First ask why.
Start with causes, conditions and goals so actions have a basis. Do not treat a sentence as isolated from its situation.
Start with why something happens, work through the actions, then examine the result. With its own process algorithms, RelVectors explores an independent path for AI. Its core process computation runs on an ordinary CPU without requiring a GPU.
Let machines do more than keep talking.
Let them keep thinking through how things unfold.
We study the relationships between causes, actions and results, organizing change into processes that can be computed, traced and checked.
Start with causes, conditions and goals so actions have a basis. Do not treat a sentence as isolated from its situation.
Design core computation around relationships and processes. The current core does not require a GPU, letting us take early steps on ordinary computers.
Start with practical tasks such as software development plans, discover gaps, collect examples and gradually develop broader capabilities.
Describe the software, its users and its essential features. Use the planner to organize goals, steps and acceptance checks, then continue with your chosen AI or developer.
Write a free development plan ↗Checking the trial service…
“I want inventory software for a small shop to track purchases, sales and stock. First outline development steps and how to check the results.”The trial uses the site's existing planning service. It produces a plan; software still needs development, testing and acceptance.
No second agent to install — connect the RelVectors algorithm to the assistant you already use.
Our independently developed algorithms are still growing. We welcome people interested in studying, testing and building this approach, giving AI rooted in Chinese thinking more reliable and broader capabilities. This is a research goal and a sincere invitation.
Explore relationships, processes and task organization. Bring questions we can discuss and test.
Bring real expressions and tasks to help identify what the model does not yet understand.
Connect algorithms to usable products and improve implementation, testing and operation.
Try real tasks and share the process and results so improvement has evidence.
People who join the development of RelVectors get their share of the dividends of the AI era.