iiinova — where your digital self keeps evolving

Infonomic @infonomic.io · 01/09/2026

Closest to home: on 14 Aug textbook authors sued OpenAI. Adoption is an institutional decision, so a substitute only has to satisfy a committee. Far easier harm to show. In short: not whether training is legal, but where you got it and whether harm can be proved. #scholcomm #ai

010

Infonomic @infonomic.io · 01/09/2026

Neither question is settled where it counts. No appeal court has ruled on any of it. The first to hear argument was the Third Circuit on 11 June, in Ross, pressing on transformativeness and market harm. No decision yet. NYT v. OpenAI still has no trial date.

100

Infonomic @infonomic.io · 01/09/2026

Then harm, which is where most cases actually turn. In Kadrey v. Meta the authors lost because they had not proved market harm, not because training is lawful. Thomson Reuters v. Ross failed for the mirror-image reason: the copying produced a directly competing product.

100

Infonomic @infonomic.io · 01/09/2026

That sequence has been said out loud. At an AI class at Stanford in 2024, Eric Schmidt described the play for an AI startup: have the model build your competitor, take the content, and hire lawyers to "clean the mess up" if it works. Bartz is the bill arriving.

100

Infonomic @infonomic.io · 01/09/2026

Provenance first. On 20 July a US court approved the $1.5bn Bartz v. Anthropic settlement, plus destruction of the pirated files. The ruling split it cleanly: training on lawfully acquired books was fair use, keeping a library of pirated copies was not.

100

Infonomic @infonomic.io · 01/09/2026

Three years into the AI copyright fight, nobody has settled whether training on copyrighted work is legal. But the cases point somewhere more specific. Courts are asking two narrower questions: where the material came from, and whether anyone can prove the output caused harm.

Series card, no. 4 in a series on attribution, consent and original content. Text reads: Where you got it. Whether it hurt. Little else is settled.

100

Infonomic @infonomic.io · 15/08/2026

Back from our first #IFLA #WLIC #WLIC2026. A great event with many excellent sessions, talks, posters as well as incredibly kind and helpful delegates. Hoping we can go to #WLIC2027

000

Infonomic @infonomic.io · 03/08/2026

Next: FORRU moves to Byline CMS, and the reason is search. A research archive is mostly attachments. Papers, chapters, field manuals, reports. Search that reads only titles and abstracts is searching the label on the box.

100

Infonomic @infonomic.io · 03/08/2026

What changed: 📚 Requests answered with a link, not a scan. Search state lives in the URL, so staff send the item asked for plus everything related to it. 🔍 First page of Google for most forest restoration topics, often first. 🌏 Bio pages in Thai and English.

100

Infonomic @infonomic.io · 03/08/2026

FORRU-CMU (Chiang Mai University) does the research behind restoring tropical forests in Thailand and SE Asia. The work was never the problem. Reaching it was - for the public, for donors, and for their own staff trying to build on what they had already published.

120

Infonomic @infonomic.io · 03/08/2026

Thirty years of forest restoration science, and for a long time the only way to get a paper was to ask someone to scan it and email you a PDF. How we turned FORRU-CMU's research archive into a searchable public library, and what we are building next 🧵👇

100

Infonomic @infonomic.io · 25/06/2026

Two caveats. Published isn't honoured: a signal only works if AI firms respect it. And none of it brings back the visit. For researchers and NGOs these standards defend the boundary, not the relevance. Necessary, not sufficient. Next: what the courts have actually settled. #scholcomm #ai

000

Infonomic @infonomic.io · 25/06/2026

The IETF is standardising one shared "Content-Usage" signal (its AIPREF group) to replace today's patchwork. It's real but not yet finished. Alongside it, C2PA Content Credentials add cryptographic provenance, and the EU AI Act's machine-readable marking rules start on 2 Aug 2026.

100

Infonomic @infonomic.io · 25/06/2026

Licensing is next. Really Simple Licensing (RSL), from the team behind RSS, became an official standard in Dec 2025. It turns robots.txt's yes/no into machine-readable terms, including pay-per-inference. The RSL Collective bundles small publishers together; 1,500+ orgs have signed on.

