Adrian Kirwan

I wish Dick Cheney was alive, so he could watch this and die again from an aneurysm

I actually laughed out loud at this one.

Fabulous PhD opportunity at Queen's University Belfast and Armagh Observatory:

Observing the Heavens from the 'Periphery': Astronomy in Ireland 1640-1830

www.qub.ac.uk/courses/post... Deadline 13 Jan #histSTM #histastro

This is such a great project!

“If you want to know how a model will behave, the only way of doing it is to run it,” Mensch also suggested. “You do need to have some empirical testing, what’s happening. Knowing the input data that has been used for training is not going to tell you whether your model is going to behave well in [a healthcare use-case], for instance. You don’t really care about what’s in the training data. You do care about the empirical behaviour of the model. So you don’t need knowledge of the training data. And if you had knowledge of the training data, it wouldn’t even teach you whether the model is going to behave well or not. So this is why I’m saying it’s neither necessary nor sufficient.”

flat out lies. these tech ceos take regulators and the masses for fools

when you wear a size 9.5 shoe but the 8.5 was on sale

What Rosalind Franklin truly contributed to the discovery of DNA’s structure Franklin was no victim in how the DNA double helix was solved. An overlooked letter and an unpublished news article, both written in 1953, reveal that she was an equal player.

On the event of James Watson's death, I highly recommend this 2023 commentary from @matthewcobb.bsky.social and Nathaniel Comfort with crucial new insights into the discovery of the double helix. (And also check out Cobb's brand new biography of Francis Crick) www.nature.com/articles/d41...

Professor of History - Madison, Wisconsin, United States Current Employees: If you are currently employed at any of the Universities of Wisconsin, log in to Workday to apply through the internal application process.Job Category:FacultyEmployment Type:Regula...

We're hiring in history of science at UW Madison! TT Assistant Professorship with a focus on water. Joint appointment between the History and Integrated Liberal Studies depts, and part of a university-wide hiring cluster on earth/sustainability science. jobs.wisc.edu/jobs/profess... #STS #HSMT

EU weighs pausing parts of landmark AI act in face of US and Big Tech pressure Brussels is set to water down part of its strict digital rule book as it aims to make bloc more competitive

As predicted, the EU is ready to dismantle its Digital Rulebook as it did with the Green Deal.

It does so upon US administration’s demand and under Big Tech pressure.

It’s a self-inflicted political decision putting the EU on a deregulatory turn on.ft.com/47vqFOl

(Call for applications) Assistant Professor Social Studies of Medicine histoiresante.blogspot.com/2025/10/post... #medhum #histSTM

As ever...

As has rightly been pointed out to me it is, in fact, not yet 2026*

Sooo... Gulliver's Travels is 300 *next* year. What do we have planned?!

*I blame overwork plus the fact that I first mentally noted this anniversary 12 years ago, which is ages to wait!
www.theguardian.com/science/the-...

If you can’t afford a vacation, an AI app will sell you pictures of one Always be manifesting.

This version of Total Recall sucks.

your timely reminder that AGI is neither clearly defined/described nor amenable to scientific or engineering principles... it remains an ideology that is rooted in white supremacy, racism and eugenics

According to my notifications, several someones rediscovered this gem earlier today. So, I figured I'd fan the flame. ICYMI.

Call for Papers – IHSA 2026 Conference at Maynooth University

The Irish History Students’ Association, with the support of the Department of History at Maynooth University, is delighted to announce the 77th Annual IHSA Conference, taking place on Saturday, 28 February 2026.

So the argument is more people doing something, means they don't see the benefit of it?

Fry said one of the browser’s best features is memory — making the browser “more personalized and more helpful to you,” as well as an agent mode, meaning that “in Atlas, ChatGPT can now take actions for you

Just a reminder: "more personalized and more helpful" is often code for "WE WILL COLLECT ALL OF YOUR DATA."

This is brilliant.

Financial Times with headline 'Silicon Valley takes stock of the AI bubble'

Just a reminder given all the discussion of AI bubbles and stock markets. 2008 recession saw a huge spike in suicides.

www.bbc.com/news/health-...

Part of the Corlea Trackway on display at the Corlea Trackway Visitor Centre. This trackway is made of oak planks laid transversely on large parallel runners.

The Corlea Trackway • Longford

This 2km ancient routeway was discovered beautifully preserved in the bog.
It was dated to c.148 BC.

A section of the trackway from Cloonbreany Townland was excavated and conserved, and is now on display in the Corlea Trackway Visitor Centre.

#Archaeology 🏺

The preserved timbers of an ancient wooden trackway. Planks are laid horizontally and pegged into place.

We had a fascinating talk on Corlea by Prof Aidan O'Sullivan.

In 'The Colossus of Roads', Aidan offers memories of the Corlea archaeological campaign by one then-young archaeologist, and the question of what did the Trackway mean to the people who built it?

Recording available for Tuatha Members

Really? Perplexity is explicitly advertising cheating with their agentic Comet browser? @marcwatkins.bsky.social found this video on their Facebook page:

www.facebook.com/story.php?st...

Let's put the pressure on for this to stop!

I agree with the technical difference you describe. I just aware that both are wrong. The first we must search ourselves for this, while the latter (as you state) means we have a single source to check against. My main point is 'misreading' is an the wrong word.

My point is as Hicks et al. (2024) it's all bullshit, i.e. it may draw on external sources but it still can tell truth from falsehoods. It's as such not a 'misreading' as only humans can do that rather in both cases it is an outcome of its architecture?

I understand that the issue with LLMs is that they statistical assemble of tokens and do not access the paper. Hence, they can produce wrong information, but many LLM will gather information via RAG and parse it through an LLM. This can produce hallucinations/incorrect information.

Is two not the same as three? Is 'misreading' not an attempt to anthropomorphise 'wrong facts'?

A meme with a character from Watchmen whose name I can’t remember, saying: It is 1982. I am one year old. Tron bombs at the box office It is 2010. I am twenty-nine years old. Tron bombs at the box office It is 2025. I am fourty-four years old. Tron bombs at the box office

This meme made me laugh.