OpenWALDO, a new open-source AI project led by Gregory Kurtzer (founder of Rocky Linux and CentOS), launched today with the goal of building a collaborative, community-governed corpus of AI training data.
Unlike existing open-weight models that lack the code and training data needed to recreate them, OpenWALDO proposes an AI Bill of Materials that includes full training data, methods, and documentation—all publicly attributed, reviewable, and correctable.
The project aims to reduce duplicate work across AI teams globally and create a shared foundation for building auditable, transparent AI models.
What happened
OpenWALDO, a new open-source AI project sponsored by Ctrl IQ Inc., launched today under the leadership of Gregory Kurtzer, founder of Rocky Linux, CentOS and Apptainer. The project aims to build a community-led corpus of AI training data, similar to Hugging Face but focused on providing not just open-weight models but also the full code, methods, recipes, and training data needed to regenerate them.
Why it matters
Open-weight models allow users to download and run AI anywhere, but they typically ship without the training data and methods needed to recreate them. OpenWALDO proposes an AI Bill of Materials that includes full training data with object references, documents, tokens, inventory and licenses, meaning every assertion is attributed, rewritable and correctable in public. This "Know before you train" approach addresses the current industry pattern where AI teams duplicate work by assembling foundation training material from scratch across the globe.
What to watch
The project is available to individuals, hobbyists, researchers, institutions and corporations. OpenWALDO aims to provide a community-built resource for archiving best practices, allowing experimentation in public, and building auditable lines that can be evaluated collectively—creating what Kurtzer describes as a durable AI commons.
Ask the AI about this article →
OpenWALDO addresses a real gap in the current open-source AI landscape. While open-weight models have been a major step forward—allowing users to run AI on consumer hardware without proprietary lock-in—they remain incomplete without the training data and methods that generated them. The body describes a widespread industry problem: AI teams routinely duplicate effort by assembling their own foundation training material from scratch, scraping recipes piecewise across the globe. OpenWALDO proposes a solution modeled on successful open-source projects: a shared infrastructure for training data that operates at scale, with full transparency about content provenance, licensing, and attribution.
Kurtzer's leadership is significant because he has built communities around other foundational open-source projects (Rocky Linux, CentOS, Apptainer). The project's framing—"Know before you train"—emphasizes public auditability: by making training data, licenses and provenance fully visible, OpenWALDO aims to address legal and ethical concerns that can arise when models are trained on data of unclear origin or rights. The body frames this as an evolution beyond open-weight models, leveraging the proven model that open-source has used to win trust against proprietary vendors' claims of security and accountability concerns.
For example, today's edition would include:
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. Q&As are published on this page for other readers too.
Israeli startup DataAgent Ltd
SK Hynix presented a custom HBM concept at SEMICON Taiwan 2026, where compute functions are placed in the base…
Nvidia reported earnings that were both remarkable and boring, reflecting its focus on avoiding a consolidated…
Anthropic has agreed to a $35bn cloud-computing contract with Lambda, a Nvidia-backed cloud provider
The Supreme Court of Japan has included about ¥60 million in its fiscal 2027 budget request for AI-related exp…
The Consumer Affairs Agency said Tuesday it will use generative AI to analyze about 900,000 annual consultatio…
