Tootfinder

A reflection on AI advances in the past decade and how scaling and time-horizon trends might point to far greater capabilities in the decade ahead (Zhengdong Wang)
https://zhengdongwang.com/2025/12/30/2025-letter.html

2025 letter
Zhengdong Wang’s personal website

What a great read and overview, recommended !
"The State Of LLMs 2025: Progress, Problems, and Predictions"
#AI

The State Of LLMs 2025: Progress, Progress, and Predictions
A 2025 review of large language models, from DeepSeek R1 and RLVR to inference-time scaling, benchmarks, architectures, and predictions for 2026.

If AI chatbot companies truly had what they claim they have (arbitrary scaling human-level intelligence)—they would use it exclusively themselves, prompting it to come up with schemes to make money and execute them.
In reality these companies all lose money (in historically unprecedented amounts), to fuel a drug dealer-like approach by giving it away for free and hoping enough people get addicted to sycophantic chatbots; with the goal to charge exorbitant fees for it in the future.

When I was a kid, I had a couple well-loved issues of G.I. Joe Special Missions, and this art from the cover of one of them was my absolute favorite. And I ran across it the other day again, and it still strikes me as extremely badass, even now. Just something about Snake Eyes scaling a wall in moonlight.
#art #comics …

I'm running into a really annoying render bug with the latest Kubuntu, which now uses Wayland. Anyone else seen this issue, and know how to fix it?
https://askubuntu.com/questions/1559959/chromium-apps-scaling-wrong-in-wayland-kubuntu-25-10

Chromium apps scaling wrong in Wayland / Kubuntu 25.10
I just installed Kubuntu 25.10 for the first time, on a fresh laptop install. 25.10 defaults to Wayland now. Now, Chromium-based apps (Chrome, Signal, and Slack so far) are showing mis-scaled; th...

I need to read it properly, but this looks 🔥 https://arxiv.org/abs/2511.16652

Evolution Strategies at the Hyperscale
We introduce Evolution Guided General Optimization via Low-rank Learning (EGGROLL), an evolution strategies (ES) algorithm designed to scale backprop-free optimization to large population sizes for modern large neural network architectures with billions of parameters. ES is a set of powerful blackbox optimisation methods that can handle non-differentiable or noisy objectives with excellent scaling potential through parallelisation. Na{ï}ve ES becomes prohibitively expensive at scale due to the…

BoN Appetit Team at LeWiDi-2025: Best-of-N Test-time Scaling Can Not Stomach Annotation Disagreements (Yet)
Tomas Ruiz, Siyao Peng, Barbara Plank, Carsten Schwemmer
https://arxiv.org/abs/2510.12516

BoN Appetit Team at LeWiDi-2025: Best-of-N Test-time Scaling Can Not Stomach Annotation Disagreements (Yet)
Test-time scaling is a family of techniques to improve LLM outputs at inference time by performing extra computation. To the best of our knowledge, test-time scaling has been limited to domains with verifiably correct answers, like mathematics and coding. We transfer test-time scaling to the LeWiDi-2025 tasks to evaluate annotation disagreements. We experiment with three test-time scaling methods: two benchmark algorithms (Model Averaging and Majority Voting), and a Best-of-N sampling method. T…

Right after graduating in 2015, I was offered a job at the #BBC, working on scaling iPlayer. Well, continuing to scale iPlayer. It's one of my great what ifs.
It's an organisation I hugely admire, for many of the reasons that it is being attacked so viciously right now. Truth matters. Finding common ground matters. Keep the faith.

Lithuania scaling down exiled Belarusian oppositon leader's protection, her office pauses work: https://benborges.xyz/2025/10/08/lithuania-scaling-down-exiled-belarusian.html

Scaling #Postgres
https://www.pgedge.com/blog/scaling-postgres

I've finally had the mental space over the holidays to make the switch to #niri off of #sway and I am already very happy with it. For one thing it seems Xwayland with scaling isn't broken anymore (of course the sway authors say it isn't broken, I'm just holding it wrong).

#freshRss release notes:
"Scaling of user statistics in Web UI and CLI, to help instances with 1k users"
I feel a bit small with my installation for two users
https://github.com/FreshRSS/FreshRSS/r

Releases · FreshRSS/FreshRSS
A free, self-hostable news aggregator…. Contribute to FreshRSS/FreshRSS development by creating an account on GitHub.

Either KDE fixes the problems with their Wayland session (faulty font rendering with fractional scaling, no ability to remap touchpad gestures that are literally baked in at compile time, and so on) or I'll be shopping for a new desktop environment once they drop X11 in 2027. I'll not be railroaded into accepting a dogshit user experience because somebody wants to dunk on the chuds. Plasma isn't that special!

ARISE: An Adaptive Resolution-Aware Metric for Test-Time Scaling Evaluation in Large Reasoning Models
Zhangyue Yin, Qiushi Sun, Zhiyuan Zeng, Zhiyuan Yu, Qipeng Guo, Xuanjing Huang, Xipeng Qiu
https://arxiv.org/abs/2510.06014

ARISE: An Adaptive Resolution-Aware Metric for Test-Time Scaling Evaluation in Large Reasoning Models
Test-time scaling has emerged as a transformative paradigm for enhancing the performance of large reasoning models, enabling dynamic allocation of computational resources during inference. However, as the landscape of reasoning models rapidly expands, a critical question remains: how can we systematically compare and evaluate the test-time scaling capabilities across different models? In this paper, we introduce ARISE (Adaptive Resolution-aware Scaling Evaluation), a novel metric specifically d…

DriveVLA-W0: World Models Amplify Data Scaling Law in Autonomous Driving
Yingyan Li, Shuyao Shang, Weisong Liu, Bing Zhan, Haochen Wang, Yuqi Wang, Yuntao Chen, Xiaoman Wang, Yasong An, Chufeng Tang, Lu Hou, Lue Fan, Zhaoxiang Zhang
https://arxiv.org/abs/2510.12796

DriveVLA-W0: World Models Amplify Data Scaling Law in Autonomous Driving
Scaling Vision-Language-Action (VLA) models on large-scale data offers a promising path to achieving a more generalized driving intelligence. However, VLA models are limited by a ``supervision deficit'': the vast model capacity is supervised by sparse, low-dimensional actions, leaving much of their representational power underutilized. To remedy this, we propose \textbf{DriveVLA-W0}, a training paradigm that employs world modeling to predict future images. This task generates a dense, self-supe…

Q&A with Ilya Sutskever about model jaggedness, why we are moving beyond the "age of scaling", SSI's plan to straight-shot superintelligence, AGI, and more (Dwarkesh Patel/Dwarkesh Podcast)
https://www.dwarkesh.com/p/ilya-sutskever-2

On Uniformly Scaling Flows: A Density-Aligned Approach to Deep One-Class Classification
Faried Abu Zaid, Tim Katzke, Emmanuel M\"uller, Daniel Neider
https://arxiv.org/abs/2510.09452

On Uniformly Scaling Flows: A Density-Aligned Approach to Deep One-Class Classification
Unsupervised anomaly detection is often framed around two widely studied paradigms. Deep one-class classification, exemplified by Deep SVDD, learns compact latent representations of normality, while density estimators realized by normalizing flows directly model the likelihood of nominal data. In this work, we show that uniformly scaling flows (USFs), normalizing flows with a constant Jacobian determinant, precisely connect these approaches. Specifically, we prove how training a USF via maximum…

Engineering atomic superradiance scaling in cavity QED system with collective and individual emission channels
Ruijin Sun, Xiang Guo, Andreas Ruschhaupt, Zhihai Wang
https://arxiv.org/abs/2510.12086

Engineering atomic superradiance scaling in cavity QED system with collective and individual emission channels
The coherent emission of multiple atoms gives rise to superradiance, a cornerstone phenomenon in quantum optics with wide-ranging applications in quantum information processing and precision metrology. Despite its importance, how the superradiant scaling with respect to the number of participating atoms can be effectively controlled remains largely unexplored. In this work, we investigate a cavity-QED system and demonstrate that atom-photon coupling can significantly alter the emission behavior…

