Iāve been using the ChatGPT version of Keichobot for a while now, so I havenāt used the original version in a while.
- Result: The idea of āClean Language as a Board Gameā was born.
summary
- Iām interested in chatbots with clean language.
- At first I saw the value in clean language.
- I saw the value in digitizing it and making it available at any time, so I created Keichobot.
- Tried to port Keichobot because I saw value in ChatGPTās language capabilities, but currently I donāt see much value in it.
- I see value in the summary function.
- The varied behavior of ChatGPT has the disadvantage of making it difficult to enforce the rules.
- Clean language is a conversation with limited options, which is different from free conversation
- More like a board game than a free conversation.
- Need to prototype and test play āClean Language as a Board Game.ā
- Once āClean Language as a Board Gameā is complete, the potential for good use of Keichobot could expand. Next action: prototype and test play āClean Language as a Board Gameā.
digest by AI
- I am interested in a chatbot using clean language to replace Keichobot.
- Variety of questions can have disadvantages that make it difficult to get people to follow the rules.
- Clean language is like a limited choice, different from a totally free conversation
- Need to prototype and test play a clean language as a board game.
- Once the clean language as a board game is completed, the good use of Keichobot could be expanded. Next action: prototype clean language and test play.
log
What would you like to see happen in this conversation? Iāve been using the ChatGPT version of Keichobot for a while now, so itās been a while since Iāve used the original version. What kind of āKeichobotā is that āKeichobotā? Rule-based decision making without using ChatGPT What kind of āChatGPTā is that āChatGPTā? Iām using a huge language model with varied output, but I canāt quite get them to follow the rules. What is the āvariationā of this āvariationā? Question variations. To begin with, Clean Language uses ā20 questionsā or ā5 basic questionsā, so variations in question wording may not be so important Is there anything else you can tell us about that āvariationā? The richness of variation may conversely be a disadvantage that makes it easier to make statements that deviate from the rules. How do you know what the āvariationsā are? Itās in the Clean Language book. What kind of āclean languageā is this āclean languageā? Clarify the userās mental picture by asking hard-to-distort questions
-
The richness of variation may conversely be a disadvantage that makes it easier to make statements that deviate from the rules.
- What do these āvariationsā look like? I thought it sounded like a chat statement, but I may have actually been mistaken. This doesnāt sound like a conversation, but a ālimited choiceā like a hand of cards in a card game or a worker placement in a board game.
-
The richness of variation may conversely be a disadvantage that makes it easier to make statements that deviate from the rules.
- Where are these āvariationsā? It is important to be human scrutinized and limited
-
The richness of variation may conversely be a disadvantage that makes it easier to make statements that deviate from the rules.
- Where are these āvariationsā? š
-
Itās in the Clean Language book.
- Is there anything else you can tell us about this āclean languageā? One of the good things that the ChatGPT version explored by not being able to do clean language well was the ability to summarize, especially the key points in the Next Action, which seems to have produced a different and better effect than clean language. What is it that you value? At first I saw value in clean language, and I saw value in making it electronic and always available, so I created Keichobot; I saw value in ChatGPTās language capabilities, so I tried to port Keichobot, but currently I donāt see much value in it; I see value in the ability to summarize, but I donāt see much value in the ability to summarize. I feel the summary function is valuable. What is the relationship between this āChatGPTā and āClean Languageā? Summary.
-
I created Keichobot because I initially saw value in clean language, and I saw value in making it available electronically at all times; I saw value in ChatGPTās language capabilities, so I tried to port Keichobot, but currently I donāt see much value in it; I see value in the ability to summarize, but I donāt see much value in the ability to summarize. I feel the summary function is valuable.
- What does this āclean languageā look like? The relationship between ChatGPT and Clean Language, the two are incompatible. This is because Clean Language is like a cooperative board game, not a natural conversation. Is that āChatGPTā the same as āClean Languageā? Is it different? It would be more beneficial to think of them as completely different.
-
ChatGPT and Clean Language, the two are incompatible. Because Clean Language is like a cooperative board game, not a natural conversation.
- Where is this āclean languageā? Clean Language as a board game is something I just came up with, so itās not anywhere in the world yet. Where is this āclean languageā? I have a vague idea in my head, it would be good to clarify this and get it out of my head.
-
Clean Language as a board game is something I just came up with, so itās not anywhere in the world yet.
- What do you hope will happen regarding this āclean languageā? Maybe when Clean Language as a board game is playable, people will better understand how to use this Keichobot well! What needs to happen to āclean languageā for that to happen? Clean language as a board game needs to be verbalized. As for āclean language,ā can it be done? I think itās a challenging assignment.
-
Clean language as a board game needs to be verbalized
- How do you know this āclean languageā? We have to repeat the process of making small prototypes and test-playing them.
-
Clean language as a board game needs to be verbalized
- Where does this āclean languageā come from?
https://keicho.netlify.app/#talk=XMZpJnByOa9qNk0vOBqU #KeichobotLog
This page is auto-translated from /nishio/ćć¼ćć²ć¼ć ćØćć¦ć®ćÆćŖć¼ć³ć©ć³ć²ć¼ćø 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. Iām very happy to spread my thought to non-Japanese readers.