Nodal Point of Thought 2023-02-19
Nodal Point of Thought 2023-02-19
Hello. What would you like to see happen in this conversation?
I had a chat with a friend about something I've been thinking about lately and we came to a lot of realizations. They are still fragmented and not connected to one another, but I have a feeling they might be connected. So I want to write it down as I speak. You be the listener.
Understood. As you said, it is a good idea to write it down first to organize our thoughts. Where would you like to start talking?
Top trading cycles and other real-world applications were made and positively evaluated. Initially, those fields were created by people interested in what can be said with mathematical certainty, but the value is found in the installation of software created by them in society, i.e., social implementation, related "No more thought without implementation." I see, so you are talking about your thoughts on social implementation. What exactly are you thinking about?
Another story, about a vague fear regarding the commoditization of a part of myself. The value of the infrastructure of exchange of goods between people is not at all damaged by LLM, which implicitly narrowed my focus and made me feel as if I were being damaged. The resolution of the concept of self needs to be raised.
One reason you might have a vague fear about a part of yourself being commoditized might be that you feel like your own value is diminished while the value of the infrastructure of goods exchange between people is high. You feel as if you are being damaged by it.
So, you mentioned the need to increase the resolution of the concept of self, have you thought about that specifically? How do you think you can get to know yourself in more detail?
I thought about it. I was, after all, good at gathering fragmentary information and creating a big, connected story. I had realized this throughout my life and had unconsciously focused on it. Most people in the marketplace are looking for stories that have been pre-compiled by others. This is expressed in words like "I want to learn systematically. On the other hand, I had a lower mental hurdle for unsystematic pieces. This difference created value.
I see that you have found that you are good at collecting bits and pieces of information and putting them together into a coherent story. While many other people are looking for stories that have been pre-compiled by others, you have a low tolerance for unorganized fragments, and this difference has created value for you.
So, what can you think about to help you get to know yourself in more detail and better define your strengths? For example, how about thinking about the situations in which you have shone in the past, what interests you, what abilities you have, etc.?
This strength will be replaced by AI within a few years. The ability to gather fragmentary information and create stories will become a commodity. This leads to the statement, "Ideas without implementation are no longer needed. The ability to create stories will become a commodity, so it is important to implement it in society. If we do not go that far, we will not be able to contribute to society.
I see that you are concerned that AI will replace your own strengths in storytelling. You feel that this may diminish the value you provide.
However, the value you yourself provide can be created by implementing it in society, even if it becomes a commodity to the point where the story is created. So it is important to contribute to society through implementation.
Then, what can you think of to implement the value you yourself provide? For example, how about thinking about how you can use your own knowledge and experience to coexist with and create synergy with AI, and how you can provide value that is needed by society?
Another story, it should have a great impact on education. Do we need a keyboard? When you want to draw a diagram and you have the Apple Pencil in your hand, it is inefficient to let go of the pen and move your hand to the keyboard to enter text. In the first place, the keyboard has made it less expensive to create digital text, so everyone is caught up in the bias to express their thoughts in text.
Okay, you are talking about the idea of keyboards in education. You are of the opinion that the keyboard facilitates text entry and creates a bias for everyone to express their ideas through text.
So what ideas do you yourself have for input methods other than the keyboard? For example, what possibilities do you see for other input methods, such as pen tablets that can draw diagrams, voice input, or handwriting input? How about thinking about how they could be incorporated into education?
For example, if you draw a diagram and give an audio explanation, the current situation is that it is transcribed as it is, but with LLM inserted here, you can cleanly separate the speech, put it in a text box, and draw out lines for the part you are drawing, and so on.
I see your idea that by drawing a diagram and explaining it with voice, it is possible to express it more intuitively, and by having LLM neatly convert that expression into text and display it with the diagram, it is possible to provide information that is easier to understand. You also seem to think that it is possible to add annotations such as drawer lines to the areas you are drawing.
So, what are your thoughts on what kind of technologies and tools would be needed to actually implement such an educational system? For example, what kind of combination of technologies, such as speech recognition or LLM development, do you think would be necessary? Also, how would you consider how this could be incorporated into the educational system?
I think it is inefficient for a teacher to speak live because it cannot be played back at double speed, and everyone should be able to play back the video at double speed at will or go back and listen to it again if they find it difficult to understand. Then, AI should convert the voice into one that is easy to listen to. The voice is like handwritten letters, and the synthetic voice is like printed letters.
I see, so you are proposing an educational system that can arbitrarily play back videos at double speed, based on your awareness that it is inefficient for teachers to speak live. And if there are people whose speech cannot tolerate the double-speed playback, it is necessary to use AI to convert their speech to an easier-to-hear format.
Also, from the perspective that the flesh voice is like handwritten text, you think that it would be easier to recognize a synthetic voice if it were spoken in a synthetic voice, since synthetic voices are like printed text.
What are some of the benefits of using synthesized speech in an educational setting? For example, how does the ability to replay a text over and over again allow students to learn at their own pace? What do you think of the technology and systems needed to use synthetic speech?
To summarize our conversation so far.
The conversation can be summarized as follows
Initially, he feared that his own strengths in storytelling would be replaced by AI.
Next, ideas for input methods other than the keyboard in education were proposed, as well as new educational systems such as arbitrary double-speed playback using video.
Finally, the benefits of using a synthetic voice that can be replayed over and over again were considered, as well as the technology and systems necessary to use synthetic voices.
---
This page is auto-translated from /nishio/思考の結節点2023-02-19 using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I'm very happy to spread my thought to non-Japanese readers.