A new $1/(1-\rho)$-scaling bound for multiserver queues via a leave-one-out technique
Yige Hong
https://arxiv.org/abs/2510.11015 https://arxiv.org/pdf/2510…

A new $1/(1-ρ)$-scaling bound for multiserver queues via a leave-one-out technique
Bounding the queue length in a multiserver queue is a central challenge in queueing theory. Even for the classic $GI/GI/n$ queue with homogeneous servers, it is highly non-trivial to derive a simple and accurate bound for the steady-state queue length that holds across all scaling regimes. A recent breakthrough by Li and Goldberg (2025) establishes bounds that scale as $1/(1-ρ)$ for any load $ρ< 1$ and number of servers $n$, which is the correct scaling in many well-known scaling regimes, inc…

Why is the UK scaling back jury trials, and why is it controversial? | Civil Rights News | Al Jazeera
https://www.aljazeera.com/news/2025/12/2/why-is-the-uk-scaling-back-jury-trials-and-why-is-it-controversial?traffic_source=rss

ElasticMoE: An Efficient Auto Scaling Method for Mixture-of-Experts Models
Gursimran Singh (Huawei Technologies Canada), Timothy Yu (Huawei Technologies Canada), Haley Li (Huawei Technologies Canada), Cheng Chen (Huawei Technologies Canada), Hanieh Sadri (Huawei Technologies Canada), Qintao Zhang (Huawei Technologies China), Yu Zhang (Huawei Technologies China), Ying Xiong (Huawei Technologies Canada), Yong Zhang (Huawei Technologies Canada), Zhenan Fan (Huawei Technologies Canada)

ElasticMoE: An Efficient Auto Scaling Method for Mixture-of-Experts Models
Mixture-of-Experts (MoE) models promise efficient scaling of large language models (LLMs) by activating only a small subset of experts per token, but their parallelized inference pipelines make elastic serving challenging. Existing strategies fall short: horizontal scaling provisions entire replicas of the current configuration, often tens to hundreds of accelerators, leading to coarse granularity, long provisioning delays, and costly overprovisioning. Vertical scaling offers finer adjustments …

I missed this one:
#12886 [css-fonts-5] Text Fitting: Default scaling limit
https://github.com/w3c/csswg-drafts/issues/12886
Essentially a discussion how responsive text can satisfy 1.4.4, especially in this fit-to-container pitch from Google.
I added a comment …

[css-fonts-5] Text Fitting: Default scaling limit · Issue #12886 · w3c/csswg-drafts
In the original issue we received valuable feedback that a native fit-to-width solution can potentially introduce a simple way for authors to violate Success Criterion 1.4.4 Resize Text, in the sam...

Privacy Enhancement in Over-the-Air Federated Learning via Adaptive Receive Scaling
Faeze Moradi Kalarde, Ben Liang, Min Dong, Yahia A. Eldemerdash Ahmed, Ho Ting Cheng
https://arxiv.org/abs/2510.03860

Privacy Enhancement in Over-the-Air Federated Learning via Adaptive Receive Scaling
In Federated Learning (FL) with over-the-air aggregation, the quality of the signal received at the server critically depends on the receive scaling factors. While a larger scaling factor can reduce the effective noise power and improve training performance, it also compromises the privacy of devices by reducing uncertainty. In this work, we aim to adaptively design the receive scaling factors across training rounds to balance the trade-off between training convergence and privacy in an FL syst…

Scaling Up: Lessons From The World's Best CEOs And Founders
Great Australian Pods Podcast Directory: https://www.greataustralianpods.com/scaling-up-lessons-from-the-worlds-best-ceos-and-founders/

David #Lammy,
“Trials are a fundamental part of our democratic settlement. Criminal trials without juries are a bad idea.”
And,
<<I have principles. If they are inconvenient, I have others,>>
Why is the UK scaling back jury trials, and why is it controversial? | Civil Rights News | Al Jazeera

Why is the UK scaling back jury trials, and why is it controversial?
Right to trial by jury seen as a touchstone of the British justice system, but severe court backlogs may prompt change.

Key Considerations for Auto-Scaling: Lessons from Benchmark Microservices
Majid Dashtbani, Ladan Tahvildari
https://arxiv.org/abs/2510.02585 https://arxiv.…

Key Considerations for Auto-Scaling: Lessons from Benchmark Microservices
Microservices have become the dominant architectural paradigm for building scalable and modular cloud-native systems. However, achieving effective auto-scaling in such systems remains a non-trivial challenge, as it depends not only on advanced scaling techniques but also on sound design, implementation, and deployment practices. Yet, these foundational aspects are often overlooked in existing benchmarks, making it difficult to evaluate autoscaling methods under realistic conditions. In this pap…

There's a type of guy whose only contribution at work is sheer volume of outputs, whether or not they serve any purpose. #AI tools ask us, "what if everyone could be that guy?"
It turns out that the result is bad for everyone. Systems lose their ability to evaluate whether outputs are fit for purpose. Shared intent disappears.
Scaling up trash only makes more trash. Work tha…

All your problems are people problems (Don't Just Do Things, Part 2)
If you want to do strategy, you must be willing to do process transformation first.

Allometric scaling of brain activity explained by avalanche criticality
Tiago S. A. N. Sim\~oes, Jos\'e S. Andrade Jr., Hans J. Herrmann, Stefano Zapperi, Lucilla de Arcangelis
https://arxiv.org/abs/2512.10834 https://arxiv.org/pdf/2512.10834 https://arxiv.org/html/2512.10834
arXiv:2512.10834v1 Announce Type: new
Abstract: Allometric scaling laws, such as Kleiber's law for metabolic rate, highlight how efficiency emerges with size across living systems. The brain, with its characteristic sublinear scaling of activity, has long posed a puzzle: why do larger brains operate with disproportionately lower firing rates? Here we show that this economy of scale is a universal outcome of avalanche dynamics. We derive analytical scaling laws directly from avalanche statistics, establishing that any system governed by critical avalanches must exhibit sublinear activity-size relations. This theoretical prediction is then verified in integrate-and-fire neuronal networks at criticality and in classical self-organized criticality models, demonstrating that the effect is not model-specific but generic. The predicted exponents align with experimental observations across mammal species, bridging dynamical criticality with the allometry of brain metabolism. Our results reveal avalanche criticality as a fundamental mechanism underlying Kleiber-like scaling in the brain.
toXiv_bot_toot

The evolution of influence operations
from crude Russian troll farms to sophisticated AI systems using large language models;
the discovery of GoLaxy documents revealing a "Smart Propaganda System" that collects millions of data points daily, builds psychological profiles, and generates resilient personas;
the fundamental challenges of measuring effectiveness;
GoLaxy's ties to Chinese intelligence agencies;
operations targeting Hong Kong's…

Unify Variables in Neural Scaling Laws for General Audio Representations via Embedding Effective Rank
Xuyao Deng, Yanjie Sun, Yong Dou, Kele Xu
https://arxiv.org/abs/2510.10948 …

Unify Variables in Neural Scaling Laws for General Audio Representations via Embedding Effective Rank
Scaling laws have profoundly shaped our understanding of model performance in computer vision and natural language processing, yet their application to general audio representation learning remains underexplored. A key challenge lies in the multifactorial nature of general audio representation-representation quality is jointly influenced by variables such as audio length, embedding dimensionality, model depth, model architecture, data volume, etc., many of which are difficult to isolate or expr…

Scaling Properties of Avalanche Activity in the Two-Dimensional Abelian Sandpile Model
Anubhav Ganguly
https://arxiv.org/abs/2510.09631 https://arxiv.org/p…

