OS DataServices

OSDS. Metadata for publishers and libraries.
MARC, KBART, ONIX. AI provenance, not slideware. Consulting, training, keynotes.

[OS DataServices @DataMercsDotNet]

@infodocket The transparency pledge covers "editorial and peer review". Signatories can make that checkable: Peer-review type, editor roles, corrections and AI policy as fields in the record, not prose on a web page. Our short audit list: https://t.co/Y9OwNlPD5D

[OS DataServices @DataMercsDotNet]

The new University Press Charter, led by Cambridge and Bristol, has over 100 signatories. On AI: Support attribution and keep people "connected to how scholarly work is used". Licensing librarians and press rights teams: Write that into the AI licence. https://t.co/37xO3zNEiE

[OS DataServices @DataMercsDotNet]

@JLA_information まえがき���「舵は手放さない」、目録業務では記録として残せます。MARC 21の883（メタデータの来歴）は、リンク先フィールドの値が機械生成か否か、生成方法、日付、信頼度を書けるフィールドです。どの値がAI由来か、後から確かめられます。

[OS DataServices @DataMercsDotNet]

@CludioNune67383 Exactly, you just quoted the hedging:
"First turnkey HJT line" = Firm order!
"planning to deliver" + "will scale to support more than 100 GW." That’s NOT a firm 100GW order!
One confirmed line (>€60m, H1 2027)!!!
Rest is aspirational. Don’t let FOMO rewrite the press release;-)

[OS DataServices @DataMercsDotNet]

#notfinancialadvice BUT please read the fine print: Singulus-SpaceX “100 GW” is one confirmed line (>€60m, H1 2027). Rest is “planning additional lines” and “will scale to support more than 100 GW of equipment.”
SpaceX buys the machines and makes the cells itself in Texas.

[Elon Musk @elonmusk]

SpaceX has ordered 100 GW/year of HJT solar manufacturing equipment from Singulus in Germany.

This is for Starmind to create a sentient sun.

[OS DataServices @DataMercsDotNet]

@Schyvii Lies das Kleingedruckte: die besorgen sich die Maschinen und bauen die Solarzellen selber in Texas - das ist gewiss kein 100GW Deal für Singulus!
...wie die Lemminge...

[OS DataServices @DataMercsDotNet]

Open weights, open data, open hardware, running onboard with no cloud leash.

Open Source, ftw!

That's the bar for library AI too:
Models and data that live in your house, not on someone else's meter.

Sovereignty isn't a feature - it's the architecture!

[Jim Fan @DrJimFan]

Introducing GR00T 2: our most advanced, open System One model for robotics. It runs entirely on your robot’s onboard compute, fully untethered, and you own every weight of it. GR00T 2 features:

1. Long-horizon tasks: dexterous manipulation and whole-body control chained over 20+ steps and 5+ minutes, such as performing chemistry lab experiments or assembling model airplanes end to end.
2. X-Adapt: adapting from easy data sources (human wearables and simulation) and deploying zero-shot to diverse real robots.
3. Unified Cognition: native integrations with frontier agents for advanced physical reasoning, robust error recovery, and safety monitoring.

It's an open-source buffet! We are releasing not only the model weights and X-Adapt recipes (coming December), but also our agentic framework, simulation engine, annotation engine, hardware design, and even training data (100K hours, that's ~11 yrs of experience), all available today!

Everything that moves will be autonomous. Anyone can become a roboticist.

Lock in, build with us!

Deep dive in thread:

[OS DataServices @DataMercsDotNet]

Tamil Nadu: The Madras High Court has ordered a criminal case over allegedly AI-generated public library books with invented authors; one was falsely credited to economist Jean Drèze. Acquisitions librarians: Check ISBN and author before paying. #libraries https://t.co/u650IjIzSg

[OS DataServices @DataMercsDotNet]

@timisstuck Agreed. Editorial offices: LSA already publishes review files (this one's is up), so reviewer disclosure could sit there as fields: tool, task, which human checked it. The draft COPE, STM and International Science Council standard asks authors the same. https://t.co/ogq91RB6Z7

[OS DataServices @DataMercsDotNet]

@dellganov 書誌側には既に置き場所があります。JAPAN/MARC（MARC21フォーマット）の084は、同一分類法の分類記号を「重出する場合」は$aを繰り返すと明記。仮想書架を請求記号ではなく084から並べれば、一冊で123と928の両方に立てられますね。

[OS DataServices @DataMercsDotNet]

@mie_furu 英国白書とJPROの比較、興味深く拝読しました。刊行112日前までにBasic基準を満たした本は平均販売数が250%高い（白書自身は相関と注意）。次は「登録率」から「項目の充足率」の公開へ、ですね。読者にもAIにも届くのは中身の項目です。

[OS DataServices @DataMercsDotNet]

@researchinfo For acquisitions librarians: COUNTER's AI best practice (updated 23 Oct) separates Agent from Regular use and adds retrieval counts beside AI_Investigations, which count only excerpts kept for the answer. Ask vendors when they will report both. https://t.co/2X1qcpjJ70

[OS DataServices @DataMercsDotNet]

After a row over AI firms bulk-buying books, the Korean Publishers Association signed MOUs with five AI data firms to supply book data for training. Rights holders' consent first, fees per contract. Rights managers: Record consent as metadata. #publishing
https://t.co/KArDK9owM6

[Charles Hoskinson @IOHK_Charles]

@dhh Using it every day and even making an Elm-based window system for it called Warlock https://t.co/rBQFBN8yNd

[OS DataServices @DataMercsDotNet]

@RetractionWatch One article with a governed version history: the kit mostly exists. NISO's revised Journal Article Versions (RP-8-2026) adds optional MAJOR.MINOR.PATCH numbers and reader-facing status labels; Crossref has 12 update types. Missing: the policy. https://t.co/xsLkD9Nx6Y

[OS DataServices @DataMercsDotNet]

@caliljp 目録のヨミを正解とし、LLMは照合の手がかりに留める設計、素晴らしいです。誤り0.1%、1万件で約7円。図書館が長年積み上げたヨミの良い使い道ですね。OPACへのAPI提供と利用規約の公開も楽しみにしています。

[OS DataServices @DataMercsDotNet]

@tandfonline Timely for Indonesia: DOAJ hears Indonesian colleagues plan to align with the new Principles of Transparency 5.0 (COPE, DOAJ, OASPA, WAME). They ask for AI policies for authors, reviewers and editors, and disclosure of AI licensing. Homework: https://t.co/90f4UlFxT6

[OS DataServices @DataMercsDotNet]

Cataloguers, take a bow. Japan's Calil released Furikake today: furigana reading aids for catalogue titles and authors, built from readings cataloguers recorded. AI only aligns them. Errors: 0.1%, vs 2.9% for an AI model given the same data. #metadata
https://t.co/pBG3B3ByLt

[OS DataServices @DataMercsDotNet]

The publishers built the machines a reading room. Cambridge, Johns Hopkins and Princeton open a licensed AI gateway. Oxford University Press tests one with Harvard and Oxford librarian Richard Ovenden. Amazon takes author IDs. Plus China.

Weekly wrap: https://t.co/90f4UlFxT6