"Matter" Creates Communication: The True Multimedia World - Yoichi Ochiai (Part 2)

Ochiai: But lately, I've been thinking about the difference between the amount of data humans hold in their brains and the amount stored in their genes. I realize humans now hold more data inside their brains. Thanks to the invention of printing, the spread of communication after the Industrial Revolution, the invention of film, and then computers. So if you asked if my whole life could fit on a single DVD now, it's starting to overflow. That makes the declining birth rate seem inevitable. Because instinctively, figuring out how to get what's inside your brain out into the world feels more important than passing on your genes.

The World of "Girls × Deep Learning"

Ochiai: That programming released online recently, combining 2D girls with deep learning, was pretty impressive. The image processing using neural networks is quite fascinating. This was all drawn by a machine trained with deep learning. Zooming in, it's definitely weird—faces distorted, eyes facing opposite directions—but if you feed it tons of 2D images, it draws pictures.

Fujita: Amazing! This is truly the future of AI.

Ochiai: The speed is pretty incredible. With 2D girls, there aren't that many elements—just hair, eyes, face, outline, hair color, clothes—so the amount of strokes needed is actually quite small. You can basically draw them by filling in areas with shapes. I was shocked.

Fujita: When it comes to what computers can do versus what only humans can do, it feels like the only thing left for humans might be making judgments like, "Connecting this part to that part seems plausible."

Ochiai: Exactly. When you assume humans might not actually be thinking as deeply as we think we are, and consider how to build a machine to sample emotions, I think a future where you feed your life log into a computer and it understands your emotions better than you do is actually coming pretty soon.

Fujita: Emotions are determined by various environmental inputs.

Ochiai: I can't help but think they are determined. After all, the human brain itself doesn't have any special individuality in its function. Emotions aren't particularly complex reactions either, and the total memory capacity isn't that huge—probably just a few gigabytes. When you think about it, it's not that impressive.

After all, the vocabulary humans can use is only about 5,000 words, right? If you made a dictionary with 5,000 words, even adding everything up, it would be at most about 120 megabytes. As for visual data, we remember having seen something, but if asked to draw it in detail, we can't really do it. That means we don't remember it properly. The connections between textual information or information expressible in words amount to at most a few megabytes – about the same data volume as a single book.

Humans aren't built to think in two dimensions; fundamentally, we're built to think in three dimensions. We only use desks or draw on walls and paper because of industrialization and gravity. Essentially, we want to live in society while skillfully handling three-dimensional things, but reinventing that means turning two-dimensional things into three-dimensional ones. Desks wouldn't exist without gravity, right? You've never seen astronauts having a meeting around a desk. Because they can't sit down. We essentially want to go to a three-dimensional world, but we haven't gotten there yet.

When I spoke with TeamLab's Toshiyuki Inoko, he said, "Things with mass are lame." It's true—when you watch a projected reflection, you don't think about the projector itself; you're just seeing the light. We've been pursuing the idea of transforming every tool into that state, thinking, "What if we made everything with holograms?" Touch, sight, sound—everything.

 

Humans will communicate through "matter"

Fujita: Probably until around 2020, we'll keep creating everything with VR and printers, but beyond that, I think the focus will shift to how we physically bring the mass of what was virtual into the real world. You often hear about digital technology getting closer and closer, or about extending human capabilities. Do you really think it's extending us?

Ochiai: We're clearly degenerating. Still, as long as computers don't disappear from this world, we'll probably be fine. We can just leave all the unnecessary functions to the computers, and in return, humans will be healthier running around outside, right? So "health" is the keyword.

Humans won't need to read maps, remember good restaurants, or have photographic memory. Machines will handle all that. Next, the idea that "you guys don't need to converse anymore" is probably another trap of AI. The notion that machines will just deliver results without conversation. You won't even need meetings. When machines gain the ability to speak any language, it becomes a tug-of-war between humans degenerating yet being enhanced.

Back around 2003, people often said only ideas had value. But ideas are infinitely available if you search Twitter; the idea itself means nothing. What's important isn't the idea, but the thing you create. In other words, I believe humans have shifted to communicating through physical objects. That's why the role of a consultant is increasingly about translating logic into code or physical form. There used to be quite a few consultants who didn't produce tangible things, but that approach has become largely meaningless.

If it doesn't become physical, it has no value anymore. Patents will probably become obsolete too, and copyrights will surely become obsolete as well. If there were a server that randomly rearranged lyrics databases to create and store billions of different song combinations, couldn't you sue every new song that came out for copyright infringement? In that kind of era, copyright would just become a shackle. In that state, what has meaning? Not the lyrics, but the sung version. It only has meaning when it becomes the sung sound; at the lyric stage, it probably has no meaning.

Fujita: The importance of production, often underestimated in Japan, will become even more crucial.

Ochiai: Next, I want to write about the aristocracy. It's a theory of aristocracy starting from the 19th century. In the past, the aristocracy guaranteed entertainment and science. Now, we're trying to return to an aristocratic society because life has become more convenient. The key is how to design the communication consumption of end-users – the differential between how much you recoup from the wealth concentrated among the aristocracy and how little you recoup from everyone else. A science of imperial rule built on gamification and platforms, a chain of hypocritical employment production. When considering what kind of social structure we're creating, I think the aristocratic society will become more materialistic, while ordinary people will likely enjoy virtual reality as a substitute. A world of VR for the poor and VR for the aristocracy – I feel like that kind of world is approaching.

Fujita: There's no end to this conversation! Today was truly fascinating. It was very stimulating.

<End>