Scaling Properties of Avalanche Activity in the Two-Dimensional Abelian Sandpile Model
We study the scaling properties of avalanche activity in the two-dimensional Abelian sandpile model. Instead of the conventional avalanche size distribution, we analyze the site activity distribution, which measures how often a site participates in avalanches when grains are added across the lattice. Using numerical simulations for system sizes up to \(L = 160\), averaged over \(10^4\) configurations, we determine the probability distribution \(P(A, L)\) of site activities. The results show tha…

Neuralink scaling? At some point, with growing numbers of Neuralink brain implant recipients, Neuralink will be busier with trying to fix broken implants than with new implants for new patients. If not, where's support? https://chatgpt.com/share/690a6c48-8f98-8004-9d69-c2a05198d…

ChatGPT - Neuralink scaling challenges
Shared via ChatGPT

SITCOM: Scaling Inference-Time COMpute for VLAs
Ayudh Saxena, Harsh Shah, Sandeep Routray, Rishi Rajesh Shah, Esha Pahwa
https://arxiv.org/abs/2510.04041 https://

SITCOM: Scaling Inference-Time COMpute for VLAs
Learning robust robotic control policies remains a major challenge due to the high cost of collecting labeled data, limited generalization to unseen environments, and difficulties in planning over long horizons. While Vision-Language-Action (VLA) models offer a promising solution by grounding natural language instructions into single-step control commands, they often lack mechanisms for lookahead and struggle with compounding errors in dynamic tasks. In this project, we introduce Scaling Infere…

Scaling Law in LLM Simulated Personality: More Detailed and Realistic Persona Profile Is All You Need
Yuqi Bai, Tianyu Huang, Kun Sun, Yuting Chen
https://arxiv.org/abs/2510.11734

Scaling Law in LLM Simulated Personality: More Detailed and Realistic Persona Profile Is All You Need
This research focuses on using large language models (LLMs) to simulate social experiments, exploring their ability to emulate human personality in virtual persona role-playing. The research develops an end-to-end evaluation framework, including individual-level analysis of stability and identifiability, as well as population-level analysis called progressive personality curves to examine the veracity and consistency of LLMs in simulating human personality. Methodologically, this research propo…

So, I guess no one has any idea how to fix Wayland vs. Chromium?
https://askubuntu.com/questions/1559959/chromium-apps-scaling-wrong-in-wayland-kubuntu-25-10

Chromium apps scaling wrong in Wayland / Kubuntu 25.10
I just installed Kubuntu 25.10 for the first time, on a fresh laptop install. 25.10 defaults to Wayland now. Now, Chromium-based apps (Chrome, Signal, and Slack so far) are showing mis-scaled; th...

Towards Inference-time Scaling for Continuous Space Reasoning
Minghan Wang, Thuy-Trang Vu, Ehsan Shareghi, Gholamreza Haffari
https://arxiv.org/abs/2510.12167 https://

Towards Inference-time Scaling for Continuous Space Reasoning
Inference-time scaling through multiple sample generation in combination with Process- or Outcome-Reward Model (PRM or ORM) re-ranking has proven effective for text-based reasoning in large language models. This paper investigates whether such established techniques can be successfully adapted to reasoning in the continuous space, using COCONUT (Hao et al. 2024) continuous space reasoning LM as the backbone. We demonstrate the feasibility of generating diverse reasoning paths through dropout-ba…

NaViL: Rethinking Scaling Properties of Native Multimodal Large Language Models under Data Constraints
Changyao Tian, Hao Li, Gen Luo, Xizhou Zhu, Weijie Su, Hanming Deng, Jinguo Zhu, Jie Shao, Ziran Zhu, Yunpeng Liu, Lewei Lu, Wenhai Wang, Hongsheng Li, Jifeng Dai
https://arxiv.org/abs/2510.08565 …

NaViL: Rethinking Scaling Properties of Native Multimodal Large Language Models under Data Constraints
Compositional training has been the de-facto paradigm in existing Multimodal Large Language Models (MLLMs), where pre-trained vision encoders are connected with pre-trained LLMs through continuous multimodal pre-training. However, the multimodal scaling property of this paradigm remains difficult to explore due to the separated training. In this paper, we focus on the native training of MLLMs in an end-to-end manner and systematically study its design space and scaling property under a practica…

Scaling crossover of the generalized Jeffreys-type law
Fugui Ma
https://arxiv.org/abs/2510.07930 https://arxiv.org/pdf/2510.07930

Scaling crossover of the generalized Jeffreys-type law
The generalized Jeffreys-type law is formulated as a multi-term time-fractional Jeffreys-type equation, whose dynamics exhibit rich scaling crossover phenomena entailing different diffusion mechanisms. In this work, we provide a novel physical explanation for the equation from first principles, beginning with a microscopic description based on the continuous-time random walk framework with a generalized waiting time distribution and further deriving the equation from an overdamped Langevin equa…

"When your number's up, your number's up", my mum used to say:
November 16, 2012. Patrick Edlinger, a celebrated climber who had braved all dangers with his bare hands, accidentally dies by falling down the stairs at home.

Scaling of Magnetic Domain Walls in Perpendicular Magnetic Anisotropy Systems
Guowen Gong, Changmin Xiong, Lijun Zhu
https://arxiv.org/abs/2510.10230 https://

Scaling of Magnetic Domain Walls in Perpendicular Magnetic Anisotropy Systems
Magnetic domain walls play a critical role in the nanoscale evolution of magnetic devices. Despite the early efforts, a complete understanding of the micromagnetic evolution of the width and the type of magnetic domain walls has still remained missing. Here, we report a combined analytical and micromagnetic simulation study and establish the scaling of the magnetic domains as a function of the exchange stiffness (A), uniaxial perpendicular magnetic anisotropy (Ku), saturation magnetization (Ms)…

Kibble-Zurek Scaling and Spatial Statistics in Quenched Binary Bose Superfluids
Subhadeep Patra, Arko Roy, Seong-Ho Shinn, Adolfo del Campo, Mithun Thudiyangal
https://arxiv.org/abs/2510.12770

Kibble-Zurek Scaling and Spatial Statistics in Quenched Binary Bose Superfluids
We study how gradually changing the chemical potential causes a two-dimensional binary Bose gas to condense from vacuum to finite density, resulting in either a mixed (miscible) or separated (immiscible) state depending on interaction strengths. In the immiscible case, random domains form, and their number, boundary length, and average size at the point of equilibration follow universal Kibble-Zurek (KZ) scaling with the cooling rate. These patterns continue to evolve in a self-similar way whil…

Pushing Test-Time Scaling Limits of Deep Search with Asymmetric Verification
Weihao Zeng, Keqing He, Chuqiao Kuang, Xiaoguang Li, Junxian He
https://arxiv.org/abs/2510.06135 htt…

Pushing Test-Time Scaling Limits of Deep Search with Asymmetric Verification
Test-time compute can be scaled both sequentially and in parallel. Sequential scaling involves lengthening the generation process, while parallel scaling involves verifying and selecting among multiple candidate outputs. Combining these two strategies has led to the most powerful AI systems, such as Grok 4 Heavy and GPT-5 Pro. In certain contexts (e.g., solving Sudoku puzzles), verifying responses can be substantially easier than generating them. This property, referred to as \emph{asymmetric v…

CMOS 2.0 - Redefining the Future of Scaling
Moritz Brunion, Navaneeth Kunhi Purayil, Francesco Dell'Atti, Sebastian Lam, Refik Bilgic, Mehdi Tahoori, Luca Benini, Julien Ryckaert
https://arxiv.org/abs/2510.04535

CMOS 2.0 - Redefining the Future of Scaling
We propose to revisit the functional scaling paradigm by capitalizing on two recent developments in advanced chip manufacturing, namely 3D wafer bonding and backside processing. This approach leads to the proposal of the CMOS 2.0 platform. The main idea is to shift the CMOS roadmap from geometric scaling to fine-grain heterogeneous 3D stacking of specialized active device layers to achieve the ultimate Power-Performance-Area and Cost gains expected from future technology generations. However, t…

