This year at The Game Awards, if your game wasn't melodramatic, mechanically innovative, beautifully presented and aggressively French, it didn't stand a chance. The Game Awards 2025 wrapped up on the evening of Thursday, December 11 with a record-breaking showing by from Sandfall Interactive. The game received the most nominations and wins in the show's 12-year history. But, we know that's not really why you're here. Between the award presentations and musical numbers, there were heaps of new game trailers, announcements and updates, and we've collected them all for you right here.

'Children's cafeterias' in Japan hit record 12,601 sites, survey reveals The number of children's cafeterias that provide free or low-cost meals mainly to children in need in Japan rose by more than 1,700 from the previous fiscal year, according to a survey by a nonprofit organization. The number of children's cafeterias, which provide free or low-cost meals mainly to children in need in Japan, has reached a record 12,601 this fiscal year, according to a survey by a nonprofit organization. The total rose by more than 1,700 from the previous fiscal year, said the survey released Thursday by Musubie, a Tokyo-based nonprofit supporting kodomo shokudō programs nationwide. The nonprofit organization said the expansion reflected efforts by central and local governments to create comfortable spaces for children. We aim to create an environment that makes it easier to start and sustain kodomo shokudō programs, Musubie head Rie Mishima said at a news conference.

OpenAI and its investor Microsoft have been sued over a Connecticut murder-suicide in the latest case to blame ChatGPT for dangerous psychological manipulation of users. OpenAI and its investor, Microsoft, have been sued over a Connecticut murder-suicide in the latest case to blame the popular ChatGPT chatbot for dangerous psychological manipulation of users. The lawsuit turns on the actions of a 56-year-old man who lived with his 83-year-old mother in Greenwich, Connecticut, and had been conversing for months with the chatbot over his fear that he was under surveillance and people were trying to kill him. In August, according to police and the state medical examiner, Stein-Erik Soelberg killed his mother, Suzanne Adams, then took his own life. Soelberg's dialogue with ChatGPT convinced him that he had made the chatbot conscious, and that he had been implanted with a "divine instrument system" in his neck and brain, which related to a "divine mission," according to a complaint filed Thursday in California Superior Court in San Francisco, where OpenAI is based.

Casey Hudson is returning to space for a brand new project. The Game Awards kicked off with a bang, showing the world premiere of . It's a brand new action role-playing game that will be directed by Casey Hudson, who previously headed up several notable BioWare games you probably know like and . There's not much to go on in the trailer, but the game simply existing is a pretty great surprise and this cinematic trailer sure looks shiny. Hudson is working with Arcanaut Studios on this project, which is described as "an epic interactive adventure across a galaxy on the brink of rebirth where every decision shapes your path towards light or darkness."

Google has announced plans to team up with the UK to invite researchers to come up with uses for the tech giant's state-of-the-art quantum chip Willow. It is one of several firms competing to develop a powerful quantum computer - which is seen as an exciting new frontier in the future of computing. Researchers hope they will be able to crack problems in fields such as chemistry and medicine which are impossible for current computers to solve. Professor Paul Stevenson of the University of Surrey - who had no involvement with the agreement - told the BBC it was great news for UK researchers. The collaboration between Google and the UK's national lab for quantum computing means more researchers will get access to the technology.

This paper presents HunyuanOCR, a commercial-grade, open-source, and lightweight (1B parameters) Vision-Language Model (VLM) dedicated to OCR tasks. The architecture comprises a Native Vision Transformer (ViT) and a lightweight LLM connected via an MLP adapter. HunyuanOCR demonstrates superior performance, outperforming commercial APIs, traditional pipelines, and larger models (e.g., Qwen3-VL-4B). Specifically, it surpasses current public solutions in perception tasks (Text Spotting, Parsing) and excels in semantic tasks (IE, Text Image Translation), securing first place in the ICDAR 2025 DIMT Challenge (Small Model Track). Furthermore, it achieves state-of-the-art (SOTA) results on OCRBench among VLMs with fewer than 3B parameters. HunyuanOCR achieves breakthroughs in three key aspects: 1) Unifying Versatility and Efficiency: We implement comprehensive support for core capabilities including spotting, parsing, IE, VQA, and translation within a lightweight framework. This addresses the limitations of narrow "OCR expert models" and inefficient "General VLMs". 2) Streamlined End-to-End Architecture: Adopting a pure end-to-end paradigm eliminates dependencies on pre-processing modules (e.g., layout analysis). This fundamentally resolves error propagation common in traditional pipelines and simplifies system deployment. 3) Data-Driven and RL Strategies: We confirm the critical role of high-quality data and, for the first time in the industry, demonstrate that Reinforcement Learning (RL) strategies yield significant performance gains in OCR tasks. HunyuanOCR is officially open-sourced on HuggingFace. We also provide a high-performance deployment solution based on vLLM, placing its production efficiency in the top tier. We hope this model will advance frontier research and provide a solid foundation for industrial applications.

