PubTech Radar

This is the first post in Big Ideas in Publishing, an occasional series exploring concepts that could reshape research communication.

Big Ideas in Publishing: Tim Vines on AI Subscriptions How creating AI-ready research content could unlock entirely new revenue streams

I caught up with Tim to discuss this concept, which emerged from his thinking about what happens when artificial intelligence becomes a primary consumer of research literature.

open.substack.com/pub/pubtechr...

Big Ideas in Publishing: Tim Vines on AI Subscriptions How creating AI-ready research content could unlock entirely new revenue streams

Tim Vines, founder of @dataseerai.bsky.social, has been thinking about the future of research publishing for some time. His latest idea, what if publishers created AI-optimized versions of research articles and sold them as premium subscriptions?

The AI powered Library Search That Refused to Search From Clarivate's Summon to Primo Research Assistant, content‑moderation layers meant mostly for chatbots are "quietly" blocking controversial topics from being searched

[Blogged] The AI powered Library Search That Refused to Search - From Clarivate's Summon to Primo Research Assistant, content‑moderation layers meant mostly for chatbots are "quietly" blocking controversial topics from being searched
aarontay.substack.com/p/the-ai-pow...

YouTube video by Cactus Communications Rethinking Science Communication: Beyond Technology Trends

Enjoyed this podcast from @nikeshgo.bsky.social, interviewing Kent Anderson and Joy Moore, co-authors of How the Internet Disrupted Science. Lots I don't agree with but interesting challenges - need to read the book when it comes out.

music.youtube.com/watch?v=MrF6...

A leaked list shows which websites were used to improve Anthropic's chatbot — and which ones were off-limits Bloomberg was allowed, and the New York Times wasn't. Anthropic said it had no knowledge of the list and that its contractor, Surge AI, created it.

Not approved list: www.documentcloud.org/documents/26...

Approved list: www.documentcloud.org/documents/26...

“The blacklist could reflect websites that made direct demands to AI companies to stop using their content, said Edward Lee, a law professor at Santa Clara University... Some sources in the blacklist have taken legal stances against AI companies using their content“

Fascinating - sources that are ok/not ok to use for third-party RLHF training of Anthropic’s models.

More at: www.businessinsider.com/anthropic-su...

Are you ready to grow?

Business Development, GTM Strategy, Thought Leadership, Sales Coaching, Competitive Analysis

Let’s talk! jason at deboerconsultancy dot com

Are you ready to grow… but could do with a little help? Let’s talk about how I can support you, whether it’s a short-term project or a longer-term relationship #Growth #Consultancy #Sales #SaaS #PaaS #ScholarlyPublishing #InformationServices

This is a really exciting metadata development. Collaborative stewardship and distributed assertion models are the right way to go and some smart people are working on it.

If the journal is myopic about b(r, t), it treats s as follows: s = q + ε′, with ε′ ∼ N (0, σ2 s ) and σ2 s = σ2 ε (e) + σ2 b(r,t) The journal’s posterior belief about q is therefore: E[q | s] = λs + (1 − λ)¯q, where λ = σ2 q σ2 q + σ

Gotta love economists. This is from a paper about researcher views on open peer review:

The Alchemist Future AI orchestration - not AI tools - will determine winners. Hum is building Alchemist to be the orchestration layer for publisher AI.

Interesting insight from Dustin Smith @humworks.bsky.social about the need for orchestration infrastructure that connects AI models, prompts, data, and interfaces to allow publishers to harness AI: blog.hum.works/posts/the-al...

Cabells, the US-based academic intelligence firm, has announced the launch of CompassAI, an AI tool to assist researchers in selecting appropriate/avoiding predatory publications: blog.cabells.com/2025/07/23/p...

Also in the mix: prompt injection and peer review, publisher-AI partnerships, crawler paywalls, metadata infrastructure, stealth prompts in preprints, benchmarking against large models, and a fresh crop of ALPSP Award finalists.

