Meeting notes between Forgejo and the Dutch government via Git commits (codeberg.org)

A web page guarding access to what appear to be meeting notes between Forgejo and the Dutch government is being blocked by an anti-scraping defense called Anubis, which presents a proof‑of‑work challenge before serving content. The site’s notice explains that Anubis (v1.21.0) implements a Hashcash‑style PoW to make automated mass scraping expensive, and that it’s a stopgap while the operator develops fingerprinting and headless‑browser detection (for example, via font‑rendering differences). The notice also warns that modern JavaScript is required and recommends disabling plugins such as JShelter that would interfere with the challenge. For the AI/ML community this is noteworthy because it highlights an escalating arms race between data holders and model builders: PoW and sophisticated fingerprinting increase the cost and fragility of large‑scale web scraping, which many teams rely on for training data. Technical takeaways are concrete—Hashcash-like PoW at the HTTP layer, browser fingerprinting via rendering quirks, and JS-dependent gating—that push scrapers toward more realistic browser emulation or legal/transparent data channels (e.g., public Git commits on forges like Forgejo). The move raises practical and ethical questions about data access, reproducibility, and the incentives to publish sensitive records in version control versus locked web pages.