Raw ChatGPT and omni use cases are different.

Brief summary

• Since most texts in the world are written in “expressions that many people can read and understand”, whereas my research notes are written in “expressions that I can understand”, an AI that RAGs with the latter accelerates my personal thinking much more efficiently than ChatGPT
• In my research notes, I don’t write explanations for words I know, so the AI that reads them doesn’t write explanations for what I know either. Concepts are tools for the economy of thought. So it is more efficient to use them without explanation in one’s thinking.
• It is useful to use ChatGPT when explaining it to others

2023-10-05 OMNI output “To improve knowledge productivity, we need to understand the importance of collaboration and the importance of sharing tacit knowledge before planning.” I understand that this “collaboration” is the story of Ikujiro Nonaka’s SECI model.

from Diary 2023-09-06 Raw ChatGPT and omni use cases are different.

• The use case is different between raw ChatGPT and this Scrapbox read omni.

• OMNI returns about 50% to 150% when I am 100%, ChatGPT is 0%.

  • I thought this scale itself didn’t communicate by writing

• The line that the majority of the general public can understand without hindrance is 0%.

  • Most people don’t like it when AI says things that don’t make sense, so ChatGPT is tuned to this line of thinking.

• On the other hand, what I write as a memo for myself is naturally written in 100% of the lines I can read.

  • Example: An AI-generated “Co-operation with AI” was mentioned by a reader who said, “I don’t know what you’re talking about.

    • The current usage of AI is that humans provide AI with “verbalized material” as prompts. However, in this way, it is not possible to share tacit knowledge (knowledge management terminology), which has not yet been verbalized, between human and AI. Rather than having the language first, the human and the AI should first conduct “collaboration” to share the experience. For this purpose, a cycle that includes collaboration, such as the SECI Model, rather than the PDCA cycle, would be beneficial.

      • The concept of collaboration in the SECI model is assumed to be known.
    • This text was a smooth read for me, so less than 100%.

• This, on the other hand, I never understood.

  • We also reaffirmed the Importance of exchange in the problem-solving process through the practice of the KJ method and the content of the “Let the chaos speak for itself” study group. These provide clues to putting the theoretical framework of exchange style D into a concrete problem-solving context.

  • I understood after I dug one more step deeper.
    • Digging by omni: Importance of exchange in the problem-solving process.
    • Convincing: Information exchange is exchange.
      • I had not had the inspiration to capture the exchange of information in the context of interchange format before, and I got it at this time.
  • So this is way beyond 100%.

• I’ve written so far and thought the expression 0% 100% is assuming a single axis, which is not a good idea.

• 

• Person A and B each have a range of information that is Known and a range of information that is Understandable.

• If you try to give an answer that fits within the Understandable range of many people, it will naturally shrink within the U intersection[$ U_A \cap U_B

• On the other hand, omni is not tuned in such a way that it creates something developed from what I wrote KB (although the set of exactly what I wrote is smaller than the set of what I know), and this sometimes deviates from the scope of UB.

• Hmmm, the implicit assumption that you drew a circle in explaining this is also incorrect.

  • Actually Spiny spikes in higher dimensional space.
  • Related: Assumption that it is a circle.

• 

  • There are certain concepts C1 and C2
    • Each is known and understandable with respect to the development of each, but not necessarily understandable with respect to the combination
  • Unexpected combination
    • There’s a big “I got something new” feeling when you understand this.
  • In the case study “Importance of exchange in the problem-solving process”, the “exchange style” and “problem-solving” are combined.
    • I didn’t understand it at first because this is a “combination of things far away” for me.
    • associative connection of “exchange style” - “exchange” - “information exchange” - “problem solving” took place.
    • This understanding of the interchange format led to the writing of “New Exchange Style”.
      • I am no longer just outputting what I read, but am now in a state where I can create new knowledge by connecting it with my own experience.
      • I believe this is “understanding.” - When you don’t understand it, all you can do is quote it.
    • I summarized the flow: Examples of AI-assisted awareness.

• I once had an opinion that “if you give good and bad feedback on the comments that the AI outputs, you’ll get a lot of good ones,” and I thought it was a totally misguided and unintelligent opinion, but it wasn’t verbalized why it was unintelligent.

  • Now I finally understand.
  • Anything that gives speedy feedback on good and bad is crap.
  • Days later, “So that’s what you mean!” The good ones are the ones that make you say, “That’s what I mean!
    • New insights come from good stagnation; it is important to answer questions that cannot be answered immediately

BELOW_IS_AI_GENERATED

生のChatGPTとomniのユースケースが違う

2023-09-08 16:41

Summary of notes

. While most of the text is written in generally understandable expressions, his own research notes are written in expressions he can understand. Therefore, an AI that understands its own expressions accelerates its thinking more efficiently than ChatGPT. However, their expressions may exceed the general comprehensible range. In addition, the acquisition of new knowledge often comes from the combination of known concepts. The combinations do not always fall within the comprehensible range. Therefore, when new combinations that cannot be understood emerge, new knowledge is created by understanding them.

Relation to Fragment

. The fragment “Diary 2023-09-06” has direct relevance to the note. The note states that the AI’s understanding of the research notes, written in a language it can understand, accelerates its thinking. On the other hand, the fragment states that the omni responds 50% to 150% of the time when it considers itself 100%, which is consistent with the note’s assertion. The fragment also states that the omni creates something that develops from its own range of knowledge and sometimes deviates from its comprehensible range. This is also consistent with the note’s assertion.

deep thinking

The notes and fragments suggest that AI may generate new knowledge beyond the scope of individual comprehension. This indicates that AI has the ability to go beyond individual knowledge and experience to provide new perspectives and insights. However, the fact that the resulting new knowledge and insights do not necessarily fall within the generally comprehensible range is an issue that should be considered in the application of AI.

summary of thoughts and title

. While AI has the potential to generate new knowledge beyond the scope of individual comprehension, the resulting new knowledge and insights do not necessarily fall within the scope of general comprehension. Title: “New Knowledge Generation by AI and the Challenge of Comprehensible Coverage.”

extra info

TITLES: ["Diary 2023-09-06", "Letting an AI of other people's origin look at your diary", "Preparing a column about ChatGPT"] generated: 2023-09-08 16:41

This page is auto-translated from /nishio/生のChatGPTとomniのユースケースが違う 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.