PubTech Radar Scan: Issue 35 New launches include Sleuth AI, Featured Notebooks in NotebookLM, LLM Citation Verifier, and CC Signals from Creative Commons. Also in the mix: AI-written peer review, publisher–AI partnerships, crawl...

New issue of Pub Tech Radar Scan: open.substack.com/pub/pubtechr...

Lots in this issue, new launches include Sleuth AI, Featured Notebooks in NotebookLM, LLM Citation Verifier, and CC Signals from Creative Commons.

I'm seeing the same in o4-mini. It makes you doubt your own sanity! Correct paper info & title, a real arxiv link, but you end up at a totally different paper. I'm still seeing a lot of hallucinations when I ask for an example of something - the examples are usually great, but sadly, also fake

Perhaps more of a show & tell rather than evaluation would be easier? I quite like articles that just give examples of what each service does and leave it to the user to evaluate based on their own criteria/use cases.

RED ALERT! How to avoid misconduct and how to investigate if you have to. This post provides an update on some recent new features in Oversight and walks through an investigation. Given the pace of development of…

New blog post. How to investigate with Oversight clearskiesadam.medium.com/red-alert-ho...

Yes please. Would be interested to know if they’re any better/worse than ChatGPT, etc and extent to which you can customise the outputs.

Things will get really interesting with @cometadata.bsky.social in the mix!

It’s a sales pitch, but nicely done. I do wish they’d included a few side-by-side examples to show the contrast more clearly. Still, more publishers should be proving their value by showing it.

BMJ Best Practice put their tool up against ChatGPT and Grok in real clinical scenarios, and (unsurprisingly) their carefully curated, evidence-based guidance came out well ahead. The AIs sounded convincing, but missed the nuance, depth, and safety insights.
👉 bestpractice.bmj.com/info/us/blog/

The future of AI-intermediated content isn’t coming, it’s already here. | Florent Daudens The future of AI-intermediated content isn’t coming, it’s already here. Google’s latest experiment with The Economist and The Atlantic proves it. Google is essentially creating curated knowledge envi...

The AI doesn’t just summarise, it interprets, contextualises and even adds editorial commentary.
Are we moving from static content → searchable content → conversational content → AI-guided knowledge exploration?
www.linkedin.com/posts/fdaude...

The future of AI-intermediated content isn’t coming, it’s already here. | Florent Daudens The future of AI-intermediated content isn’t coming, it’s already here. Google’s latest experiment with The Economist and The Atlantic proves it. Google is essentially creating curated knowledge envi...

Florent Daudens’ latest post on LinkedIn highlights how Google’s NotebookLM partnerships with The Economist, The Atlantic and other premium publishers aren’t just experiments, they’re likely previews of how we’ll engage with information in the years ahead.

They binned off XML in favour of HTML (I’m still slightly unconvinced, but hey, simplicity wins), and reshaped how publishing systems get built. Coko showed the world that shared infrastructure and community-led development can actually work. Onward to Adam’s next chapter. 🚀

What began as a Shuttleworth-funded experiment grew into a portfolio of innovative platforms and projects like INK, Phenom, Editoria, Kotahi, Ketty, Wax, Micropubs, and the Open Publishing Awards, to name a few.

Ten Years of Coko(!) I saw a post on LinkedIn recently—someone congratulating me on my 10-year anniversary at Coko. I hadn’t realized it had been that long, but they’re right.

Adam Hyde’s gloriously sprawling, passionate love letter to a decade of building open, collaborative publishing tools is well worth a read: www.linkedin.com/pulse/ten-ye...

🧠 Curious how XML automation and AI are reshaping scholarly publishing? Join industry legends at @scholarlypub.bsky.social's free Innovation Showcase on July 24: customer.sspnet.org/Portal/iCore...

Are you a researcher? Are you using any AI tools to process your raw data other than the usual (chatGPT, Gemini, Claude)?

If so, what?

For white paper/company/patient info/blog type info it seems to come up with a plausible title for the item and then bungs it onto the end of the company’s main url.