Tootfinder

Ich habe gerade ein Bild gesehen von Thomas Brezina.
Bin ich so alt oder sah der schon in meiner Kindheit aus wie Willem Dafoe?!?
#amdamdes #thomasbrezina #willemdafoe

ACE-G: Improving Generalization of Scene Coordinate Regression Through Query Pre-Training
Leonard Bruns, Axel Barroso-Laguna, Tommaso Cavallari, \'Aron Monszpart, Sowmya Munukutla, Victor Adrian Prisacariu, Eric Brachmann
https://arxiv.org/abs/2510.11605

ACE-G: Improving Generalization of Scene Coordinate Regression Through Query Pre-Training
Scene coordinate regression (SCR) has established itself as a promising learning-based approach to visual relocalization. After mere minutes of scene-specific training, SCR models estimate camera poses of query images with high accuracy. Still, SCR methods fall short of the generalization capabilities of more classical feature-matching approaches. When imaging conditions of query images, such as lighting or viewpoint, are too different from the training views, SCR models fail. Failing to genera…

A Bourgain-Brezis-Mironescu result for fractional thin films
Andrea Braides, Margherita Solci
https://arxiv.org/abs/2508.08874 https://arxiv.org/pdf/2508.0…

A Bourgain-Brezis-Mironescu result for fractional thin films
We consider the limit of squared $H^s$-Gagliardo seminorms on thin domains of the form $Ω_\varepsilon=ω\times(0,\varepsilon)$ in $\mathbb R^d$. When $\varepsilon$ is fixed, multiplying by $1-s$ such seminorms have been proved to converge as $s\to 1^-$ to a dimensional constant $c_d$ times the Dirichlet integral on $Ω_\varepsilon$ by Bourgain, Brezis and Mironescu. In its turn such Dirichlet integrals divided by $\varepsilon$ converge as $\varepsilon\to 0$ to a dimensionally reduced Dirichlet…

Unlocking Altermagnetism in Antiferromagnetic 2D Films via Adsorption
Dong Liu, Sike Zeng, Ji-Hai Liao, Yu-Jun Zhao
https://arxiv.org/abs/2507.09518 https:…

Unlocking Altermagnetism in Antiferromagnetic 2D Films via Adsorption
Altermagnets, characterized by zero net magnetization and momentum-dependent spin splitting, have recently garnered significant attention due to their potential applications in a variety of fields. Here, we propose a symmetry-engineering strategy to unlock altermagnetism in two dimensional (2D) antiferromagnetic systems via surface adsorption of atoms or molecules. By employing spin group theory, we systematically demonstrate that selectively breaking symmetry operations, specifically those pro…

Beim Schreiben neige ich (anders als beim mündlichen Brabbeln) zur Kürze. Aber zu Gabriele Mainettis neuem Film LA CITTÀ PROIBITA (deutscher Titel aus unklaren Gründen KUNG FU IN ROME <- ja, in englischer Schreibweise...) gab es überraschend viel zu sagen. Mein Text im Perlentaucher.
https://www.

Fabian Tietke - Filmkritik zu "Kung-Fu in Rome" von Gabriele Mainetti - Im Kino
Fabian Tietke: Im Kino vom 08.09.2025

Filing: Lime's UK revenue rose 75% YoY to £111.3M in 2024, while its pre-tax profit fell to £1.71M, partly due to a £51.3M "reseller fee" paid to its US parent (Athena Browning/Zag Daily)
https://zagdaily.com/trends/limes-uk-revenue…

For the first time in years, I’m not feeling any FOMO about not preordering anything on iPhone day.
I’ll pick up the new watch when Costco has them, might as well save a buck. My series 7 is starting to struggle with the battery.
But there’s absolutely no hype there, just a practical upgrade whenever.
Thanks for reminding me how much of a capitalist bootlicker you are Tim, otherwise the spell might not have broken.

a good blog but the most relevant line is "GPT-5 may be a moderate quantitative improvement (and it may be cheaper) but it still fails in all the same qualitative ways as its predecessors" , very true but indeed now using it a few days and i notice those moderate improvements.. And i was already aware of all it's failings.
For me in day-to-day use it is better and that is wat counts for me at least. Oh and always (yes always) check outcomes before you use them.

GPT-5: Overdue, overhyped and underwhelming. And that’s not the worst of it.
A new release botched … and a breaking research new paper that spells trouble

The “Neural Data” Goldilocks Problem: Defining “Neural Data” in U.S. State Privacy Laws
https://fpf.org/blog/the-neural-data-goldilocks-problem-defining-neural-data-in-u-s-state-privacy-laws/
@…

GPT-5's release was underwhelming, offering incremental improvements and failing to meet expectations, showing that pure scaling simply isn't the path to AGI (Gary Marcus/Marcus on AI)
https://garymarcus.substack.com/p/gpt-5-overdue-overhyped-and-underwhel…

GPT-5: Overdue, overhyped and underwhelming. And that’s not the worst of it.
A new release botched … and a breaking research new paper that spells trouble