When we started MyForest, we were examining the enormous data within the human/AI information space. We wondered: how do we maximize information bandwidth and redefine interaction between humans and AI for comprehensive content co-creation?

For a single topic, you can brainstorm with AI on an enormous canvas. But how about complex ones? Book creation? Deep research? How large would that space need to be?

How do we define information bandwidth? Everyone has their own information bandwidth. It's not about presenting as much information as possible in front of a user; more isn't always better. If a user cannot handle or navigate through it, they might give up. Successful management of information flow is crucial.

An efficient information bandwidth optimization would:

• Deliver the right amount of information to the user at the right time, allowing them to utilize it most effectively.

• Avoid hiding information in a way that the user cannot find it. State-of-the-art chat-based AI interfaces often do this, leading to frustration when information is hidden in a deep chain and can no longer be retrieved.

So, what is the information we're talking about? It's multidimensional data created around the entire content creation process. The information space encompasses research, conversations, chats, and, in the end, your content creation—be it a long story, a picture book, or a complex report.

At present, content creators have limited access to the entire information space, categorized into two types:

Linear Slice:

A simplified slice (trimmed from a deep tree) at a point of content, capturing most human/AI interactions, e.g., ChatGPT. It's extremely painful to locate or retrieve information hidden in a deep branch generated during AI brainstorming.

When I don't know "what," I'm not ready to write, and it's easier to just chat. However, access to idea space is limited and tangled, making it hard to retrospectively explore comprehensively.

Final Stop Slice:

A slice along content that only captures the final stop of human/AI interaction, e.g., Notion, Quillbot, Sudowrite, Google Docs. It doesn't capture the full co-creation journey, often losing valuable ideation and iterative processes.

Remember, the journey of creation doesn't begin with writing. It begins with ideation.

Currently, there's no application efficiently embracing multidimensional data while providing the best writing environment for accessing the idea space along content creation.

MyForest's Approach:

MyForest aims to provide a comprehensive solution that captures the full spectrum of the content creation process, from initial ideation to final output.

Human/AI interaction iteratively conducts research and ideation, creating a multidimensional space that could be challenging to navigate. We utilize the content structure to guide the human/AI interaction, serving as a bridge between the ideation graphs, and effectively slicing the multidimensional data to make it more manageable for humans.

The structured content also provides a base for fine context control on large and comprehensive projects, which offers greater transparency to humans regarding large context windows, achieves better results for AI, and requires less manual prompt tuning. MyForest AI treats the AI as a true collaborator, with the goal of minimizing manual efforts in prompt tuning and maximizing context understanding for AI.

Our vision is to create an immersive experience for human-AI interaction and embrace the nature of multidimensional data. We aim to make information easy to retrieve—slice and present effectively according to users' intentions and capacities. The information space should be transparent to all parties, in a human/human/AI team, utilizing AI collaboration tools.