Fine-tuning large-scale text-to-video diffusion models to add new generative controls, such as those over physical camera parameters (e.g., shutter speed or aperture), typically requires vast, high-fidelity datasets that are difficult to acquire. In this work, we propose a data-efficient fine-tuning strategy that learns these controls from sparse, low-quality synthetic data. W e show that not only does fine-tuning on such simple data enable the desired controls, it actually yields superior results to models fine-tuned on pho-torealistic "real" data. Beyond demonstrating these results, we provide a framework that justifies this phenomenon both intuitively and quantitatively.

Central to long-form text generation in vertical domains is the "impossible trinity" confronting current large language models (LLMs): the simultaneous achievement of low hallucination, deep logical coherence, and personalized expression. This study establishes that this bottleneck arises from existing generative paradigms succumbing to the Statistical Smoothing Trap, a phenomenon that overlooks the high-entropy information acquisition and structured cognitive processes integral to expert-level writing. To address this limitation, we propose the DeepNews Framework, an agentic workflow that explicitly models the implicit cognitive processes of seasoned financial journalists. The framework integrates three core modules: first, a dual-granularity retrieval mechanism grounded in information foraging theory, which enforces a 10:1 saturated information input ratio to mitigate hallucinatory outputs; second, schema-guided strategic planning, a process leveraging domain expert knowledge bases (narrative schemas) and Atomic Blocks to forge a robust logical skeleton; third, adversarial constraint prompting, a technique deploying tactics including Rhythm Break and Logic Fog to disrupt the probabilistic smoothness inherent in model-generated text. Experiments delineate a salient Knowledge Cliff in deep financial reporting: content truthfulness collapses when retrieved context falls below 15,000 characters, while a high-redundancy input exceeding 30,000 characters stabilizes the Hallucination-Free Rate (HFR) above 85%. In an ecological validity blind test conducted with a top-tier Chinese technology media outlet, the DeepNews system--built on a previous-generation model (DeepSeek-V3-0324)-achieved a 25% submission acceptance rate, significantly outperforming the 0% acceptance rate of zero-shot generation by a state-of-the-art (SOTA) model (GPT-5).

General-purpose audio representations aim to map acoustically variable instances of the same event to nearby points, resolving content identity in a zero-shot setting. Unlike supervised classification benchmarks that measure adaptability via parameter updates, we introduce VocSim, a training-free benchmark probing the intrinsic geometric alignment of frozen embeddings. VocSim aggregates 125k single-source clips from 19 corpora spanning human speech, animal vocalizations, and environmental sounds. By restricting to single-source audio, we isolate content representation from the confound of source separation. We evaluate embeddings using Precision@k for local purity and the Global Separation Rate (GSR) for point-wise class separation. To calibrate GSR, we report lift over an empirical permutation baseline. Across diverse foundation models, a simple pipeline, frozen Whisper encoder features, time-frequency pooling, and label-free PCA, yields strong zero-shot performance. However, VocSim also uncovers a consistent generalization gap. On blind, low-resource speech, local retrieval drops sharply. While performance remains statistically distinguishable from chance, the absolute geometric structure collapses, indicating a failure to generalize to unseen phonotactics. As external validation, our top embeddings predict avian perceptual similarity, improve bioacoustic classification, and achieve state-of-the-art results on the HEAR benchmark. We posit that the intrinsic geometric quality measured here proxies utility in unlisted downstream applications. We release data, code, and a public leaderboard to standardize the evaluation of intrinsic audio geometry.

Studying how recommendation systems reshape social networks is difficult on live platforms: confounds abound, and controlled experiments risk user harm. We present an agent-based simulator where content production, tie formation, and a graph attention network (GAT) recommender co-evolve in a closed loop. We calibrate parameters using Mastodon data and validate out-of-sample against Bluesky (4--6\% error on structural metrics; 10--15\% on held-out temporal splits). Across 18 configurations at 100 agents, we find that \emph{activation timing} affects outcomes: introducing recommendations at $t=10$ vs.\ $t=40$ decreases transitivity by 10\% while engagement differs by $<$8\%. Delaying activation increases content diversity by 9\% while reducing modularity by 4\%. Scaling experiments ($n$ up to 5,000) show the effect persists but attenuates. Jacobian analysis confirms local stability under bounded reactance parameters. We release configuration schemas and reproduction scripts.