Geometric Model Selection for Latent Space Network Models: Hypothesis Testing via Multidimensional Scaling and Resampling Techniques
Jieyun Wang, Anna L. Smith
https://arxiv.org/abs/2510.06136

Geometric Model Selection for Latent Space Network Models: Hypothesis Testing via Multidimensional Scaling and Resampling Techniques
Latent space models assume that network ties are more likely between nodes that are closer together in an underlying latent space. Euclidean space is a popular choice for the underlying geometry, but hyperbolic geometry can mimic more realistic patterns of ties in complex networks. To identify the underlying geometry, past research has applied non-Euclidean extensions of multidimensional scaling (MDS) to the observed geodesic distances: the shortest path lengths between nodes. The difference in…

"Scientists Develop Cigarette Butt Asphalt to Build Stronger Roads"
#Roads #Recycling
https://hap…

Scientists Develop Cigarette Butt Asphalt to Build Stronger Roads | Happy Eco News
Scientists in Spain and Italy are developing cigarette butt asphalt by turning discarded cigarette filters into road-strengthening pellets. The innovation makes asphalt more durable, extends road lifespans, and reduces toxic waste in the environment. Pilot projects show promise for scaling, offering cities a sustainable solution to improve infrastructure and cut pollution.

Well, that's two ways of putting it.
https://www.theverge.com/news/823750/european-union-ai-act-gdpr-changes

Europe is scaling back its landmark privacy and AI laws
The European Commission has proposed changes weakening EU privacy legislation and delaying parts of the AI Act to make the bloc more competitive globally.

Monthly Rural-Urban Scaling of Road Accidents in England, Wales and Scotland (2019-2023)
Isabel Copsey, Quentin Hanley, Jack Sutton
https://arxiv.org/abs/2510.07351 https://

Monthly Rural-Urban Scaling of Road Accidents in England, Wales and Scotland (2019-2023)
Road traffic accidents remain a major public health challenge worldwide, with urbanisation and population density identified as key factors influencing risk. This study analyses monthly accident data from 2009 to 2023 across 632 parliamentary constituencies in England, Wales, and Scotland, using an area-normalised approach based on population density. Segmented power law models consistently identified breakpoints separating sublinear rural from superlinear urban scaling behaviours. Seasonal var…

Superradiance and Superabsorption Engine of $N$ Two-Level Systems: $N^{2}$-Power Scaling at Near-Unity Efficiency
L. F. Alves da Silva, H. Sanchez, M. A. Ponte, M. H. Y. Moussa, Norton G. de Almeida
https://arxiv.org/abs/2510.12017

Superradiance and Superabsorption Engine of $N$ Two-Level Systems: $N^{2}$-Power Scaling at Near-Unity Efficiency
We present a thermal engine that exploits the \emph{cooperative superradiance} and \emph{superabsorption} of a sample of \(N\) two-level atoms. This engine operates using a single cold reservoir via cycles of collective pumping followed by decay. Using an effective mean-field Hamiltonian to describe the many-body dynamics, we design optimized drive pulses that preserve adiabaticity and achieve an average power output scaling quadratically with the system size, \(P \propto N^2\). An experimental…

Zephyrus: Scaling Gateways Beyond the Petabit-Era with DPU-Augmented Hierarchical Co-Offloading
Yuemeng Xu, Haoran Chen, Jiarui Guo, Mingwei Cui, Qiuheng Yin, Cheng Dong, Daxiang Kang, Xian Wu, Chenmin Sun, Peng He, Yang Gao, Lirong Lai, Kai Wang, Hongyu Wu, Tong Yang, Xiyun Xu
https://arxiv.org/abs/2510.11043

Zephyrus: Scaling Gateways Beyond the Petabit-Era with DPU-Augmented Hierarchical Co-Offloading
Operating at petabit-scale, ByteDance's cloud gateways are deployed at critical aggregation points to orchestrate a wide array of business traffic. However, this massive scale imposes significant resource pressure on our previous-generation cloud gateways, rendering them unsustainable in the face of ever-growing cloud-network traffic. As the DPU market rapidly expands, we see a promising path to meet our escalating business traffic demands by integrating DPUs with our established Tofino-based g…

Scaling Homomorphic Applications in Deployment
Ryan Marinelli, Angelica Chowdhury
https://arxiv.org/abs/2510.02376 https://arxiv.org/pdf/2510.02376

Scaling Homomorphic Applications in Deployment
In this endeavor, a proof-of-concept homomorphic application is developed to determine the production readiness of encryption ecosystems. A movie recommendation app is implemented for this purpose and productionized through containerization and orchestration. By tuning deployment configurations, the computational limitations of Fully Homomorphic Encryption (FHE) are mitigated through additional infrastructure optimizations Index Terms: Reinforcement Learning, Orchestration, Homomorphic Encryp…

Universal scaling of shear thickening suspensions under acoustic perturbation
Anna R. Barth, Navneet Singh, Stephen J. Thornton, Pranav Kakhandiki, Edward Y. X. Ong, Meera Ramaswamy, Abhishek M. Shetty, Bulbul Chakraborty, James P. Sethna, Itai Cohen
https://arxiv.org/abs/2510.11820

Universal scaling of shear thickening suspensions under acoustic perturbation
Tuning shear thickening behavior is a longstanding problem in the field of dense suspensions. Acoustic perturbations offer a convenient way to control shear thickening in real time, opening the door to a new class of smart materials. However, complete control over shear thickening requires a quantitative description for how suspension viscosity varies under acoustic perturbation. Here, we achieve this goal by experimentally probing suspensions with acoustic perturbations and incorporating their…

xLSTM Scaling Laws: Competitive Performance with Linear Time-Complexity
Maximilian Beck, Kajetan Schweighofer, Sebastian B\"ock, Sebastian Lehner, Sepp Hochreiter
https://arxiv.org/abs/2510.02228 …

xLSTM Scaling Laws: Competitive Performance with Linear Time-Complexity
Scaling laws play a central role in the success of Large Language Models (LLMs), enabling the prediction of model performance relative to compute budgets prior to training. While Transformers have been the dominant architecture, recent alternatives such as xLSTM offer linear complexity with respect to context length while remaining competitive in the billion-parameter regime. We conduct a comparative investigation on the scaling behavior of Transformers and xLSTM along the following lines, prov…

“Google also announced the expansion of its 24/7 Carbon-Free Energy (CFE) partnership with Engie in Germany [..] now scaling up to contribute to Germany's energy transition and grid stability.
[..] Google's German operations are projected to run at or near 85% carbon-free energy in 2026.”

Google Announces €5.5 Billion Investment in Germany, including AI Infrastructure, through 2029
BERLIN — November 11, 2025 — Google today announced a new €5.5 billion investment (2026-2029) in infrastructure and offices in Germany — including a new data center in Dietzenbach,...

I gotta say, #cosmicdesktop gets many things right even in its alpha state. Using it right now. X11 apps and Electron garbage all work fine with fractional scaling. Can't say that about #gnome49. Oh, and quarter tiling by default! Optimistic for what the future brings for it.

