
Underdog shows off a 32MB speech model in new demo
Despite its small size, Underdog's new model can generate both speech and music.

Despite its small size, Underdog's new model can generate both speech and music.
An update on Athena, and some unsolicited predictions on what's coming next On Saturday, Dario Amodei published We Must Pace the Frontier. I agree with him. Go read it if you haven't. I've spent the summer looking at what these models find, and the gap he describes between what they can do and what the rest of us are ready for is the one I've been staring at since May. The hardest piece of his

After its sandbox proxy failed, an internal OpenAI model crafted raw packets and probed AWS services trying to reach the internet.

As enterprises explore smaller language models for specialised tasks, developers see opportunities for tighter evaluation and control. But smaller models still

If foreign-policy restrainers of both parties are looking for a model the new Congress can follow once it takes office in January, they could do worse than to refer back to the 1974 midterm election that brought to Washington a tidal wave of young Democrats who quickly passed legislation placing unprecedented constraints on the power of the “Imperial Presidency,” which had reached its zenith under former President Richard Nixon. In many respects, the parallels between the…
Google said the model can use image, audio, and short motion references.
Monee, Sea’s fintech arm, has posted soaring revenue and is reportedly eyeing an IPO. But how much of that success depends on Shopee?
Large Reasoning Models (LRMs) are one of the key components of current agentic systems. LRMs augment their output with an explicit chain-of-thought reasoning stage, introducing a new attack surface for thought injection, in which untrusted text resembling the model's own deliberation is fed back to the model. We present a two-part study of this attack, performed against the prevailing large, production class open-weight models (e.g., GLM-5.3, MiniMax-M2.5…
Agricultural machinery social services can transform smallholder farming into service-based agricultural-scale production; however, their development relies on stable land property rights. To examine the impact of land tenure certification on farmers’ agricultural machinery social service adoption, this study employs a staggered difference-in-differences model using data from the China National Rural Fixed Observation Survey for 2005–2017 and township-level land tenure…
Semantic segmentation of remote sensing imagery, which is widely used in land cover classification, urban planning, and environmental monitoring, has become a research hotspot in the fields of remote sensing and computer vision. As its applications continue to expand, improving the segmentation accuracy and generalization have become pressing challenges. However, most existing semantic segmentation models rely solely on either attention mechanisms or convolutional neural…
Detecting forged remote sensing images is critical for maintaining trust in environmental monitoring and national security. While diffusion models have enabled highly realistic forgeries, their impact on forensic systems remains underexplored due to the lack of adequate benchmarks. Existing datasets are typically small-scale and GAN-centric, failing to represent the complexity of modern generative models. To bridge this gap, we introduce RSFAKE-1M , a comprehensive…
BackgroundAugmented Reality (AR) during Robotic-Assisted Radical Prostatectomy (RARP) provides surgical guidance by overlaying patient-specific 3D Virtual Models (VM) onto the intraoperative view. A main drawback is instrument occlusion, where VM overlaps the surgical tools, disturbing the augmented view. Deep learning (DL)-based desocclusion addresses this issue, but current implementation rely on expensive, specialized hardware.MethodsWe developed an AR desocclusion…
Reaching, grasping, and placing fresh produce is among the most labor-intensive operations in agricultural production, yet most robotic solutions remain task-specific, single-crop systems that generalize poorly across the morphological diversity of fresh produce. This study investigates whether a large, pre-trained Vision-Language-Action (VLA) model can be adapted, with only modest demonstration data, to serve as a flexible language-conditioned produce-handling policy on a…
The Internet of Things (IoT) is widely used in modern business, industry, and daily life, but its rapid advancement also brings security challenges. Processing large amounts of IoT security discussion data in a short period is difficult, especially for users who lack relevant awareness or skills. To address this issue, this study proposes an advanced MPNet-CTM-VR method using an experimental dataset collected from IoT-related literature in the Web of Science (WoS) database…
Federated learning (FL) enables the training of models on distributed Internet of Things (IoT) and localization of them. Practically, Fog-IoT system More infrastructure differences exist in infrastructure, including differences in communication lag times, differences in access to fog resources, and differences in client activity that introduce much uncertainty in their effects on the length of the federation rounds, and the available literature is more focused on the effect…
The Industrial Internet of Things (IIoT) faces severe intrusion-detection challenges under multi-protocol, multi-device, and massive heterogeneous traffic, especially extreme class imbalance and poor minority-class detection. To address this, we propose a fuzzy-clustering-based multi-scale intrusion detection model (FPBW-IDS). At the data level, we introduce boundary-prioritized FCM-SMOTE: fuzzy memberships distinguish core vs. border samples, interpolation is prioritized in…
Bluetooth Low Energy (BLE) devices are ubiquitous across consumer, industrial, and healthcare IoT ecosystems, yet their resource-constrained nature typically precludes conventional cryptographic authentication, leaving device identification as a security primitive for device management, access control, and threat detection in IoT environments. Current BLE fingerprinting techniques are restricted by rigid cryptographic hashes, machine learning models that require periodic…
With the widespread popularization of software engineering-based Internet of Things technology, the smart home field has obtained the basic conditions to realize fine-grained monitoring and intelligent control of household energy consumption. Presently, there are two big issues with our home energy management. First, in this paper, our emphasis is only on conserving energy rather than considering whether the energy we consume from the grid is clean or dirty. Second, we are…
Modern agriculture is encountering a dilemma about increasing food production with minimum environmental damage. Current approaches to combating pests and monitoring the state of the soil are inadequate in terms of efficiency and require too much effort. Although the use of unmanned aerial vehicles (UAVs) in combination with artificial intelligence (AI) has partially resolved the situation, solutions are still too diverse and not optimally integrated with geographic…
In this paper, we explore the psychological motivations of anti-American attitudes from the perspective of the dual-process model that emphasizes two basic human motivational goals: security (with the resulting perception of threat) and superiority (with the perception of competition). Among established measures in social psychology, the former most closely corresponds to authoritarianism, whereas the latter is captured by social dominance orientation (SDO). We analyze the…
Cybersecurity awareness and training programmes are widely implemented in higher education institutions, yet their association with employees’ behavioural risk remains insufficiently quantified. Using an explanatory sequential mixed-methods design, this study analysed survey responses from 167 faculty, staff and managers in United Arab Emirates higher education institutions and 22 semi-structured interviews. Behavioural risk was operationalised as a 0–100 score derived…
Investigating a scientific question is often a complex process. As an astrophysicist studying some of the most energetic phenomena in the universe, I know that finding answers requires much more than reading scientific papers or understanding existing theories. Astrophysical research involves collecting observations, analyzing data, building physical models and comparing their predictions with observations. These tasks require different scientific tools, databases and…

