So hows the existential dread everyone? Where on a scale of super-El-Nino to Killer-Robots are we in terms of "Avoiding the news and hiding in a garden shed with chocolate"?

Well this looks like an interesting PhD project... People should definitely apply for it...! bbsrcdtp.lifesci.cam.ac.uk/project-refe...

1 yr Post doc in my lab! A fun collaborative project building theory of mind in AI. You'll need to pass basic security check - do get in touch with any questions! www.cam.ac.uk/jobs/researc...

🇪🇸🇪🇸🇪🇸🇪🇸🇪🇸🇪🇸🇪🇸

FIFA frantically changing the rules so if you go to full time at 0-0, the team with the most red cards wins #WorldCupFinal

More than 2,700 people died in the May and June heatwaves. Extreme heat is a growing threat to our health and safety. New buildings must be designed to protect people from extreme heat. https://www.theguardian.com/politics/2026/jul/15/england-risks-building-death-traps-overheating-crisis-plans

England risks building new ‘death traps’ as experts warn of overheating crisis

Vulnerable people more at risk as research finds only half of local authority plans require cooling strategies

theguardian.com

Thank you to coauthors Matteo G. Mecattaf, @lucycheke.bsky.social , John Burden and Winnie Street. This work was completed partially in the ERA AI safety fellowship, and partially during my PhD: Take a look at the paper at arxiv.org/abs/2606.31916!

Theory of Mind and Persuasion Beyond Conversation: Assessing the Capacity of LLMs to Induce Belief States via Planning and Action

Theory of Mind (ToM) benchmarks for Large Language Models (LLMs) typically rely on passive question-answering formats, but the deployment of LLMs in increasingly agentic and autonomous forms demands n...

arxiv.org

There's a lot further to go to understand the full extent of NCP-ToM: Real scenarios will have much more complexity, both in the environment and in those being persuaded. We hope our work inspires others evaluating social intelligence in models to consider the NCP-ToM aspects of their scenarios.

Our design also allowed us to pair each NCP-ToM task (Agentic) with a conventional ToM task (Q&A). We thought that the Agentic tasks might be strictly more difficult than the Q&A ones, but that broke down for GPT-5, which had lots of cases where it failed the Q&A task but passed the Agentic.

We evaluated six frontier models, including GPT-5, Gemini 2.5 Pro, Claude models, and a cohort of human participants, across 600 task instances. GPT-5 was the only model to outperform human participants on our task. However, humans were still more robust to changes in the context than models.

We inverted conventional ToM tasks to task the model with inducing belief states in simulated characters, by curating what they observe. Achieving many beliefs simultaneously requires a lot of planning, so we call the capability we're testing Non-Conversational Planning Theory of Mind (NCP-ToM).

New work on AI Theory of Mind: We ask: *How capable are models at inducing belief states without using conversation?* This is important to track as it unlocks many AI agent use-cases, but also potential harms. In our task, recent generations of models exceeded human performance 🧵

AI generated videos are becoming so life like, but it just means that more and more videos are giving me an uncanny valley ick.

Europe's record-breaking June heat wave continues. The climate crisis is here. Yet Farage, like Germany's AfD, the RN in France, and other European far-right parties calls for net-zero targets yo be dropped and more licences and investment in oil & gas. www.theguardian.com/environment/... (1/2)

European heatwave is worst ever and impossible without climate crisis, scientists say

Study also finds high humidity means people in hundreds of cities are enduring their worst ever heat stress

theguardian.com

little recognised hot weather needs: a proper computer mouse. signed, a lady with an unpleasantly sweaty track pad.

wait, is bluesky suddenly gonna be full of under 16s

Ladies First is such a bad film

Said a spokesperson for chimps and elephants, "Fuck you, we're not a problem to solve" www.npr.org/2026/06/07/n...

Bumblebees have tiny brains but they can solve problems like chimps and elephants

New research suggests the fuzzy insects may be capable of spontaneously solving problems the way animals with much larger brains do.

npr.org

Giles 😭😭😭

Angela Rayner was dragged through every front page for weeks... HMRC's verdict: NOT deliberate. NOT even careless. Farage pockets £5 million and chooses not to declare it So where's the wall-to-wall coverage? Tell me this isn't a rigged game.

This ⬇️ 💯 is extremely important and it's essential that it'd be widely watched. It dates back several years (you can see Thiel is much younger) and he really did say that. Nowadays he is less direct but it's important to understand his view of things.

Dr Abeba Birhane@abeba.blacksky.app · 5mo ago

remember what Thiel wants

Nature research paper: General scales unlock AI evaluation with explanatory and predictive power go.nature.com/4telRF0

General scales unlock AI evaluation with explanatory and predictive power - Nature

A fully automated methodology based on rubrics capturing a broad range of cognitive and intellectual demands is illustrated using LLMs and tasks, demonstrating a new way to evaluate the capabilities of AI systems and anticipate their performance.

go.nature.com

Does anyone have experience of CFS like symptoms that lasted like 3-4 months then just went away? I wonder if (hope that?) there are a lot of these people but they are invisible as not in support groups / at the dr?

New Paper Out! This one has been a *long* time in the making, with the data collected over 10 years ago! We found that different aspects of the "obese syndrome" predict different elements of memory performance. www.tandfonline.com/doi/full/10....

Adiposity, diet or inactivity – which is the culprit in obesity-related memory deficits in human adults?

Obesity, diet and physical activity have independently been shown to impact cognition and memory. However, these factors have rarely been considered together in humans, and thus their relative cont...

tandfonline.com

we are weeks away from 'Trump attempts to make horse a Senator' and the only thing stopping him is that he is too much of a sociopath to form an attachment to animals

The Daily Beast@thedailybeast.bsky.social · 7mo ago

Trump is obsessed with buying his aides shoes, and they’re too worried about igniting his wrath to take them off.

if you told someone they were just like Noah from "Nobody wants this" would they think you were: a) insulting them b) flirting c) calling them ernest and considerate d) calling them a bit annoying

why are all these skiers so beautiful? like seriously do sponsors only support the hot ones?