100

Infonomic @infonomic.io · 25/06/2026

The fastest mover is infrastructure. Since July 2025 Cloudflare blocks AI crawlers by default on new sites. Its Pay Per Crawl lets you allow, charge or block each one, and it now returns over a billion "402 Payment Required" replies to AI bots a day. "Do not train" is now a free default.

100

Infonomic @infonomic.io · 25/06/2026

Posts 1 and 2 in this series laid out the problem: original work ingested without credit, and AI answers replacing the visit to the source. So what is actually being built in response? More than you might think, and less than you would hope. A short tour of the standards.

110

Infonomic @infonomic.io · 21/06/2026

🤖 AI-assisted text + richtext editing across collections. ⚡ Faster: tagged data cache, prefetch, lighter bundle, dynamic sitemap. Coming Next: hasMany Relations, Search and MCP.

000

Infonomic @infonomic.io · 21/06/2026

Auditability: document-level audit log, per-document history, and a system-wide Activity area. 🌍 i18n: six new admin languages (ES, DE, IT, zh-Hans, KO, joining EN/FR), content-locale resolution, hreflang, and per-page locale switching.

100

Infonomic @infonomic.io · 21/06/2026

📝 Markdown surface (for humans and agents) A .md version of every published page, served via .md URLs, rel=alternate links, and Accept: text/markdown, plus llms.txt and a URL index. The editor speaks Markdown too: source toggle, as-you-type shortcuts, table + admonition transforms.

100

Infonomic @infonomic.io · 21/06/2026

📁 Document Trees (new) Hierarchical, single-parent ordered trees via a simple tree: true flag. Drag-to-reorder + drag-to-re-parent in the admin, auto-placement, self-healing of unplaced docs, and clean public URLs for HTML and .md. Our docs are now a live tree.

100

Infonomic @infonomic.io · 21/06/2026

Byline is coming along and definitely sucks less than a month ago. 🚀 A new website with much better docs (link below), built with Byline + TanStack Start + Base UI. A rundown of what shipped this month 🧵👇 bylinecms.app #tanstack #react #opensource #headlesscms #contentmanagement

bylinecms.app

Byline CMS

A developer-friendly, open-source headless CMS — built with versioning, editorial workflow, and content translation.

110

Infonomic @infonomic.io · 16/06/2026

An AI answer that draws on research without naming the institution doesn't just lose them a visitor. It makes them disappear in plain sight while their work is in active use. Next: what the standards bodies are building in response. #scholcomm #openweb #ai

000

Infonomic @infonomic.io · 16/06/2026

For ad-funded publishers this is a revenue crisis, severe and, for some, fatal. For researchers, NGOs and archives it's different: their credibility depends on being recognised as the source. A paraphrase without attribution breaks the chain that makes the work legible as knowledge.

100

Infonomic @infonomic.io · 16/06/2026

Only 1% of users click the citation links inside an AI Overview. Ahrefs: position-one click-through for informational queries (where knowledge matters most) fell from 7.6% to 1.6%. Being ranked first is no longer enough to be found.

100

Infonomic @infonomic.io · 16/06/2026

Similarweb: zero-click searches (queries ending with no click to any website) hit 69% by May 2025. Above 80% by 2026. Pew Research (68,000 real queries): click-through halves when an AI summary appears. 8% versus 15% without one.

100

Infonomic @infonomic.io · 16/06/2026

For twenty-five years the web ran on a deal: publishers let crawlers in; search engines sent readers back as traffic. Nobody wrote it down. AI answer engines have ended it unilaterally. Here's what the data shows, and why it hits researchers and NGOs differently than news sites.

100

Infonomic @infonomic.io · 10/06/2026

Some machinery is emerging: collective licensing to bundle the long tail (RSL Collective, TollBit, ProRata), and machine-readable rights signals that now carry legal weight in the EU. Early and unproven, but the direction is set. More in posts to come. #scholcomm #openweb #ai

020

Infonomic @infonomic.io · 10/06/2026

For researchers, NGOs and archives, credibility is the asset. The real risk isn't lost ad clicks. It's invisibility and mis-attribution, as AI answers quietly replace the visit to the source. A paraphrase with no name attached makes your work disappear in plain sight.

