SHIRO & Co. Announces 'Thought Cache,' a Local Knowledge Base to Preserve Fleeting Thoughts in the Age of Generative AI

## SHIRO & Co. Announces 'Thought Cache,' a Local Knowledge Base to Preserve 'Thoughts Before They Are Forgotten' in the Generative AI Era

SHIRO & Co. has announced 'Thought Cache,' a local knowledge base designed to accumulate fragments of individual thoughts, observations, conversations, photos, audio, and gut feelings to connect them to future editorial, publishing, and folkloric projects. Thought Cache is not a completed product but an experimental initiative currently operating in a local environment.

Thought Cache serves as an experimental knowledge base for saving, fermenting, and re-editing unorganized thoughts and as-yet-unnamed questions that arise in daily life, rather than immediately converting them into finished products.

While generative AI has made it easy to create text, images, audio, and video, individual nuanced thoughts—what humans notice, feel, and risk forgetting in daily life—are rapidly washed away within SNS, chats, cloud services, and timelines.

Thought Cache is a foundation for saving these 'thoughts before they are forgotten,' starting from a local PC, and connecting them to future writing, publishing, archiving, and folkloric observations.

### What is Thought Cache?

Thought Cache is a local knowledge base for saving, nurturing, connecting, and re-editing individual fragments of thought.

Unlike typical note-taking apps or generative AI tools that aim to instantly organize, summarize, and output information, Thought Cache focuses on saving fragments that do not have defined meanings yet.

Thoughts in Thought Cache are not finished sentences. They are seeds of inquiry, unresolved gut feelings, and small records that hold the potential to connect with other fragments later.

Daily observations, words remaining from conversations, gut feelings from photos or audio, fragments born from books or music, and questions remaining from dialogues with generative AI. These materials are accumulated as 'Seeds,' grouped as 'Beds' when necessary, and their relationships are visualized as a 'Galaxy.'

### Background: Why is 'Preservation of Thought' Necessary in the Generative AI Era?

With the spread of generative AI, everyone can create text and images in a short time. However, as the amount of generated content increases, individual micro-thoughts, gut feelings, hesitations, and observations behind it become harder to see.

Humans do not think only in finished sentences.

Inexpressible sensations. A landscape that caught one's attention. A sentence left from a conversation. A time remembered while listening to music. A sign seen on the street or the negative space in a photo. A question that hasn't reached a conclusion yet.

Such fragments have unknown value immediately. However, when revisited after some time, they might become writing, projects, research, or potential future folkloric material.

Thought Cache is a mechanism to save these 'things that haven't become meaning yet' before they are lost.

### Main Features of Thought Cache

1. Seed: Saving Fragments of Thought

In Thought Cache, daily fragments are saved as 'Seeds.'

Seeds do not need to be finished sentences. Short phrases, questions, gut feelings, fragments of conversation, or notes on photos and audio are sufficient. Saved Seeds are revisited, connected with other Seeds, and developed into drafts of articles or projects as needed.

This process is closer to a garden for fermenting thoughts than efficient task management.

2. Beds: Grouping Related Thoughts

Thought Cache features a function called 'Beds' to group related Seeds.

Beds are not just folders. They are thought nurseries for cultivating Seeds gathered around a single theme or question over time.

For example, in 'Beds' like 'Linguistic Relativity' or 'Epistemology of Writing,' multiple Seeds gather around a single question, forming a flow of thought.

3. Thesis / Anti-thesis: Leaving Conflicts Without Simplifying Questions

In Thought Cache, a 'Thesis / Anti-thesis' structure is implemented for each Bed, introducing a mechanism to leave conflicts without simplifying questions.