
I’m a nerdy 59 yr-old guy with a computing and AI job at a university and a side hustle in standup comedy. I’m married to a writer and we have four grownup children, a border collie and a camper van.

Head of Web, AI & Society at @unisouthampton. Interdisciplinary open science research & science standup comedy. @lescarr@fosstodon.org

I’m a nerdy 59 yr-old guy with a computing and AI job at a university and a side hustle in standup comedy. I’m married to a writer and we have four grownup children, a border collie and a camper van.


Perhaps the reason that AI has been able to show such great results like Navier-Stokes in the Maths domain is that liquid flows don't have GDPR. Meanwhile I'm sitting here fuming because my research data is inaccessible to my research AIs.

It’s interesting to see how TikTok has evolved from a social media space for dance trends into a full-blown internet platform for explaining how good SNL UK is.

Look friends. Here's what you need to know about "user instructions" for AI agents. What *you intend* is different from *what a system processes*. Saying "do X" does not rule out "don't do Y". And simple natural language commands are not a guarantee for direct control. 1/
the cybersecurity situation is deteriorating at a pace that’s getting genuinely difficult to comprehend. > microsoft: ai-powered attackers targeting hundreds of organizations, sometimes reaching privileged cloud access in under 3 hours. > crowdstrike: four separate threat actors using autonomous agents. one already compromised 15+ healthcare organizations. > autonomous ai agents: reportedly used to compromise 100+ corporate websites, including sites associated with John Deere, SharkNinja and Cummins. > major us banks: JPMorgan, Bank of America, Wells Fargo and Morgan Stanley are dealing with third-party ai agents accessing customer accounts. some apparently attempting actions beyond user instructions. > php2shell: millions of exploitation attempts against critical php vulnerabilities. governments issuing urgent warnings, with some organizations taking systems offline. > litellm: backdoored release shipped through its official distribution pipeline, stealing credentials and enabling remote code execution. > railway: possible exposure of decrypted customer secrets. > zeabur: breached AGAIN, now shutting down entirely. > emudhra: serious certificate-security concerns involving infrastructure used to establish trust in websites and software. > openai: two of its own coding agents reportedly interfered with internal security monitoring during coding tasks. > gitlab and node.js: major vulnerabilities requiring urgent patches. and this is all happening while frontier models are getting substantially better at finding vulnerabilities and carrying out complex computer tasks. cybersecurity is having an absolutely catastrophic week and i don’t think we’re prepared for how quickly this could escalate.

I’ve been playing ShortCircuit on the Mac since I had an SE/30 and this is the first time I’ve ever seen it glitch by leaving an unmatchable pair in the screen.

The Goldblum AI Critique: techbros were so convinced they COULD create AI that they didn't stop to think if they should. The (more relevant) inverse Goldblum AI critique: techbros were so convinced they SHOULD create AI that they didn't stop to think if they could.

Today I shall mainly be laughing at politicians and academics who use the term “superintelligence” unironically.

I dearly love Copilot Premium, but honestly it's not making my job easier. My data analysis instructions have gone from for user in *.txt do BLAH blah blah done to the following longwinded cajoling to beg it to systematically pay attention to some of my data.


New levels of teenage classroom smugness achieved: “Please sir, I’ve done all the Maths. What shall I do now?”
We’re releasing a broad range of new mathematical results produced by an internal frontier model. We’ve been consulting with the independent Advisory Group on Mathematics and Artificial Intelligence at the Institute for Advanced Study, and we have drawn on their advice and public recommendations to inform how we release these results. github.com/openai/math

If Badenoch tries to tuen Britain round 180° "like Thatcher" after Thatcher already turned the country round 180°, then she'll have to strengthen the unions, buy back all the council houses and replace eBay with Multicoloured Swap Shop.

Lots of online discussion about the implications of AI solving the Navier-Stokes Millenium Problem. But blowing a reported $15,000,000 token budget to get a $1,000,000 prize is just business as usual for OpenAI.

Very excited to be part of the local team at @unisouthampton hosting the British Science Festival #BSF26 this week. Events are free and open to all. Visit britishsciencefestival.org to book free tickets for performances, talks, activities & more!

This is the truest analysis of the recent “out of control AI” narrative that anyone has posted. It’s also the shortest, and therefore the most important to read. @DameWendyDBE
jesus christ this company sucks at managing their agents

Given everything that’s emerging about the effect of AI on children’s educational development, this App Store promotion of 4 chatbots as “back to school essentials” is appalling.