AI News

DeepL wants real-time translation to finally capture sarcasm and mood shifts View the LinkedIn Profile of Jonathan Kemper Nov 11, 2026 DeepL Voice's real-time translations are about to sound a lot more natural. The company's new Aure model carries over tone, emotion, emphasis, and speaking pace from the original speaker's voice into the target language. Until now, DeepL Voice generated translated speech from a static voice sample captured at the start of a

US data brokers are selling half a billion dollars in RL training environments to Chinese AI labs View the LinkedIn Profile of Maximilian Schreiner Nov 11, 2026 Nano Banana Pro prompted by THE DECODER Key Points At least nine US firms are selling $500 million worth of AI training data to Chinese labs, according to a ChinaTalk investigation. Chinese buyers use the same training environments built for OpenAI or Anthropic, cutting their own compute costs and

Anthropic says AI is eroding Nvidia's CUDA lead, cutting an AMD chip switch from months to three days View the LinkedIn Profile of Matthias Bastian Nov 11, 2026 Nano Banana Pro prompted by THE DECODER Key Points Anthropic claims it has overcome its dependence on Nvidia through a multi-chip strategy. According to co-founder Tom Brown, the company now trains and runs Claude on all four major chip families, treating them as a single, interchangeable compute p

Helsing built an AI system that prepares targets for weapons systems and says every decision will carry a human name View the LinkedIn Profile of Maximilian Schreiner Nov 11, 2026 Helsing Key Points Defense company Helsing has unveiled "Clarity," an AI platform that aggregates sensor feeds and reports to identify military targets and recommend weapons systems. Helsing emphasizes European sovereignty and mandatory human authorization for every strike, posit

Googles CodeMender is designed to automatically find and fix security flaws in software View the LinkedIn Profile of Maximilian Schreiner Oct 7, 2025 Midjourney prompted by THE DECODER AI News Without the Hype – Curated by Humans Subscribe to THE DECODER for ad-free reading, a weekly AI newsletter, our exclusive "AI Radar" frontier report six times a year, full archive access, and access to our comment section. Subscribe now

Google Deepmind's AlphaProof Code verified Signal's cryptographic core in four weeks instead of years Nov 11, 2026 Google Deepmind has unveiled AlphaProof Code, an AI system that mathematically proves software meets its security specs without errors. Researchers demonstrated the technology on Signal's encryption protocol, completing in four weeks work that typically takes years. Earlier verification efforts stopped at the theoretical design stage, accordin

US corporate AI spending hits a new high in October, but growth slows to 1.65 percent View the LinkedIn Profile of Matthias Bastian Nov 11, 2026 Nano Banana Pro prompted by THE DECODER Key Points US corporate AI spending rose just 1.65 percent month over month in October 2026 to a new high, according to the Ramp AI Index, with the top one percent of companies paying a median of $10,200 per employee per month while the typical company spent $16.32. One perc

Google joins the decision-model wave with Gemini Decision 1, scoring 1,000 questions in a single pass View the LinkedIn Profile of Matthias Bastian Nov 11, 2026 Google Deepmind is introducing the Gemini Decisions API, an interface built for fast, closed-ended decisions like yes/no questions, multiple choice, and scale-based ratings. The underlying model, Gemini Decision 1, processes up to 1,000 independent questions in a single pass across text, images, au

Everything we know so far about the Nvidia spy scandal targeting CEO Jensen HuangSee all StoriesPosted Nov 12, 2026 at 4:00 AM UTCRExternal LinkRichard LawlerIf Jensen Huang is done at Nvidia, who will be the next CEO? Johny Srouji?With reports that the company’s board is meeting to oust Huang from the company he co-founded, Bloomberg reports the board has been working with an unnamed outside firm to identify candidates and is considering naming a temporar

