where we can publish a paper on how to make this easier to induce

*6:the ancient Romans would have called him a slave

*7:By the way, it ends with a standing ovation. I happened to be there as one of the speakers, but the content was so highbrow that I asked the people I was with after the talk if they understood it. I would like to add that many of them commented that they could not comprehend what was being said. I think more people might understand somewhat now. LOL

*8:user interface

*9:graphic user interface

*10:personal computer

*11:Computer & Communication

*12:Apple, MS, Alphabet, Amazon, Meta