Last week I had the opportunity to attend an executive round table on "securing the agent workforce", and then later in the week I had the chance to have lunch with one of the co-founders of one of the hot AI startups, he was doing a tour of London, but is based in the US. The roundtable was organised by HotTopics https://hottopics.ht/...

I've recently been interviewed on a few podcasts about, of course, AI. I gave the keynote at publisherspeak earlier this year (talk - https://mulvany.net/talks/2025-05-publisherspeak-keynote-london/annotated_talk.html), and soon after sat down with Sowyma, and it was a really fun conversation. https://www.youtube.com/watch?v=4BiZ-8IsEOs

I wrote three posts about vibe coding this week. https://world.hey.com/ian.mulvany/notes-on-vibe-coding-1-this-is-pretty-amazing-tool-to-help-my-son-learn-about-circles-e849677c - in which I show an app I built for my son. It's neat, and is entirely deployed via a Claude artefact. https://world.hey.com/ian.mulvany/notes-on-vibe-coding-...

I’ve been doing a ton of vibe coding. A ton. I have a fairly large app that I’ve built with significant functionality. For my money it is as production ready as many things out there. I've been really following two modes - make it as robust as possible vs just make the thing already. For making it as robust as possible I have some patt...

Over the summer I was trying to figure out how to share a small flask app that I built. I wanted some colleges to see it, but it was sitting on my hard drive. I had an idea - maybe I could convert this local python app into an Atom app? This would allow my colleagues to efficiently run it as a desktop app. The app has some screens that...

The truth is we don't know yet. This morning I read a post by Jerry Normann - https://joincolossus.com/article/ai-will-not-make-you-rich/, arguing that AI is going to be more like shipping containers than like microprocessors, and as a result value creation will be diffuse, will flow to consumers, but will not create any large value fo...

My wife is an academic, her university had a course on about use of GenAI, here are some of the tools that were used on the course: • https://www.humata.ai/ • https://scispace.com/ • https://consensus.app/ • https://www.joinsuperhuman.io/tools • https://scite.ai/ • https://elicit.com/ The founder of https://www.world-brain-scholar.eu/ ...

Yesterday I made a mini app for my son to learn about circles. He's 13 and is doing equations for circles in school. I thought I'd tell him about conic sections, and the I thought, could I get an app built that does that. After a little ping ponging between Claude and GPT, I got something pretty neat. You can play with it here - https:...

I have been a strong proponent of the idea that we cannot identify LLM generated text on the basis that I could push any instruction set into an LLM to get it to mock a style, or take specific instructions into account. This post from Adam Day - https://clearskiesadam.medium.com/genai-detection-that-actually-works-edc562581fed, led me ...

found on 2025–07–12 Reflections on OpenAI by Calvin French-Owen This is an insightful read, but what makes it interesting for me is the question around what abstractions do for us. Here are some thoughts from me. This post inspired me to spend a bit of time this morning learning about Pydantic and FastAPI. I know a lot of python, but I...

I started listening to an episode of the rest his history this morning about the Irish War of Independence. In the podcast they mentioned that there were some earlier episodes about Ireland and I was quite interested in trying to find them, but the experience of looking through the past episodes on the Podcast app was terrible. So, wit...

On Monday I hosted a panel about how to create infrastructure that can support a trustworthy literature. I had an amazing set of panelists: • Kaitlin Thaney, Executive Director of Invest in Open Infrastructure • Hylke Koers from STM, and one of the leads of the STM integrity Hub • Jimmy Wales, founder of Wikipedia While I am sure that ...

I’ve been at the Royal Society conference on the future of scientific publishing at the start of this week, and I had the immense pleasure to chair the panel on infrastructure. I want to note my very first impression of the meeting. I have north of 20 pages of notes to work through, and some really interesting recommendations and sugge...

This post is a really important one for those of us working in scholarly publishing - https://gradientflow.substack.com/p/a-new-framework-for-ai-knowledge It talks about distinc patters for how we might interact with agents for knowledge work. The post describes the scholar agent, analysis agent, and facilitator agent patterns - go rea...

I had a couple of followups to my blog post from last night, so here is a quick update to that post. Dave Flanagan has created a plugin to Simon Willisons llm tool that checks references from LLMs against Crossref https://github.com/DWFlanagan/llm-citation-verifier Chris Reid pointed me towards https://www.undermind.ai/ and https://pla...

