Tootfinder

What Does Your Benchmark Really Measure? A Framework for Robust Inference of AI Capabilities
Nathanael Jo, Ashia Wilson
https://arxiv.org/abs/2509.19590 https://

What Does Your Benchmark Really Measure? A Framework for Robust Inference of AI Capabilities
Evaluations of generative models on benchmark data are now ubiquitous, and their outcomes critically shape public and scientific expectations of AI's capabilities. Yet growing skepticism surrounds their reliability. How can we know that a reported accuracy genuinely reflects a model's true performance? Evaluations are often presented as simple measurements, but in reality they are inferences: to treat benchmark scores as evidence of capability is already to assume a theory of what capability is…

OmniScene: Attention-Augmented Multimodal 4D Scene Understanding for Autonomous Driving
Pei Liu, Hongliang Lu, Haichao Liu, Haipeng Liu, Xin Liu, Ruoyu Yao, Shengbo Eben Li, Jun Ma
https://arxiv.org/abs/2509.19973

OmniScene: Attention-Augmented Multimodal 4D Scene Understanding for Autonomous Driving
Human vision is capable of transforming two-dimensional observations into an egocentric three-dimensional scene understanding, which underpins the ability to translate complex scenes and exhibit adaptive behaviors. This capability, however, remains lacking in current autonomous driving systems, where mainstream approaches primarily rely on depth-based 3D reconstruction rather than true scene understanding. To address this limitation, we propose a novel human-like framework called OmniScene. Fir…

Hey everyone,
@…’s mom needs your help:
“My mother has been suffering from chronic diabetes, and some time ago, one of her toes had to be amputated. During the war, her condition has worsened, and she now relies on a walker to move. 💔
Today, she fell inside our tent, and her foot got injured again – it’s swollen and bruised badly…

Help Bara'a and her elderly parents in Gaza
My name is Bara'a El-Mughraby. Me and my older brother are the sole breadwinners for my family of five in Al-Shati refugee camp in North Gaza. Me, my elderly parents, my sister and my brother. For months, we have endured the unimaginable and did everything we can to survive, but now our situation worsened a lot and has become a matter of life or death. The constant bombardment and the recent evacuation orders from the IDF mean we have to leave our home—but we don't have the capabilities nor t…

CyberSOCEval: Benchmarking LLMs Capabilities for Malware Analysis and Threat Intelligence Reasoning
Lauren Deason, Adam Bali, Ciprian Bejean, Diana Bolocan, James Crnkovich, Ioana Croitoru, Krishna Durai, Chase Midler, Calin Miron, David Molnar, Brad Moon, Bruno Ostarcevic, Alberto Peltea, Matt Rosenberg, Catalin Sandu, Arthur Saputkin, Sagar Shah, Daniel Stan, Ernest Szocs, Shengye Wan, Spencer Whitman, Sven Krasser, Joshua Saxe

CyberSOCEval: Benchmarking LLMs Capabilities for Malware Analysis and Threat Intelligence Reasoning
Today's cyber defenders are overwhelmed by a deluge of security alerts, threat intelligence signals, and shifting business context, creating an urgent need for AI systems to enhance operational security work. While Large Language Models (LLMs) have the potential to automate and scale Security Operations Center (SOC) operations, existing evaluations do not fully assess the scenarios most relevant to real-world defenders. This lack of informed evaluation impacts both AI developers and those apply…

Federation of Agents: A Semantics-Aware Communication Fabric for Large-Scale Agentic AI
Lorenzo Giusti, Ole Anton Werner, Riccardo Taiello, Matilde Carvalho Costa, Emre Tosun, Andrea Protani, Marc Molina, Rodrigo Lopes de Almeida, Paolo Cacace, Diogo Reis Santos, Luigi Serio
https://arxiv.org/abs/2509.20175…

Federation of Agents: A Semantics-Aware Communication Fabric for Large-Scale Agentic AI
We present Federation of Agents (FoA), a distributed orchestration framework that transforms static multi-agent coordination into dynamic, capability-driven collaboration. FoA introduces Versioned Capability Vectors (VCVs): machine-readable profiles that make agent capabilities searchable through semantic embeddings, enabling agents to advertise their capabilities, cost, and limitations. Our aarchitecturecombines three key innovations: (1) semantic routing that matches tasks to agents over shar…

EnAnchored-X2X: English-Anchored Optimization for Many-to-Many Translation
Sen Yang, Yu Bao, Yu Lu, Jiajun Chen, Shujian Huang, Shanbo Cheng
https://arxiv.org/abs/2509.19770 htt…

EnAnchored-X2X: English-Anchored Optimization for Many-to-Many Translation
Large language models (LLMs) have demonstrated strong machine translation capabilities for English-centric language pairs but underperform in direct non-English (x2x) translation. This work addresses this limitation through a synthetic data generation framework that leverages models' established English-to-x (en2x) capabilities. By extending English parallel corpora into omnidirectional datasets and developing an English-referenced quality evaluation proxy, we enable effective collection of hig…

Migrating a large accretion of complex blog post content between platforms is a draining & thoroughly demoralising experience. It turns out that we, as web developers, are capable of creating byzantine platforms which give diligent blog writers the ability to create hours and hours of unrewarding work that is so cryptically & diversely broken that no script can be written to make sense of the mess.
Enjoy your holidays if you're in a cultural context where that sort of thin…

In 2014, NASA awarded a Commercial Crew Transportation Capability contract to Boeing to fly astronauts to and from the International Space Station with its #Starliner spacecraft; as part of its contract, Boeing was awarded up to six crewed flights to the orbital complex - after a thorough evaluation, the definitive order has been adjusted to four missions, with the remaining two available as options: https://www.nasa.gov/blogs/spacestation/2025/11/24/nasa-boeing-modify-commercial-crew-contract/. The next Starliner flight, known as Starliner-1, will be used by NASA to deliver necessary cargo to the orbital laboratory and allow in-flight validation of the system upgrades implemented following the Crew Flight Test mission last year - NASA and Boeing are targeting no earlier than April 2026 to fly the uncrewed Starliner-1 pending completion of rigorous test, certification, and mission readiness activities.

South Korea, the EU, the UK, Saudi Arabia, the UAE, and others aim to build "sovereign AI"; Gartner expects $1.5T in AI spending in 2025, up 50% from 2024 (Jiyoung Sohn/Wall Street Journal)
https://www.

‘Sovereign AI’ Takes Off as Countries Seek to Avoid Overreliance on Superpowers
South Korea and others believe their homegrown tech sectors are strong enough to build up their AI capabilities

Flexion, which is building a full autonomy stack for humanoid robots, raised a $50M Series A from DST Global Partners, NVentures, and others, after a $7M seed (Mary Ann Azevedo/Crunchbase News)
https://news.crunchbase.com/venture/robotic-brain-building-star…

Exclusive: Founded By Ex-Nvidia Researchers, Flexion Lands $50M To Build The 'Brain' for Humanoid Robots
A startup founded by former Nvidia researchers that’s “building the brain for humanoid and human-capable robots,” Flexion has raised $50 million in funding, it tells Crunchbase News exclusively.