100

Infonomic @infonomic.io · 10/06/2026

When Taylor & Francis licensed academic content to Microsoft for a reported $10M, authors were neither consulted nor paid, and could not opt out. The backlash wasn't mainly about money. It was about consent: being used, mis-cited, or stripped of context.

100

Infonomic @infonomic.io · 10/06/2026

Their work is ingested all the same. Independent publishers, academics and cultural archives sit in a long tail that is too small to sue and too small to license individually. No seat at the table, no opt-out, and so far, no cheque.

100

Infonomic @infonomic.io · 10/06/2026

In the AI copyright fight there are two kinds of rights-holders: those big enough to sue, and those big enough to sign. The New York Times is suing. News Corp signed for a reported $250M. Most of the people who actually create original work (researchers, NGOs, archives) are neither.

100

Infonomic @infonomic.io · 09/06/2026

Native markdown endpoints coming soon!

000

Infonomic @infonomic.io · 09/06/2026

All six locales have correct <html lang>, hreflang alternates, canonical URLs and sitemap entries, reflecting the published locale rather than the source. Built on TanStack Start, Lexical, and Base UI. bylinecms.app/zh-CN/about-... #i18n #headlessCMS #collections #glam

bylinecms.app

Byline CMS - 关于 Byline

Byline 是一个建立在三大支柱之上的内容管理系统——版本控制、工作流和内容翻译——旨在协同工作而无需权衡。我们致力于真正的数据所有权，并以开放方式构建。

100

Infonomic @infonomic.io · 09/06/2026

Byline CMS now serves its about page in six languages. English and French are full interface locales - 'sticky', so the surrounding app and admin chrome follow you as you navigate. Spanish, German, Chinese and Thai are content-only: the page translates, but the interface stays put.

100

Infonomic @infonomic.io · 08/06/2026

So the HTML doc and the text/x-component payload collapse into one cache entry — whichever's cached first wins. Fix: a Cache Rule to bypass cache on RSC requests. Watch the order: Cache Rules are "last match wins," unlike WAF rules. #nextjs #cloudflare

010

Infonomic @infonomic.io · 08/06/2026

TIL: content negotiation at one URL when your CDN ignores Vary headers looks like this 👇 The Next.js RSC flight payload, dumped straight onto the visitor's screen. Ouch. Next sets Vary: rsc correctly — Cloudflare ignores Vary on everything but Accept-Encoding. 🧵

100

Infonomic @infonomic.io · 20/01/2026

We still believe in the open web — and in curated public knowledge. Because whether you love AI, hate it, or feel uncertain about it… AI runs on content. #collections #archives #knowledgemanagement #content #ai infonomic.io/platforms/co...

infonomic.io

Collections Management - Infonomic

Affordable collections management and digital libraries with rich metadata, controlled vocabularies, excellent multilingual support and full ownership.

011

Infonomic @infonomic.io · 15/06/2025

Introducing our "Giving Back" program. If you're a non-profit, civil society group, or NGO we may be able to help with pro-bono consulting, substantially reduced fees, or even fully funded solutions - Learn more ➜ infonomic.io/giving-back #NGO #non-profit #civil-society

infonomic.io

Giving Back - Infonomic

Giving back to support social causes that make a difference.

001

Infonomic @infonomic.io · 23/05/2025

Very pleased to have been able to support Pratthanadee. pratthanadee.org A great organization, working hard to support underprivileged women and girls in Thailand. #gender #women #thailand

pratthanadee.org

Pratthanadee Foundation - Home

Committed to advancing girls' education and promoting gender equality through impactful educational programs and mentorship.

032

Infonomic @infonomic.io · 13/11/2024

We've just removed the social share link and icon to Twitter / X from all of our websites and client properties.

010