Tootfinder

Apparently, looking for The Darkest of the Hillside Thickets’ cover of “Walking on the Moon” on #Apple #Music is a cry for help. Judgy much, Apple? 🤔

Climate Station ausprobiert: Mit PSVR 2 durch die Klimageschichte
Sonys kostenlose PS5-App "Climate Station" visualisiert den Klimawandel. Wir haben sie ausprobiert und klären, welchen Mehrwert Playstation VR 2 bringt.

I've never written a novel or any other intensely-plotted work of fiction, but anyone who has read or watched it played a lot of stories can probably also recognize that some authors just aren't good at endings. They're great at setting things in motion, at keeping the twists and turns coming, at the soap opera style of drama. But they just don't have the craft necessary to tie things together into a satisfying conclusion. I imagine it's much harder than the process of getting things going out keeping them moving, since you both have to wind down all the various threads you've spun up and balance satisfaction with believability.
I just finished Girl Gone Viral by Arvin Ahmadi, and it has a bad ending. The beginning is fine, the middle has plenty of drama to keep you wanting to see what happens, but the ending is murky, unsatisfying, and manages neither veracity nor satisfaction (even discounting the biggest next step that might reasonably have been left there to make room for a sequel).
Given the other issues with the book, from poor politics, to inauthentic characters, to a techno-optimism that feels as bitter in this moment as it is far from the mark in its predictions, I can't recommended it, despite having read through to the end.
#AmReading

Paul Simonon with an impromptu duet of "Guns of Brixton" with some "punk rock kids" playing the song on the street 🤩
https://www.tiktok.com/@the_whops/video/7515347559120948502

Leaky Thoughts: Large Reasoning Models Are Not Private Thinkers
Tommaso Green, Martin Gubri, Haritz Puerto, Sangdoo Yun, Seong Joon Oh
https://arxiv.org/abs/2506.15674

Leaky Thoughts: Large Reasoning Models Are Not Private Thinkers
We study privacy leakage in the reasoning traces of large reasoning models used as personal agents. Unlike final outputs, reasoning traces are often assumed to be internal and safe. We challenge this assumption by showing that reasoning traces frequently contain sensitive user data, which can be extracted via prompt injections or accidentally leak into outputs. Through probing and agentic evaluations, we demonstrate that test-time compute approaches, particularly increased reasoning steps, ampl…

Thoughts from the @braiduk.bsky.social event thus far (shaped through my own interests and thinking)
People's desire to shape AI in society was very present (but I missed a sense of grassroots work eg AI4LAM in GLAMs, other equivalents)
Science fiction and other creative expressions can help society think through potential models for AI
AI is an extension, intensification of existing technologies
Clearly tensions, temptations between values and efficiencies, which lin…

Getting ready for tomorrow's #ClimateClassic: road cycling tour of 250 km, from @… in De Bilt to Groningen, along the 1-metre-above-sealevel line. In our special #ShowYourClimateStripes

heise | Pixelprotz mit vielen Ports
Der 27-Zoll-Monitor PD2730S mit 5K-Auflösung zeigt sich äußerst anschlussfreudig. Mac & i hat ihn getestet.
https://www.heise.de/tests/Pixelprotz-mit-

Our partner conference, MICES – Mix-Camp E-Commerce Search, takes place on June 18th at TUECHTIG in Berlin. Focusing on e-commerce search, MICES brings together experts from various fields such as IT, product management, UX design, search management, information retrieval, data science and search engine vendors to discuss challenges, share ideas, best practices and case studies in e-commerce search.
Register for free and learn more here:

How Well Can Reasoning Models Identify and Recover from Unhelpful Thoughts?
Sohee Yang, Sang-Woo Lee, Nora Kassner, Daniela Gottesman, Sebastian Riedel, Mor Geva
https://arxiv.org/abs/2506.10979

How Well Can Reasoning Models Identify and Recover from Unhelpful Thoughts?
Recent reasoning models show the ability to reflect, backtrack, and self-validate their reasoning, which is crucial in spotting mistakes and arriving at accurate solutions. A natural question that arises is how effectively models can perform such self-reevaluation. We tackle this question by investigating how well reasoning models identify and recover from four types of unhelpful thoughts: uninformative rambling thoughts, thoughts irrelevant to the question, thoughts misdirecting the question a…