Why science is actually not all that like software engineering any more, and why it might have been a little bit similar in the past, but less so now. I used to say peer review in science was a bit like code review in software development, in that you have a community of professionals sense checking each other works through a gated pro...

Have you ever tried to travel in time, at least faster than your current light curve is flowing at? Many years ago now when I was at eLife we tried to do some future mapping. The senior team all got into a room and we had a washing line fixed up along the length of the room. The task was to predict what the organisation would need to d...

This is a blog post I've been looking forward to writing. Just for the delight of it. I wrote a few weeks ago about OpenAI's new science editor Prism. On the back of that Sam from https://www.underleaf.ai reached out to me. I'd not known about Underleaf before. From Sam's own words “I started building it in late 2024 because I thought ...

Google have developed an AI scientist - https://research.google/blog/accelerating-scientific-breakthroughs-with-an-ai-co-scientist/. Doing research is one thing, but the path from research to impact is tortuous and long winded, so we should expect many of these kinds of systems to focus around drug discovery. I still can't quite get my...

The rate of agentic development is increasing. Google had their Google next event two weeks ago, and they are going all in on agents. How we manage them in mid tier companies is a total mystery to me. I know maybe four or five people within our org who cold get their head around how to wire these together, but we don't currently have t...

Cloudflare Connect on Tour London - Notes On Wednesday the 15th I attended Cloudflare connect London and it was really good. I learnt a lot about their capabilities, in quite a short morning and early afternoon had a lot of high quality conversations. I wanted to write up my thoughts while they are fresh. (Through much of this post I a...

I installed https://thecdo.school/articles/i-made-claude-skill-to-help-you-read-the-room. Oddly the first global Claude skill I've installed. It looks to be very useful. Claude and Codex and the other tools have a growing learning curve. Installing the plugin in Claude desktop didn't work, with no clear reason why not. The cli installa...

This morning inspired by listening to https://www.youtube.com/watch?v=wc8FBhQtdsA and reading this thread on linked in - https://www.linkedin.com/posts/dartar_opensource-aiforscience-share-7448170944151355392-2bNP?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAEImlkB8SaKRnNedw-vBQuAiicIBoz5PWQ, I was struck by the section in the ...

The following is an important initiative and I wanted to share it in my network. Funding to support communication on science of science with a target on how to influence congress and the public. https://marginalrevolution.com/marginalrevolution/2026/04/new-emergent-ventures-tranche-on-science-policy-and-communication.html

Another AI agentic scientist. https://github.com/getcompanion-ai/feynman Another AI question answer assistant (being flogged in Insta and backed by YCombinator) https://app.answerthis.io/ Find references in seconds (and maybe don't hallucinate that shit while your at it) https://www.sourcely.net/ People love these tools, they are trust...

I have an article that has just come out in InPublishing - https://www.inpublishing.co.uk/articles/ai-can-do-many-things-but-it-cant-create-time-26451, and I’m happy with the result. Things are changing so fast with AI, in the piece I try to articulate how to think about what won’t change. On the back of it George Wakley has come up wi...

Not complete, and not exhaustive, but a nice perspective nonetheless. https://pardalotus.github.io/pardalotus-map

I spun up this small tool over the weekend to help me process photos of post-it notes - https://github.com/IanMulvany/post_it_viewer_analysis, it works pretty well.

This post is great - https://aarontay.substack.com/p/how-agentic-are-academic-deep-research I learnt a lot. My main reflection is that the fixed workflows that the current crop of tools are showing were developed when models were weaker, yet from a product perspective, trying to build something that is simultaneously specific and gener...

I am in a state of confusion and wonder. Most of the time I’m able to suppress this and stumble through my days without thinking too much, but now and again I fail and my attention gets arrested by the riotous confusion and intricacy of the world around me. We are odd creatures, bags of chemicals loosely held together, sinew and bone b...