Scaling up is a hallmark of large models and has driven leaps in language model technology over the past few years. Yet in fact, large language model architectures, as their most fundamental computational foundation, have also continued to evolve in recent years. With the introduction of GQA, MLA, and various sparse attention modules; linear attention networks such as Mamba and Gated DeltaNet; DeepSeek’s fine-grained MoE; and multi-token prediction and Loo

Editor | PandaLately, while scrolling through Douyin, you’re likely to come across a type of video like this: three to five minutes long, with no on-camera presenters or live-action footage, just words and graphics that continually grow, break apart, and recombine on screen. For example, the widely known line “Bright moonlight before my bed” is used to explain what “surprise” is, leading into the concept of “information theory”; a few stick figures + a few

To meet users’ need to independently select models for different tasks, Doubao Work officially integrated multiple third-party models, including Kimi 3.1 Pro, GLM 5.3, and DeepSeek V4.1 Flash, on November 12, and launched a limited-time special offer. Users can select the appropriate model to complete tasks based on their actual work needs.Previously, Doubao Work had made third-party models available to enterprise customers. To help users better understand

The 3rd Wiley Star Scientist Award (Wiley Rising Star Scientist Award) officially launched and opened for applications on November 7, 2026. This year’s award focuses on the field of materials science, while continuing to offer the AI Special Award. The Journal Ranking Table Team of the National Science Library, Chinese Academy of Sciences will provide support for the selection process, and renowned scholars in relevant fields will be invited to form the ex

In robotic manipulation research, Vision-Language-Action (VLA) models and World Action Models (WAM) provide robots with different pathways for action generation. However, when a task extends from a single pick-and-place operation to continuous construction and bimanual coordination, success depends not only on whether the current action can be completed, but also on whether the conditions established earlier can continue to support subsequent execution.An

By incorporating spatial understanding into the continual learning process of foundation models, can robots better cope with changes in the real world?As a robot reaches out to grasp a fruit or vegetable, someone moves the target away. An object that was already in place is thrown back onto the table. In another set of experiments, the tabletop continuously rises and lowers, and the gripper must keep up with changes in the target’s height. In both grasping

@suwa_nobuinJQIT CorporationIf you write “This code is yours” in CLAUDE.md, will it consolidate duplicate code? I tested it 132 times with Claude Code, Codex, and Gemini CLI, and only Claude consolidated it.AIAI agentsClaudeClaudeCodeGenerative AI4Last updated at 2026-11-12Posted at 2026-11-12When you ask AI to add a feature, it copies existing code, resulting in yet another similar function. This is a common sight in code reviews. “AI usage generates a fr

@blinkgroup_jp(Blink Group)What Is the Difference Between MCP and A2A? Their Roles in AI Agent Architecture and Criteria for Choosing Between ThemAILLMMCPArchitectureSystem Design0Posted at 2026-11-12 Challenges Readers Face When considering the introduction of AI agents and system integration, there is an increasing number of cases in which people struggle with the differences and how to choose between the concepts of "MCP (Model Context Protocol)" and "A

@Aki-Dortu(Aki Dortu)AWS's Major AI Services, Organized by Use Case [2026 Edition]AWSAIGenerative AIAmazon BedrockMachine Learning0Posted at 2026-11-12 Introduction AWS offers a wide range of AI-related services, from services for developing generative AI applications to image recognition, speech processing, and machine learning model development. However, the roles of these services vary considerably. Some use existing AI models, some train custom models,

@alibabacloud-saFrom the “Car Wash Problem” to an Answering System Based on Multi-Model Consensus: End-to-End Implementation with Model StudioModelStudioLLMAIAgentTAG24@@TAG260

@Hideki_Tamae(Hideki Tamae)6. When I counted how many times the trigger fired between the previous article publication and this run, the count was this. The number of articles produced was 0.AIAI agentClaudeCodeHarnessautomation0Posted at 2026-11-12 Introduction 6. This is the number for this run. The last time an article was published to Zenn/Qiita was November 8, 2026 at 15:29 (JST). Today, this scheduled task ran on November 12, 2026 at 9:40 (JST). Duri