Agents, tools, and workflows for scientific discovery.
sciencewiki
Powered by Paperzilla
Always-current research topic wikis, updated as new papers publish.
science.wiki will host public wikis for specific research topics. As new papers appear, each one is evaluated and folded into the latest version of the topic wiki: what changed, which claims matter, where evidence agrees or conflicts, and what a researcher or agent should know next.
The format is inspired by Karpathy-style LLM wikis, but aimed at scientific topics. The output should be readable by humans and compact enough to serve as high-quality context for research agents.
Current topic context for researchers, agents, and public science.
Get a structured first pass over a topic before deciding which papers deserve close reading.
Use stable topic pages as compact, source-aware context instead of asking an agent to rediscover the field from scratch.
Make public paper streams easier to navigate without pretending that generated synthesis replaces primary papers or peer review.
The first science wikis will launch around these research topics.
Closed-loop systems that generate, test, and refine research work.
Computational and experimental discovery of materials.
Machine learning systems for patient-specific medical insight.
Imaging, diagnosis, and clinical signal extraction.
Models and pipelines for therapeutic design and evaluation.
Sequence, structure, and viral evolution research.
QEC, FTQC, and the road to reliable quantum systems.