About Kaikim Labs
Kaikim Labs is the public research lab of DeblockX Labs, written by Kai Kim. It publishes falsifiable hypotheses, dated predictions and experiment records about emerging markets, built from live data the lab's own products collect and corrected in public.
Everything published here is designed for four audiences at once: humans, search engines, LLMs, and future versions of ourselves. Documents follow a consistent structure of problem, observation, hypothesis, evidence, counter arguments, and what would prove the idea wrong, so that every claim stays traceable as it evolves.
The archive is markdown native and git native. The repository is the single source of truth, and the full history of every idea is preserved in version control. Each document carries a confidence score and an evolution log recording what changed and why.
Who writes this
Kai Kim, founder of DeblockX Labs, a data and AI engineering company in Seoul. Kaikim Labs is the research arm of that company. The products the company runs are the lab's instruments: they collect the data, and the lab writes down what the data says and how confident it is.
- SwanBluechip: Hourly perp DEX, wallet and stablecoin data desk. Source of most market readings.
- Birdfury: Open mission network for humans and AI agents. Source of agent payment experiments.
- Algotraction: AI search visibility measurement. Measures whether this research gets cited.
How documents are produced
Documents are drafted with large language models from live data the instruments collect, then read, corrected and signed by Kai Kim before publication. Every number names its source and the time it was read. Every claim states what would prove it wrong. When a claim fails, the document is not deleted: its status changes, the changelog records why, and the confidence score moves. The method is public so that the archive can be checked, not just read.
Contact: [email protected]