I don't know if you know the Avengers movie from 2012, our hero are defending the city from an alien onslaught, they are in a desperate situation and need to have the help of Hulk, but Hulk's alter ego Bruce Banner arrives on the scene. Captain America says to Bruce "Now might be a really good time for you to get angry" "That's my secr...

What this means is that you can get Claude to create micro apps or websites, and now share them, and allow people to interact with them. This is a lot of fun, but could also be a productivity hack within organisations. Here are some of the ones liked when looking over them after the announcement. code as art https://claude.ai/artifacts...

Documentation | Kuzu by Kuzu “Kuzu is an embedded graph database built for query speed and scalability. It is optimized for handling complex join-heavy analytical workloads on very large graphs, with the following core feature set: Property Graph data model and Cypher query language Embedded (in-process) integration with applications C...

I have continued to collect interesting links but noticed that I've not posted out any since the start of March, so I'm going to work through these over the next few weeks. This is the first batch. Taming LLMs found on 2025–03–02 Can peer review accolade awards motivate reviewers? A large-scale quasi-natural experiment | Humanities and...

I’ve finally gotten around to reading this excellent piece https://scholarlykitchen.sspnet.org/2025/02/27/creating-the-publishing-platforms-that-next-gen-professionals-expect/ I really recommend it. I like this a lot and I don’t have a huge amount to add, other than perhaps these thoughts. There are a lot of potential features offered ...

A nice write up of a survey that springer nature conducted asking researchers how they felt about research assessment - https://www.springernature.com/gp/librarians/the-link/research-management-blogpost/research-assessment-insights-6600-researchers/27769018 However where does the agency in research assessment lie? Is it not with resear...

https://www.nature.com/articles/d41586-023-04043-w Nature reports that deep seek has discovered some results that exceed the ability of people to have created! Hurrah, so now we can sit back and let the AIs take over the act of science? Not so, not so at all. For these devices have no agency, they are just tools. It would be akin to sa...

https://stitch.withgoogle.com/ is a gen ai design prototyping tool. I’ve just given it a quick test to design an integrity dashboard. I didn’t provide much detail or much context. It was a bit slower than I expected, and I didn’t find the design to be particularly amazing. I’ve found Claude and Claude code to be very good at creating U...

This came up in a conversation with a colleague today, and also at a round table conversation at a conference earlier this week. I am mostly re-sharing links from Simon Willisons blog, but they are all worth reading. MIT Tech Review: https://www.technologyreview.com/2025/05/20/1116327/ai-energy-usage-climate-footprint-big-tech/ Conclus...

A colleague pointed me to this yesterday - https://gemini.google/overview/gemini-in-chrome/. (Of course there is so much happening at the moment that it is very hard to keep up). As well as this we also had yesterday the announcement of Jony Ive joining OpenAI - video here - https://openai.com/sam-and-jony/. We also had Google preview ...

I gave a talk at the publisherspeak event yesterday here in London. I wrote a somewhat tongue in cheek post at the beginning of the year about what might be true about scholarly publishing (https://world.hey.com/ian.mulvany/what-things-might-be-true-about-scholarly-publishing-11d227f3), and I've been wanting to put a more constructive ...

In reading about ORCID trust markers I found the information on the ORCID site to be a little hard to parse, so I sent everything I could find over to Claude and I had Claude create some united documentation for me. This is a lay overview of trust markers: https://claude.ai/public/artifacts/809826cf-3941-45cf-86b8-dbd38264b2ad And this...

I'm working on a talk for next week. Claude and ChatGPT helped me make this quite neat timeline of events in the open access movement: https://claude.ai/public/artifacts/a3e89cb7-c1d9-42a1-bd72-0ac161b39d17.

I fed in my sketch notes to GPT to ask it for a title for my talk, and it came up with a really interesting set of options, I'm actually genuinely impressed by these potential titles. I'm not normally into posting pure AI generated content, but these titles do a good job of capturing some of the aspects of my notes. Now I just have to ...

Hacker news threw up a story today: https://www.theregister.com/2025/05/08/google_gemini_update_prevents_disabling/ Gemini's latest model update has made the model refuse to discuss topics like sexual assault and rape. A developer is working on a platform to help victims, and their current app is not not working. It seems like there wa...