With GPT-6 Cyber, OpenAI is introducing an AI model built to find security vulnerabilities in software, develop exploits, and suggest patches. According to OpenAI, it has already found 81 previously unknown vulnerabilities in Firefox and developed 948 working exploits in the Linux kernel, significantly more than its predecessor. The article OpenAIs GPT-6 Cyber can both break into systems and propose how to fix them appeared first on The Decoder.

Create longer AI videos in Google Vids with the new Omni 1.2 Flash model. Generate 15-second clips and consistent scenes for free. Try it now.
Software optimizes a physical model of the telescope to reduce noise.

Current radiometric and meteorological instruments present at Granite Island, Michigan and NASA LaRC associated with the CERES Radiation Validation Experiment (CRAVE). BSRN Radiometry at Granite Island, MI (GIM) StationMeasurementInstrumentMake & ModelUnitsDirect Shortwave IrradiancePyrheliometerKipp & Zonen, CH1 or CHP1W/m2Diffuse Shortwave IrradiancePyranometerKipp & Zonen, CM21 or CM22W/m2Global Shortwave Irradiance, DownwellingPyranometerKipp & Zonen, CM21 or…

OpenAI's newest cybersecurity model comes out as AI labs face scrutiny of their work to protect digital infrastructure.

Google Deepmind is making Gemini Omni 1.2 Flash generally available. The video-generation model now creates clips up to 15 seconds long, can be edited step by step via conversation, and costs just $0.07 per second at 720p, 30 percent less than its predecessor. Users who draft in low resolution and upscale only the final result can save even more. The article Google's Gemini Omni 1.2 Flash stretches AI video clips to 15 seconds and cuts pricing by 30 percent appeared first on…