Are neural scaling laws leading quantum chemistry astray?
Siwoo Lee, Adji Bousso Dieng
https://arxiv.org/abs/2509.26397 https://arxiv.org/pdf/2509.26397

Are neural scaling laws leading quantum chemistry astray?
Neural scaling laws are driving the machine learning community toward training ever-larger foundation models across domains, assuring high accuracy and transferable representations for extrapolative tasks. We test this promise in quantum chemistry by scaling model capacity and training data from quantum chemical calculations. As a generalization task, we evaluate the resulting models' predictions of the bond dissociation energy of neutral H$_2$, the simplest possible molecule. We find that, reg…

Theory of Scaling Laws for In-Context Regression: Depth, Width, Context and Time
Blake Bordelon, Mary I. Letey, Cengiz Pehlevan
https://arxiv.org/abs/2510.01098 https://

Theory of Scaling Laws for In-Context Regression: Depth, Width, Context and Time
We study in-context learning (ICL) of linear regression in a deep linear self-attention model, characterizing how performance depends on various computational and statistical resources (width, depth, number of training steps, batch size and data per context). In a joint limit where data dimension, context length, and residual stream width scale proportionally, we analyze the limiting asymptotics for three ICL settings: (1) isotropic covariates and tasks (ISO), (2) fixed and structured covarianc…

Dual Data Scaling for Robust Two-Stage User-Defined Keyword Spotting
Zhiqi Ai, Han Cheng, Yuxin Wang, Shiyi Mu, Shugong Xu, Yongjin Zhou
https://arxiv.org/abs/2510.10740 https:/…

Dual Data Scaling for Robust Two-Stage User-Defined Keyword Spotting
In this paper, we propose DS-KWS, a two-stage framework for robust user-defined keyword spotting. It combines a CTC-based method with a streaming phoneme search module to locate candidate segments, followed by a QbyT-based method with a phoneme matcher module for verification at both the phoneme and utterance levels. To further improve performance, we introduce a dual data scaling strategy: (1) expanding the ASR corpus from 460 to 1,460 hours to strengthen the acoustic model; and (2) leveraging…

Lingxi: Repository-Level Issue Resolution Framework Enhanced by Procedural Knowledge Guided Scaling
Xu Yang, Jiayuan Zhou, Michael Pacheco, Wenhan Zhu, Pengfei He, Shaowei Wang, Kui Liu, Ruiqi Pan
https://arxiv.org/abs/2510.11838

Lingxi: Repository-Level Issue Resolution Framework Enhanced by Procedural Knowledge Guided Scaling
Driven by the advancements of Large Language Models (LLMs), LLM-powered agents are making significant improvements in software engineering tasks, yet struggle with complex, repository-level issue resolution. Existing agent-based methods have two key limitations. First, they lack of procedural knowledge (i.e., how an issue is fixed step-by-step and rationales behind it) to learn and leverage for issue resolution. Second, they rely on massive computational power to blindly explore the solution sp…

Prompting Test-Time Scaling Is A Strong LLM Reasoning Data Augmentation
Sondos Mahmoud Bsharat, Zhiqiang Shen
https://arxiv.org/abs/2510.09599 https://arxi…

Prompting Test-Time Scaling Is A Strong LLM Reasoning Data Augmentation
Large language models (LLMs) have demonstrated impressive reasoning capabilities when provided with chain-of-thought exemplars, but curating large reasoning datasets remains laborious and resource-intensive. In this work, we introduce Prompting Test-Time Scaling (P-TTS), a simple yet effective inference-time data augmentation strategy for enhancing LLM reasoning through finetuning. Rather than collecting thousands or even millions of examples, P-TTS leverages a small pool of only 90 manually se…

The EU unveils proposed updates to GDPR, including simplifying cookie permission pop-ups, and plans to water down the AI Act, after US and tech company pressure (The Verge)
https://www.theverge.com/news/823750/european-union-ai-act-gdpr-changes

Europe is scaling back its landmark privacy and AI laws
The European Commission has proposed changes weakening EU privacy legislation and delaying parts of the AI Act to make the bloc more competitive globally.

On the Role of Temperature Sampling in Test-Time Scaling
Yuheng Wu, Azalia Mirhoseini, Thierry Tambe
https://arxiv.org/abs/2510.02611 https://arxiv.org/pdf…

On the Role of Temperature Sampling in Test-Time Scaling
Large language models (LLMs) can improve reasoning at inference time through test-time scaling (TTS), where multiple reasoning traces are generated and the best one is selected. Prior work shows that increasing the number of samples K steadily improves accuracy. In this paper, we demonstrate that this trend does not hold indefinitely: at large K, further scaling yields no gains, and certain hard questions remain unsolved regardless of the number of traces. Interestingly, we find that different …

Comparing Cross-Platform Performance via Node-to-Node Scaling Studies
Kenneth Weiss, Thomas M. Stitt, Daryl Hawkins, Olga Pearce, Stephanie Brink, Robert N. Rieben
https://arxiv.org/abs/2510.12166

Comparing Cross-Platform Performance via Node-to-Node Scaling Studies
Due to the increasing diversity of high-performance computing architectures, researchers and practitioners are increasingly interested in comparing a code's performance and scalability across different platforms. However, there is a lack of available guidance on how to actually set up and analyze such cross-platform studies. In this paper, we contend that the natural base unit of computing for such studies is a single compute node on each platform and offer guidance in setting up, running, and …

BroRL: Scaling Reinforcement Learning via Broadened Exploration
Jian Hu, Mingjie Liu, Ximing Lu, Fang Wu, Zaid Harchaoui, Shizhe Diao, Yejin Choi, Pavlo Molchanov, Jun Yang, Jan Kautz, Yi Dong
https://arxiv.org/abs/2510.01180

BroRL: Scaling Reinforcement Learning via Broadened Exploration
Reinforcement Learning with Verifiable Rewards (RLVR) has emerged as a key ingredient for unlocking complex reasoning capabilities in large language models. Recent work ProRL has shown promise in scaling RL by increasing the number of training steps. However, performance plateaus after thousands of steps, with clear diminishing returns from allocating more computation to additional training. In this work, we investigate a complementary paradigm for scaling RL, BroR-Lincreasing the number of rol…

A Tauberian approach to metric scaling limits of random discrete structures, with an application to random planar maps
William Fleurat
https://arxiv.org/abs/2510.05078 https://

A Tauberian approach to metric scaling limits of random discrete structures, with an application to random planar maps
We prove sandwich theorems and a Tauberian theorem in the space of compact metric measure spaces, endowed with the Gromov-Hausdorff-Prokhorov (GHP) topology. These results hold with respect to a close relative of Gromov's Lipschitz order. As a proof-of-concept of a general method to prove metric scaling limits of random discrete structures, we give an application to the theory of random planar maps: the Brownian sphere is the scaling limit in the GHP topology of irreducible quandrangulations. O…

Scaling up AI requires staggering amounts of power and water
— especially when considering that many areas are already dealing with strained grids or drought conditions.
Even when optimized, a single hyperscale facility can draw as muchpower as a mid-sized city
and millions of gallons of water annually.
Professor Romany Webb, deputy director of Columbia University's Sabin Center for Climate Change Law, explained the challenge:
"Data centers are incred…

New map and research expose disturbing side effects of AI industry boom: 'There are going to be challenges'
As part of the latest wave of Big Tech's spending on AI, Google announced a $9 billion investment to build a new data center campus in Virginia.

BLAZER: Bootstrapping LLM-based Manipulation Agents with Zero-Shot Data Generation
Rocktim Jyoti Das, Harsh Singh, Diana Turmakhan, Muhammad Abdullah Sohail, Mingfei Han, Preslav Nakov, Fabio Pizzati, Ivan Laptev
https://arxiv.org/abs/2510.08572

BLAZER: Bootstrapping LLM-based Manipulation Agents with Zero-Shot Data Generation
Scaling data and models has played a pivotal role in the remarkable progress of computer vision and language. Inspired by these domains, recent efforts in robotics have similarly focused on scaling both data and model size to develop more generalizable and robust policies. However, unlike vision and language, robotics lacks access to internet-scale demonstrations across diverse robotic tasks and environments. As a result, the scale of existing datasets typically suffers from the need for manual…

AutoDAN-Reasoning: Enhancing Strategies Exploration based Jailbreak Attacks with Test-Time Scaling
Xiaogeng Liu, Chaowei Xiao
https://arxiv.org/abs/2510.05379 https://

AutoDAN-Reasoning: Enhancing Strategies Exploration based Jailbreak Attacks with Test-Time Scaling
Recent advancements in jailbreaking large language models (LLMs), such as AutoDAN-Turbo, have demonstrated the power of automated strategy discovery. AutoDAN-Turbo employs a lifelong learning agent to build a rich library of attack strategies from scratch. While highly effective, its test-time generation process involves sampling a strategy and generating a single corresponding attack prompt, which may not fully exploit the potential of the learned strategy library. In this paper, we propose to…

