I don't think people are ready for this kind of joke yet.

Which is pretty crucial when broadsheets, tabloids and broadcasters decide to focus upon British and European wildfires this summer. @eciu.net

The UK media landscape overall is increasingly treating net zero and adjacent policies as a political/economic topic rather than a climate‑science one, the Energy and Climate Intelligence Unit found in December 2025. eciu.net/media/press-...

With the new Ethnic Unity Law coming into force this month, with purposefully vague wording on how that legislation affects Chinese diaspora overseas, this doesn't fill me with warm fuzzy feelings.

We don't talk enough about stuff that's of personal interest to me. And it's amazing how few people know the information I've just learned

Particularly when judgements of value are made by people who haven't lifted a finger for a manual labour job in their lives.

Plenty of decent writing out there about the intrinsic difficulties of placing value on different forms of employment these days, played out in equal pay employment tribunals. notes.archie-hall.com/p/against-eq...

Against 'equal value'

A tricky argument, but an essential one

notes.archie-hall.com

My mother's view

Robert Smith@bondhack.ft.com · 2w ago

‘Tice said, however, that it was “not surprising no one reads the FT”, adding: “it’s so boring”.’ www.ft.com/content/4abf...

Enduring tension between populism and actual boring governance.

🎉 And now there’s some good news. The government has committed to implementing Rycroft’s recommendations in the Representation of the People Bill about to go through its final stages in the Commons.

And we’re not alone. In fact, the Rycroft Review into foreign interference made this recommendation and also called for the Electoral Commission to mandate political parties to submit their annual reports and accounts and their campaign spending returns in a standardised format.

The Rycroft Review: Report of the independent review into countering foreign financial influence and interference in UK politics

gov.uk

But democracy isn’t just about elections. Full transparency & consistent disclosures across all parties is key for assessing the impact of big money on our politics. 📣 This is why we’re calling for a change in the law to require political parties to declare campaign expenditure all year round. 🗓️

🔎 But if you read the small print of Labour’s accounts, this is because electoral law offers a loophole where campaign spending can be interpreted to only relate to a party’s expenditure in national elections. search.electoralcommission.org.uk/English/Acco...

🤔 It is very hard to believe that Labour spent nothing on campaigning following its record £27m splurge in the previous general election year…

The Electoral Commission has now published the annual accounts of the largest political parties. But something seems amiss when you look at how much each party spent on campaigning in 2025 🤔 ➡️ Reform: £5.1m 🌳 Conservatives: £2.4m 🔶 Liberal Democrats: £694k 🌹 Labour: £0

What could possibly go wrong?

AI is now making new viruses - Engadget

engadget.com

- but it's still worth taking into account how this might impact British universities/research institutions that are reliant on funding and research capabilities domiciled in or linked to China, including PLA-linked establishments.

by prohibiting collaboration with these where NSF funding is involved. As mentioned by others, the Science.org piece has exaggerated those numbers slightly - x.com/i/status/207...

This is the NSF (U.S. National Science Foundation) notifying its domestic research community that, in aligning with existing DOD/DOE policies, NSF is now taking the risk-elimination rather than the previous risk-mitigation approach to certain international entities - www.nsf.gov/funding/info...

Prohibition on Collaborations with Restricted Entities

nsf.gov

The implications for British research institutions are, needless to say, pretty big and will potentially reveal the true extent of exposure that our universities have to Chinese research collaborations and funding.

Johns Hopkins Berman Institute of Bioethics@bermaninstitute.bsky.social · last mo.

New NSF policy would ban almost all collaborations with Chinese scientists. Agency drops attempt to mitigate risk in favor of outright prohibition on working with most Chinese research institutions. | Science | AAAS www.science.org/content/arti...

The U.S. National Science Foundation (NSF, previously the US closest equivalent to DSIT until, well, y'know) has announced an investment of $380 million to 20 teams so they can establish a nationwide network of AI-enabled automated laboratories across the United States. www.nsf.gov/tip/updates/...

NSF announces $400M investment in new national network of AI-programmable cloud laboratories in alignment with the U.S. government's Genesis Mission

nsf.gov

Grateful to Juliet Samuel for the scalpel drawn across British academia's over-reliance on Chinese research institutions affiliated with the PLA, www.thetimes.com/article/47f1...

Chinese money still has universities in thrall

Despite warnings that Beijing is a ‘genuine and increasing cyber risk’, British science’s love affair with China continues

thetimes.com

Probably time for @rusi.bsky.social to roll this one out again. www.rusi.org/explore-our-...

We Need to Talk About the Insider Risk from AI

As AI systems gain human-like capabilities, organisations must tackle emerging risks from AI insiders, drawing on insights from managing human insider risks to enhance security.

rusi.org

x.com/i/status/208...

Excellent case study for research into agentic AI-enabled insider risks.

Ethan Mollick@emollick.bsky.social · 2w ago

Yes, the AIs were given a cybersecurity challenge, with internet access enabled and safety filters disabled But the extent to which Mythos 5 pursued its mission (fake identities, social engineering, inserting malicious code into a real open-source project) is notable www.aisi.gov.uk/blog/inciden...

So I'm taking this with a huge pinch of salt while considering the realities of the insider risk present.

As evidenced here. www.bbc.co.uk/news/article...

However - yet again - a human was behind the motivation (to complete the tasks set). So although the methodology of reaching that point was taken autonomously, the initial motivation/prompt to do so wasn't engineered by the AI.

Usagi@usagi3939.bsky.social · 2w ago

All that being said, this excerpt from AISI's report gives me the heeby-jeebies. bsky.app/profile/usag...

It has to be independent in order to evaluate frontier models. There's an important distinction here between the suspected hype of "dangerous autonomous" models presented by the actual suppliers, and an independent test house reporting it's seen concerning results.