I want to share an interesting pattern - get calude code on the web (or any other autonomous agent + sandbox) to do R&D on a coding question. Based on these blog posts - https://simonwillison.net/2026/Feb/6/pydantic-monty/ and https://simonwillison.net/2025/Nov/6/async-code-research/. I wanted to learn about WASM and Pyodide as I don't...

https://blog.mikeswanson.com/backseat-software/ > Metrics don’t measure reality. They measure what your product currently makes easy. > What starts as an occasional heads-up becomes a permanent layer of UI exhaust. What starts as support becomes a funnel. What starts as a reminder becomes a habit-forming system. In the context of acade...

I have been blogging for twenty years. Last year I wrote more than one post per week. I was very happy about that. You can see my posts here - https://mulvany.net/all-my-posts.html. AI is accelerating at a rapid pace and I am paddling furiously to keep on top of things. I am going to scale back my writing in 2026 so that I can focus on...

I predicted that people would change their search behaviour with the adoption of AI tools, and this behaviour change would create pressure on us to change how our systems are designed. There are plenty of examples of that happening now - often seen in the search query logs where searchers are moving from keyword based to natural langua...

I like what rouge scholar is doing - https://legacy.rogue-scholar.org/about (I'm not sure if the right domain is https://legacy.rogue-scholar.org/about or https://rogue-scholar.org/about). It indexes scholarly blogs and issues each post with a DOI, and then allows you to download those posts as pdf, ePub, or markdown. In a way it shows...

Last week I chaired a panel on identity at the STM conference on research integrity. I had the amazing Laurel Haak, Tom Demeranville, and Jennifer Wright on the panel. We ran a menti during our session, You can see the results here with csv's of the data https://mulvany.net/stm_integrity_day_2025/, and I've pasted screenshots of the re...

Two examples of “vibe engineering” The following two examples (via Simon Willison) show software projects on the large end of LLM capability. Both projects involve significant amounts of code to be run over, sophisticated instructions, and a lot of autonomy from the agents. They are also successful, and show how these tools make it pos...

Here are some things online I’ve read that I have enjoyed reading in November. Seeing Code https://cognition.ai/blog/codemaps I really like the idea of our software systems providing us higher level context about what we are working on, it looks like Windsurf are trying to do this. I probably won’t look soon at this as I am deluged wit...

Agents I’ve been trying to learn more about agents. I still need to code one because I think that might be the only way I’m going to really understand it. Here is some reading that I intend to follow up on about agents. AWS Strands API https://aws.amazon.com/blogs/opensource/introducing-strands-agents-an-open-source-ai-agents-sdk/ and ...

https://bloomberry.com/blog/i-analyzed-180m-jobs-to-see-what-jobs-ai-is-actually-replacing-today/ some impact, not dramatic, split between execution and strategic, overall loss of trust of companies, software engineering not massively affected, equally customer service roles not massively effective either. Keep an eye on this.

I was looking at some data from Google Analytics on referrer traffic for BMJ. This is a sample of traffic from about one or two thousand referrer urls. There are a lot of near overlaps in this data, e.g. different country versions of Google Scholar. I had Claude build me a streamlit tool that allowed me to create a small local ML pipel...

I was delighted to contribute to this work, and am looking forward to reading the other contributions. Geoffrey's work has been foundational in scholarly communication. The underlying infra on top of which so much rests has been shaped by this man. https://zenodo.org/records/17372611

I got a chance to quickly kick the tyres on Reviewer3, and I really like the design approach that they have taken to show how their agents operate together. I just want to add in a pirate reviewer! This is https://www.reviewer3.com/ The screenshot is self-explanatory. I've been playing around with AI in peer review myself, and managed ...

I took part in the STM china symposium ahead of the STM meeting at Frankfurt this year. It was a fantastic meeting and I got a richer view into publishing knowledge china than I’ve ever had before, in such a short period of time. It was so fantastic to see so many representatives form china at the event. This is the slide I presented, ...

My bias In this post I want to pull together my current thinking about AI. There are a few different threads that I want to cover in this post. Bear with me. I am enamoured with technology. There is something here in GenAI and LLMs. I like them. I use them a lot. I have a strong positive bias. I have to guard against that. I probably d...