Extreme events scaling in self-organized critical models
Abdul Quadir, Haider Hasan Jafri
https://arxiv.org/abs/2510.08733 https://arxiv.org/pdf/2510.08733…

Extreme events scaling in self-organized critical models
We study extreme events of avalanche activities in finite-size two-dimensional self-organized critical (SOC) models, specifically the stochastic Manna model (SMM) and the Bak-Tang-Weisenfeld (BTW) sandpile model. Employing the approach of block maxima, the study numerically reveals that the distributions for extreme avalanche size and area follow the generalized extreme value (GEV) distribution. The extreme avalanche size follows the Gumbel distribution with shape parameter $ξ=0$ while in the …

DeepPrune: Parallel Scaling without Inter-trace Redundancy
Shangqing Tu, Yaxuan Li, Yushi Bai, Lei Hou, Juanzi Li
https://arxiv.org/abs/2510.08483 https://…

DeepPrune: Parallel Scaling without Inter-trace Redundancy
Parallel scaling has emerged as a powerful paradigm to enhance reasoning capabilities in large language models (LLMs) by generating multiple Chain-of-Thought (CoT) traces simultaneously. However, this approach introduces significant computational inefficiency due to inter-trace redundancy -- our analysis reveals that over 80% of parallel reasoning traces yield identical final answers, representing substantial wasted computation. To address this critical efficiency bottleneck, we propose DeepPru…

Generalized Parallel Scaling with Interdependent Generations
Harry Dong, David Brandfonbrener, Eryk Helenowski, Yun He, Mrinal Kumar, Han Fang, Yuejie Chi, Karthik Abinav Sankararaman
https://arxiv.org/abs/2510.01143

Generalized Parallel Scaling with Interdependent Generations
Parallel LLM inference scaling involves sampling a set of $N>1$ responses for a single input prompt. However, these $N$ parallel responses tend to be generated independently from each other, partitioning compute resources and leaving potentially useful information in one generation untapped by others. This is in contrast to response length scaling where past computation is used in all future steps. For higher quality responses and response sets, we propose Bridge to generate interdependent resp…

Go with Your Gut: Scaling Confidence for Autoregressive Image Generation
Harold Haodong Chen, Xianfeng Wu, Wen-Jie Shu, Rongjin Guo, Disen Lan, Harry Yang, Ying-Cong Chen
https://arxiv.org/abs/2509.26376

Go with Your Gut: Scaling Confidence for Autoregressive Image Generation
Test-time scaling (TTS) has demonstrated remarkable success in enhancing large language models, yet its application to next-token prediction (NTP) autoregressive (AR) image generation remains largely uncharted. Existing TTS approaches for visual AR (VAR), which rely on frequent partial decoding and external reward models, are ill-suited for NTP-based image generation due to the inherent incompleteness of intermediate decoding results. To bridge this gap, we introduce ScalingAR, the first TTS fr…

Anthropic commits to buy $30B in Azure capacity in a new deal with Microsoft and Nvidia, which commit to invest up to $5B and $10B, respectively, in Anthropic (Microsoft)
https://blogs.microsoft.com/blog/2025/11/18/microsoft-nvidia-…

Microsoft, NVIDIA and Anthropic announce strategic partnerships - The Official Microsoft Blog
Anthropic to scale Claude on Azure Anthropic to adopt NVIDIA architecture NVIDIA and Microsoft to invest in Anthropic Today Microsoft, NVIDIA and Anthropic announced new strategic partnerships. Anthropic is scaling its rapidly-growing Claude AI model on Microsoft Azure, powered by NVIDIA, which will broaden access to Claude and provide Azure enterprise customers with expanded model choice and new capabilities. Anthropic...

Critical attention scaling in long-context transformers
Shi Chen, Zhengjiang Lin, Yury Polyanskiy, Philippe Rigollet
https://arxiv.org/abs/2510.05554 https://

Critical attention scaling in long-context transformers
As large language models scale to longer contexts, attention layers suffer from a fundamental pathology: attention scores collapse toward uniformity as context length $n$ increases, causing tokens to cluster excessively, a phenomenon known as rank-collapse. While $\textit{attention scaling}$ effectively addresses this deficiency by rescaling attention scores with a polylogarithmic factor $β_n$, theoretical justification for this approach remains lacking. We analyze a simplified yet tractable…

Quenching, Fast and Slow: Breaking Kibble-Zurek Universal Scaling by Jumping along Geodesics
Thi Ha Kyaw, Guillermo Romero, Gaurav Saxena
https://arxiv.org/abs/2510.08528 https:…

Quenching, Fast and Slow: Breaking Kibble-Zurek Universal Scaling by Jumping along Geodesics
A major drawback of adiabatic quantum computing (AQC) is fulfilling the energy gap constraint, which requires the total evolution time to scale inversely with the square of the minimum energy gap. Failure to satisfy this condition violates the adiabatic approximation, potentially undermining computational accuracy. Recently, several approaches have been proposed to circumvent this constraint. One promising approach is to use the family of adiabatic shortcut procedures to fast-forward AQC. One c…

CodeChemist: Functional Knowledge Transfer for Low-Resource Code Generation via Test-Time Scaling
Kaixin Wang, Tianlin Li, Xiaoyu Zhang, Aishan Liu, Xianglong Liu, Ziqi Liu, Zhiqiang Zhang, Jun Zhou, and Bin Shi
https://arxiv.org/abs/2510.00501

CodeChemist: Functional Knowledge Transfer for Low-Resource Code Generation via Test-Time Scaling
Code Large Language Models (CodeLLMs) are increasingly used in code generation tasks across a wide range of applications. However, their performance is often inconsistent across different programming languages (PLs), with low-resource PLs suffering the most due to limited training data. In this paper, we present CodeChemist, a novel and efficient framework for test-time scaling that enables functional knowledge transfer from high-resource to low-resource PLs using generated test cases. CodeChem…

Generalized Parallel Scaling with Interdependent Generations
Harry Dong, David Brandfonbrener, Eryk Helenowski, Yun He, Mrinal Kumar, Han Fang, Yuejie Chi, Karthik Abinav Sankararaman
https://arxiv.org/abs/2510.01143

Generalized Parallel Scaling with Interdependent Generations
Parallel LLM inference scaling involves sampling a set of $N>1$ responses for a single input prompt. However, these $N$ parallel responses tend to be generated independently from each other, partitioning compute resources and leaving potentially useful information in one generation untapped by others. This is in contrast to response length scaling where past computation is used in all future steps. For higher quality responses and response sets, we propose Bridge to generate interdependent resp…

DiTSinger: Scaling Singing Voice Synthesis with Diffusion Transformer and Implicit Alignment
Zongcai Du, Guilin Deng, Xiaofeng Guo, Xin Gao, Linke Li, Kaichang Cheng, Fubo Han, Siyu Yang, Peng Liu, Pan Zhong, Qiang Fu
https://arxiv.org/abs/2510.09016

DiTSinger: Scaling Singing Voice Synthesis with Diffusion Transformer and Implicit Alignment
Recent progress in diffusion-based Singing Voice Synthesis (SVS) demonstrates strong expressiveness but remains limited by data scarcity and model scalability. We introduce a two-stage pipeline: a compact seed set of human-sung recordings is constructed by pairing fixed melodies with diverse LLM-generated lyrics, and melody-specific models are trained to synthesize over 500 hours of high-quality Chinese singing data. Building on this corpus, we propose DiTSinger, a Diffusion Transformer with Ro…

Scaling Language-Centric Omnimodal Representation Learning
Chenghao Xiao, Hou Pong Chan, Hao Zhang, Weiwen Xu, Mahani Aljunied, Yu Rong
https://arxiv.org/abs/2510.11693 https://…

