After chatting with ChatGPT, I asked him to summarize the chat in bullet points at the end, and I looked at the summary and made minor changes.
summary
- Storing "dead text" in Scrapbox promotes rediscovery and continued use of the text.
- Some users perceive value in the original text itself, so destructively "[starting a fire (with flint, by rubbing sticks together, etc.)" is a reduction in value and is psychologically resistant.
- I have imported articles from Hatena Diary in the past and occasionally cite them. ref Import Hatena Diaries to Scrapbox with past dates.
- I believe that storing in Scrapbox will lead to high value in the future by having my intellectual production in one place and easily accessible for reference from AI. The #OnePocketOnePrinciple
- Dialogue with Books, I was looking for a way to understand the information in books without holding myself back too much.
- We believe that we can gain a better understanding of books by interacting with them in a way that the language model emulate me and I emulate the author of the book.
What would you like to see happen in this conversation?
I'll tell you a story that isn't coherent, so I hope it takes shape.
How do you hope it will take shape? Do you have any specific goals or directions?
It could be a Scrapbox page. I was thinking about putting things into Scrapbox that are often considered "dead text" like books, past Tweets, meeting minutes, chat logs, etc.
You are interested in collecting past "dead texts" in Scrapbox. What types of "dead texts" do you plan to collect? What are your reasons for collecting them and what do you hope to do with them?
Dead text is initially link-poor. So it is by search that they are rediscovered. What happens when you read a search hit and find it interesting? If you try to retain the original at this time, part of it will be quoted and copied into a new page, and the original page will remain unchanged. This is typical dead text behavior. On the other hand, Scrapbox's standard cut-out function removes selected parts of the original text and moves them to a new page, a design decision that keeps Scrapbox from becoming a dead text warehouse, but I wonder if actual users would choose "don't cut out"?
I see, you believe that using Scrapbox to store "dead text" will promote rediscovery and continued use of the text. I also understood that Scrapbox's cut-out function is designed to transfer information to a new page, but does not leave information on the original page, making it less likely to become a warehouse of dead text. However, are you saying that you are concerned that actual users may not use the cutout feature and leave it on the original page? Do you have any ideas as to why that might happen?
Certain types of text are valued by the user for the fact that they are in their original form. For these users, "cutting out" is an act that reduces the value of the original text, and there is psychological resistance.
I see, so for some users, there is psychological resistance because the original text itself is considered valuable and they feel that cutting it out would diminish its value. In that case, what types of text would you expect to collect in Scrapbox? Also, what are your thoughts on the benefits of storing them in Scrapbox?
There are advantages to being in one place and being the subject of a collective search, so we want to put "everything" in the search box, and there are advantages to that.
I see, so you feel that storing a variety of texts in Scrapbox is beneficial because it allows you to manage them all in one place and makes them more searchable. Then, are there any restrictions on the types of text that can be stored in Scrapbox? Also, do you have any specific goals you would like to achieve by storing in Scrapbox?
Specifically, I've imported all of my Hatena Diary articles from the past, and I occasionally unearth and quote from them. Since I recently exported my past statements from Twitter and Slideshare has made a change to put ads between slides, I'd like to consolidate those contents into Scrapbox
Okay, specifically, I have already imported Hatena Diary articles in the past and have quoted them occasionally, the recent deterioration of Slideshare's advertising has made it difficult to retrieve slides from Slideshare, and I have been using Twitter's past statements. I have found that I would like to export them. What benefits do you see in centralizing these contents in Scrapbox? And what specific goals are you trying to achieve by storing them in Scrapbox?
Based on recent experiments with Scrapbox in Bing AI, I believe that having my intellectual production in one place and easily referenceable by AI will lead to high value in the near future.
I see, you believe that centralizing one's intellectual production in Scrapbox, which can be easily referenced by AI, will be of high value in the future. Is that a decision based on the fact that there are experiments with Bing AI to target Scrapbox? Can you tell us more about such experiments? Also, what new ideas and discoveries do you think you will gain by consolidating your intellectual production into Scrapbox?
For example, when reading a book, if you have a strong sense of self, you will omit information that does not fit your own framework of thought, making the reading itself futile. On the other hand, if you try to emulate the author's thinking, you can acquire knowledge more efficiently, but it tends to be a concept that is not connected to your own knowledge and is not grounded in the world. I was wondering if this problem could be solved through dialogue with books, but instead of the language model doing the book side, why not have the language model emulate me and I emulate the book's authors and have a dialogue with them?
Okay, so you feel that you need to interact with books to deepen your own knowledge, but you were addressing the issue that if you suppress yourself and follow only the information in books, you may miss information that does not fit your way of thinking. You also seem to be of the opinion that the language model emulates itself and that by interacting with books, you can gain a better understanding of them. What advantages do you see in adopting such a method? Also, what specific steps do you think would be necessary to make this happen?
Summarize in bullet-point style what I said in this chat.
Using Scrapbox to store "dead text" promotes rediscovery and continued use of the text.
- For some users, there is psychological resistance because the original text itself has value and they feel that cutting it out will diminish its value.
- I have already imported Hatena Diary articles from the past, and occasionally cite them.
- By storing in Scrapbox, I believe that having my intellectual production in one place and easily accessible for reference from AI will lead to high value in the future.
- ...interacted with the books and was looking for a way to understand the information in the books without holding myself back too much.
- We believe that the language model emulates itself and interacts with the book to achieve a better understanding.
This page is auto-translated from /nishio/思考の結節点2023-02-23 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.