Dera News

Google DeepMind & Gemini

Google DeepMind research, Gemini model releases, AlphaFold, Veo, Imagen, AI Overviews, and Google's full AI ecosystem strategy.

15 articles
1 stories
Last 30 days
(36 total matching)
Google article thumbnail
3 sources
Google logo
Google

Google Gemini Rolls Out To Cars

Google’s Gemini AI assistant is hitting the road in millions of vehicles

Google's Gemini AI assistant is coming to vehicles with integrated Google systems, marking a significant upgrade from the previous Google Assistant. This move promises more natural conversational AI for drivers, enabling enhanced vehicle control and information retrieval. The rollout begins in the US, with plans for broader availability and language support, and will reach both new and existing compatible vehicles via software updates.

📰TechCrunch · The Verge
Google article thumbnail
Google and 1 others
Google

Metal-Sci: A Scientific Compute Benchmark for Evolutionary LLM Kernel Search on Apple Silicon

Authors: Víctor Gallego arXiv Links arXiv | PDF AI summary Abstract A benchmark for optimizing scientific computing kernels on Apple Silicon is paired with an automated search framework that uses a large language model in a evolutionary loop to optimize performance while identifying hidden regressions through out-of-distribution evaluation. AI-generated summary 摘要 Apple Silicon 上优化科学计算内核的基准与自动搜索框架相结合,该框架在进化循环中使用大型语言模型来优化性能,同时通过分布外评估识别隐藏的回归。 AI 生成的摘要 Abstract We present Metal-Sci, a 10-task bench...

Trending
Google article thumbnail
Google logo
Google

Gemini can now pull from Google Photos to generate personalized images

Google Gemini is getting a new feature that lets it tap into your Google Photos library to generate highly personalized images. This means Gemini can now learn your specific tastes and lifestyle from your photos, using that intel to create images tailored to your prompts, like designing your 'dream home' or illustrating 'essentials for a deserted island.'

Trending
OpenAI article thumbnail
OpenAI and 2 others
OpenAI

IntentGrasp: A Comprehensive Benchmark for Intent Understanding

Institution: University of British Columbia | Authors: Yuwei Yin, Chuyuan Li, Giuseppe Carenini arXiv Links arXiv | PDF AI summary Abstract IntentGrasp is a benchmark for evaluating large language models' intent understanding capability, demonstrating poor performance across 20 models and showing significant improvements with intentional fine-tuning. AI-generated summary 摘要 IntentGrasp 是评估大型语言模型意图理解能力的基准,展示了 20 个模型的较差性能,并通过有意的微调显示出显着的改进。 AI 生成的摘要 Abstract Accurately understanding the intent behi...

Trending
Google article thumbnail
Google and 3 others
Google

Meet GitHub Spec-Kit: An Open Source Toolkit for Spec-Driven Development with AI Coding Agents

If you have spent time using AI coding agents — GitHub Copilot, Claude Code, Gemini CLI — you have probably run into this situation: you describe what you want, the agent generates a block of code that looks correct, compiles, and then subtly misses the actual intent. This “vibe-coding” approach can work for quick prototypes The post Meet GitHub Spec-Kit: An Open Source Toolkit for Spec-Driven Development with AI Coding Agents appeared first on MarkTechPost.

Notable
Google article thumbnail
Google logo
Google

Google introduces "Skills" in Chrome to make Gemini prompts instantly reusable

Google Chrome just rolled out a new 'Skills' feature, making Gemini much more powerful for SMBs. This allows you to save and instantly reuse your most-used prompts, cutting down on repetitive input. From calculating nutrition to summarizing long documents, 'Skills' promises to streamline workflows and boost productivity, making it a must-try for any business leader.

Notable
OpenAI article thumbnail
OpenAI and 2 others
OpenAI

Soohak: A Mathematician-Curated Benchmark for Evaluating Research-level Math Capabilities of LLMs

A new benchmark called 'Soohak,' developed by 64 mathematicians, challenges the most advanced LLMs with 439 complex math problems. Initial results show even top models like Gemini-3-Pro and GPT-5 scoring low, highlighting a major gap in their ability to handle research-level mathematical reasoning and, crucially, to recognize and refuse to answer inappropriate questions.

Notable
Anthropic article thumbnail
OpenAI and 3 others
Anthropic

Running Claude Code or Claude in Chrome? Here's the audit matrix for every blind spot your security stack misses

Between May 6 and 7, four security research teams published findings about Anthropic’s Claude that most outlets covered as three separate stories. One involved a water utility in Mexico, another targeted a Chrome extension, and a third hijacked OAuth tokens through Claude Code. In one case, Claude identified a water utility’s SCADA gateway without being told to look for one. These are not three bugs. They are one architectural question playing out on three surfaces. No single patch released so f...

Notable
Google article thumbnail
Google logo
Google

Google DeepMind Releases Gemini Robotics-ER 1.6: Bringing Enhanced Embodied Reasoning and Instrument Reading to Physical AI

Google DeepMind has released Gemini Robotics-ER 1.6, a major upgrade to the 'cognitive brain' of robots. This new model drastically improves how robots understand their environment, plan complex tasks, and even read industrial gauges. It's designed to give robots better judgment for real-world operations, moving beyond simple task execution to more intelligent decision-making in diverse settings.

Notable
Google article thumbnail
Google and 2 others
Google

Spooked by Mythos, Trump suddenly realized AI safety testing might be good

In a surprising pivot, former President Trump has agreed with major AI players like Google DeepMind and Microsoft on the necessity of AI safety testing. This shift comes after Anthropic's decision to halt the release of its advanced 'Claude Mythos' model due to potential misuse risks, prompting a reevaluation of his 'no regulation' stance.

Notable
Google article thumbnail
Google logo
Google

DeepMind’s David Silver just raised $1.1B to build an AI that learns without human data

David Silver, a renowned former Google DeepMind researcher, has launched a new AI lab called Ineffable Intelligence. The company quickly secured a massive $1.1 billion in funding, reaching a valuation of $5.1 billion. Their ambitious goal is to develop AI models that learn independently, without relying on human-created data, marking a significant shift in AI development.

Google article thumbnail
Google logo
Google

Google adds Nano Banana-powered image generation to Gemini’s Personal Intelligence

Google's AI assistant, Gemini, just got a significant upgrade in its image generation capabilities. Thanks to 'Nano Banana' technology, Gemini can now tap into your Google account data from services like Gmail and Google Photos to create personalized images. This means less prompting and more intuitive image creation, potentially streamlining content generation for SMBs.

Notable
Google article thumbnail
Google and 2 others
Google

Three AI coding agents leaked secrets through a single prompt injection. One vendor's system card predicted it

A security researcher, working with colleagues at Johns Hopkins University, opened a GitHub pull request, typed a malicious instruction into the PR title, and watched Anthropic’s Claude Code Security Review action post its own API key as a comment. The same prompt injection worked on Google’s Gemini CLI Action and GitHub’s Copilot Agent (Microsoft). No external infrastructure required. Aonan Guan, the researcher who discovered the vulnerability, alongside Johns Hopkins colleagues Zhengyu Liu and...

Notable