Scaling Language-Centric Omnimodal Representation Learning
Recent multimodal embedding approaches leveraging multimodal large language models (MLLMs) fine-tuned with contrastive learning (CL) have shown promising results, yet the underlying reasons behind their superiority remain underexplored. This work argues that a crucial advantage of MLLM-based approaches stems from implicit cross-modal alignment achieved during generative pretraining, where the language decoder learns to exploit multimodal signals within a shared representation space for generati…

Resolution scaling governs DINOv3 transfer performance in chest radiograph classification
Soroosh Tayebi Arasteh, Mina Shaigan, Christiane Kuhl, Jakob Nikolas Kather, Sven Nebelung, Daniel Truhn
https://arxiv.org/abs/2510.07191

Resolution scaling governs DINOv3 transfer performance in chest radiograph classification
Self-supervised learning (SSL) has advanced visual representation learning, but its value in chest radiography, a high-volume imaging modality with fine-grained findings, remains unclear. Meta's DINOv3 extends earlier SSL models through Gram-anchored self-distillation. Whether these design choices improve transfer learning for chest radiography has not been systematically tested. We benchmarked DINOv3 against DINOv2 and ImageNet initialization across seven datasets (n>814,000). Two representati…

Best-of-Majority: Minimax-Optimal Strategy for Pass@$k$ Inference Scaling
Qiwei Di, Kaixuan Ji, Xuheng Li, Heyang Zhao, Quanquan Gu
https://arxiv.org/abs/2510.03199 https://

Best-of-Majority: Minimax-Optimal Strategy for Pass@$k$ Inference Scaling
LLM inference often generates a batch of candidates for a prompt and selects one via strategies like majority voting or Best-of- N (BoN). For difficult tasks, this single-shot selection often underperforms. Consequently, evaluations commonly report Pass@$k$: the agent may submit up to $k$ responses, and only the best of them is used when computing regret. Motivated by this, we study inference scaling in the more general Pass@$k$ inference setting, and prove that neither majority voting nor BoN …

Q&A with Rivian founder and CEO RJ Scaringe on founding Rivian in 2009, production challenges, the VW partnership, autonomy, AI, EVs, chips, CarPlay, and more (Ben Thompson/Stratechery)
https://stratechery.com/2025/an-interv

An Interview with Rivian CEO RJ Scaringe About Building a Car Company and Autonomy
An interview with Rivian founder and CEO RJ Scaringe about starting a car company, scaling, and building for an autonomous future.

Verifier-free Test-Time Sampling for Vision Language Action Models
Suhyeok Jang, Dongyoung Kim, Changyeon Kim, Youngsuk Kim, Jinwoo Shin
https://arxiv.org/abs/2510.05681 https:/…

Verifier-free Test-Time Sampling for Vision Language Action Models
Vision-Language-Action models (VLAs) have demonstrated remarkable performance in robot control. However, they remain fundamentally limited in tasks that require high precision due to their single-inference paradigm. While test-time scaling approaches using external verifiers have shown promise, they require additional training and fail to generalize to unseen conditions. We propose Masking Distribution Guided Selection (MG-Select), a novel test-time scaling framework for VLAs that leverages the…

TaTToo: Tool-Grounded Thinking PRM for Test-Time Scaling in Tabular Reasoning
Jiaru Zou, Soumya Roy, Vinay Kumar Verma, Ziyi Wang, David Wipf, Pan Lu, Sumit Negi, James Zou, Jingrui He
https://arxiv.org/abs/2510.06217

TaTToo: Tool-Grounded Thinking PRM for Test-Time Scaling in Tabular Reasoning
Process Reward Models (PRMs) have recently emerged as a powerful framework for enhancing the reasoning capabilities of large reasoning models (LRMs), particularly in the context of test-time scaling (TTS). However, their potential for supervising LRMs on tabular reasoning domains remains underexplored. Through detailed empirical analyses, we identify that existing PRMs, though widely adopted for supervising text-only reasoning steps, struggle with table-specific operations such as sub-table ret…

Nonlinear Heisenberg Limit via Uncertainty Principle in Quantum Metrology
Binke Xia, Jingzheng Huang, Yuxiang Yang, Guihua Zeng
https://arxiv.org/abs/2510.09216 https://

Nonlinear Heisenberg Limit via Uncertainty Principle in Quantum Metrology
The Heisenberg limit is acknowledged as the ultimate precision limit in quantum metrology, traditionally implying that root mean square errors of parameter estimation decrease linearly with the time T of evolution and the number N of quantum gates or probes. However, this conventional perspective fails to interpret recent studies of "super-Heisenberg" scaling, where precision improves faster than linearly with T and N. In this work, we revisit the Heisenberg scaling by leveraging the position-m…

Multi-Dimensional Autoscaling of Stream Processing Services on Edge Devices
Boris Sedlak, Philipp Raith, Andrea Morichetta, V\'ictor Casamayor Pujol, Schahram Dustdar
https://arxiv.org/abs/2510.06882

Multi-Dimensional Autoscaling of Stream Processing Services on Edge Devices
Edge devices have limited resources, which inevitably leads to situations where stream processing services cannot satisfy their needs. While existing autoscaling mechanisms focus entirely on resource scaling, Edge devices require alternative ways to sustain the Service Level Objectives (SLOs) of competing services. To address these issues, we introduce a Multi-dimensional Autoscaling Platform (MUDAP) that supports fine-grained vertical scaling across both service- and resource-level dimensions.…

Crypto investor Roger Ver, aka "Bitcoin Jesus", reaches a deferred prosecution agreement with the US DOJ and pays ~$50M to resolve a tax evasion indictment (Ben Weiss/Fortune)
https://fortune.com/crypto/2025/10/14/roger…

‘Bitcoin Jesus’ reaches $50 million deal with DOJ to dismiss tax evasion charges
Roger Ver’s agreement with prosecutors is the latest sign that President Donald Trump’s administration is scaling back its crackdown on crypto.

Auto-scaling Continuous Memory for GUI Agent
Wenyi Wu, Kun Zhou, Ruoxin Yuan, Vivian Yu, Stephen Wang, Zhiting Hu, Biwei Huang
https://arxiv.org/abs/2510.09038 https://

Auto-scaling Continuous Memory for GUI Agent
We study how to endow GUI agents with scalable memory that help generalize across unfamiliar interfaces and long-horizon tasks. Prior GUI agents compress past trajectories into text tokens, which balloons context length and misses decisive visual cues (e.g., exact widget size and position). We propose a continuous memory that encodes each GUI trajectory into a fixed-length sequence of continuous embeddings using the VLM itself as an encoder; these embeddings are plugged directly into the backbo…

MLE-Smith: Scaling MLE Tasks with Automated Multi-Agent Pipeline
Rushi Qiang, Yuchen Zhuang, Anikait Singh, Percy Liang, Chao Zhang, Sherry Yang, Bo Dai
https://arxiv.org/abs/2510.07307

MLE-Smith: Scaling MLE Tasks with Automated Multi-Agent Pipeline
While Language Models (LMs) have made significant progress in automating machine learning engineering (MLE), the acquisition of high-quality MLE training data is significantly constrained. Current MLE benchmarks suffer from low scalability and limited applicability because they rely on static, manually curated tasks, demanding extensive time and manual effort to produce. We introduce MLE-Smith, a fully automated multi-agent pipeline, to transform raw datasets into competition-style MLE challeng…

Scaling LLM Multi-turn RL with End-to-end Summarization-based Context Management
Miao Lu, Weiwei Sun, Weihua Du, Zhan Ling, Xuesong Yao, Kang Liu, Jiecao Chen
https://arxiv.org/abs/2510.06727

Scaling LLM Multi-turn RL with End-to-end Summarization-based Context Management
We study reinforcement learning (RL) fine-tuning of large language model (LLM) agents for long-horizon multi-turn tool use, where context length quickly becomes a fundamental bottleneck. Existing RL pipelines can suffer from degraded instruction following, excessive rollout costs, and most importantly, strict context limits. To address these challenges, we introduce summarization-based context management to training. In specific, it periodically compresses the tool using history by LLM-generate…

Recursive Self-Aggregation Unlocks Deep Thinking in Large Language Models
Siddarth Venkatraman, Vineet Jain, Sarthak Mittal, Vedant Shah, Johan Obando-Ceron, Yoshua Bengio, Brian R. Bartoldson, Bhavya Kailkhura, Guillaume Lajoie, Glen Berseth, Nikolay Malkin, Moksh Jain
https://arxiv.org/abs/2509.26626…

Recursive Self-Aggregation Unlocks Deep Thinking in Large Language Models
Test-time scaling methods improve the capabilities of large language models (LLMs) by increasing the amount of compute used during inference to make a prediction. Inference-time compute can be scaled in parallel by choosing among multiple independent solutions or sequentially through self-refinement. We propose Recursive Self-Aggregation (RSA), a test-time scaling method inspired by evolutionary methods that combines the benefits of both parallel and sequential scaling. Each step of RSA refines…

Shape Happens: Automatic Feature Manifold Discovery in LLMs via Supervised Multi-Dimensional Scaling
Federico Tiblias, Irina Bigoulaeva, Jingcheng Niu, Simone Balloccu, Iryna Gurevych
https://arxiv.org/abs/2510.01025

Shape Happens: Automatic Feature Manifold Discovery in LLMs via Supervised Multi-Dimensional Scaling
The linear representation hypothesis states that language models (LMs) encode concepts as directions in their latent space, forming organized, multidimensional manifolds. Prior efforts focus on discovering specific geometries for specific features, and thus lack generalization. We introduce Supervised Multi-Dimensional Scaling (SMDS), a model-agnostic method to automatically discover feature manifolds. We apply SMDS to temporal reasoning as a case study, finding that different features form var…

Test-Time Scaling in Diffusion LLMs via Hidden Semi-Autoregressive Experts
Jihoon Lee, Hoyeon Moon, Kevin Zhai, Arun Kumar Chithanar, Anit Kumar Sahu, Soummya Kar, Chul Lee, Souradip Chakraborty, Amrit Singh Bedi
https://arxiv.org/abs/2510.05040

Test-Time Scaling in Diffusion LLMs via Hidden Semi-Autoregressive Experts
Diffusion-based large language models (dLLMs) are trained flexibly to model extreme dependence in the data distribution; however, how to best utilize this information at inference time remains an open problem. In this work, we uncover an interesting property of these models: dLLMs trained on textual data implicitly learn a mixture of semi-autoregressive experts, where different generation orders reveal different specialized behaviors. We show that committing to any single, fixed inference time …

Revisiting Direct Speech-to-Text Translation with Speech LLMs: Better Scaling than CoT Prompting?
Oriol Pareras, Gerard I. G\'allego, Federico Costa, Cristina Espa\~na-Bonet, Javier Hernando
https://arxiv.org/abs/2510.03093

Revisiting Direct Speech-to-Text Translation with Speech LLMs: Better Scaling than CoT Prompting?
Recent work on Speech-to-Text Translation (S2TT) has focused on LLM-based models, introducing the increasingly adopted Chain-of-Thought (CoT) prompting, where the model is guided to first transcribe the speech and then translate it. CoT typically outperforms direct prompting primarily because it can exploit abundant Automatic Speech Recognition (ASR) and Text-to-Text Translation (T2TT) datasets to explicitly model its steps. In this paper, we systematically compare CoT and Direct prompting unde…

Shape Happens: Automatic Feature Manifold Discovery in LLMs via Supervised Multi-Dimensional Scaling
Federico Tiblias, Irina Bigoulaeva, Jingcheng Niu, Simone Balloccu, Iryna Gurevych
https://arxiv.org/abs/2510.01025

Shape Happens: Automatic Feature Manifold Discovery in LLMs via Supervised Multi-Dimensional Scaling
The linear representation hypothesis states that language models (LMs) encode concepts as directions in their latent space, forming organized, multidimensional manifolds. Prior efforts focus on discovering specific geometries for specific features, and thus lack generalization. We introduce Supervised Multi-Dimensional Scaling (SMDS), a model-agnostic method to automatically discover feature manifolds. We apply SMDS to temporal reasoning as a case study, finding that different features form var…

An interview with Sam Altman and OpenAI President Greg Brockman on the tepid initial reception to GPT-5's launch, scaling, reinforcement learning, AGI, and more (Steven Levy/Wired)
https://www.wired.com/story/sam-altman-says-the-gpt-5-haters-got-it-all-wron…

Sam Altman Says the GPT-5 Haters Got It All Wrong
OpenAI's CEO explains that its large language model has been misunderstood—and that he's changed his attitude to AGI.

D2E: Scaling Vision-Action Pretraining on Desktop Data for Transfer to Embodied AI
Suwhan Choi, Jaeyoon Jung, Haebin Seong, Minchan Kim, Minyeong Kim, Yongjun Cho, Yoonshik Kim, Yubeen Park, Youngjae Yu, Yunsung Lee
https://arxiv.org/abs/2510.05684

D2E: Scaling Vision-Action Pretraining on Desktop Data for Transfer to Embodied AI
Large language models leverage internet-scale text data, yet embodied AI remains constrained by the prohibitive costs of physical trajectory collection. Desktop environments -- particularly gaming -- offer a compelling alternative: they provide rich sensorimotor interactions at scale while maintaining the structured observation-action coupling essential for embodied learning. We present D2E (Desktop to Embodied AI), a framework that demonstrates desktop interactions can serve as an effective pr…

Parallel Scaling Law: Unveiling Reasoning Generalization through A Cross-Linguistic Perspective
Wen Yang, Junhong Wu, Chong Li, Chengqing Zong, Jiajun Zhang
https://arxiv.org/abs/2510.02272

Parallel Scaling Law: Unveiling Reasoning Generalization through A Cross-Linguistic Perspective
Recent advancements in Reinforcement Post-Training (RPT) have significantly enhanced the capabilities of Large Reasoning Models (LRMs), sparking increased interest in the generalization of RL-based reasoning. While existing work has primarily focused on investigating its generalization across tasks or modalities, this study proposes a novel cross-linguistic perspective to investigate reasoning generalization. This raises a crucial question: $\textit{Does the reasoning capability achieved from E…

Scaling Spoken Language Models with Syllabic Speech Tokenization
Nicholas Lee, Cheol Jun Cho, Alan W Black, Gopala K. Anumanchipalli
https://arxiv.org/abs/2509.26634 https://

Scaling Spoken Language Models with Syllabic Speech Tokenization
Spoken language models (SLMs) typically discretize speech into high-frame-rate tokens extracted from SSL speech models. As the most successful LMs are based on the Transformer architecture, processing these long token streams with self-attention is expensive, as attention scales quadratically with sequence length. A recent SSL work introduces acoustic tokenization of speech at the syllable level, which is more interpretable and potentially more scalable with significant compression in token len…

Laminar: A Scalable Asynchronous RL Post-Training Framework
Guangming Sheng, Yuxuan Tong, Borui Wan, Wang Zhang, Chaobo Jia, Xibin Wu, Yuqi Wu, Xiang Li, Chi Zhang, Yanghua Peng, Haibin Lin, Xin Liu, Chuan Wu
https://arxiv.org/abs/2510.12633

Laminar: A Scalable Asynchronous RL Post-Training Framework
Reinforcement learning (RL) post-training for Large Language Models (LLMs) is now scaling to large clusters and running for extended durations to enhance model reasoning performance. However, the scalability of existing RL frameworks is limited, as extreme long-tail skewness in RL trajectory generation causes severe GPU underutilization. Current asynchronous RL systems attempt to mitigate this, but they rely on global weight synchronization between the actor and all rollouts, which creates a ri…