Google has updated its AI Overviews with a new "Save" feature and simplified language option. These updates are available for US users in English. Google is also testing a new link display feature and expanding AI Overviews to six new countries.

Geekbench AI, a tool for evaluating AI-intensive tasks, analyzes CPU, GPU, and NPU capabilities. It assesses performance using processing speed, accuracy, and supports frameworks like ONNX and CoreML. Scores consider accuracy and alignment with ground truth, crucial for assessing AI task completion. It's important to test devices with AI capabilities, like Copilot Plus PC and smartphones, to verify performance in real-world tasks. Geekbench AI is downloadable for Windows, macOS, Linux, Android,

Lambda and Nous Research collaborated to launch Hermes 3, boasting 405 billion parameters and "Amnesia Mode." Founded in 2023, Nous Research focuses on open-source code and large language models. Hermes 3 features long-term context retention, multi-turn dialogue management, and powerful proxy capabilities. It utilizes Lambda's 1-Click Cluster infrastructure and Neural Magic's FP8 quantization technique for efficiency. Offering free access to the AI community for testing.

Google releases Imagen 3, an AI text-to-image generator with improved details. Users can access it on Google's AI Test Kitchen for better image generation and editing capabilities.

Apple is developing a $1000 desktop robot with a large touch screen, Siri voice assistant, and home security features. Expected release in 2026.

From Google's Gemini to OpenAI's GPT-4o, generative AI models have varying levels of fictional content. A study found that models with less fiction perform better but trust should still be limited due to unreliable content.

Approximately 50% of job seekers are using AI like ChatGPT for resumes and applications. Employers are overwhelmed by low-quality AI-generated applications. Concerns arise about AI's impact on the recruitment process.

Snowflake introduces Cortex Analyst, a service using large language models for accurate insights. Users ask questions in English, system converts to SQL, queries data, and provides answers. Integration with Snowflake Data Cloud ensures accurate responses. Snowflake aims to enhance features for more interactive experiences.

Elon Musk's Grok launched an AI image feature with weak security, powered by Black Forest Labs' FLUX.1 model. Black Forest Labs surpassed competitors in image quality and plans to release a text-to-video model. Musk backs Black Forest Labs for its unrestrained approach and criticism of excessive security in AI models. The partnership has led to concerns about misuse and spread of AI-generated fake images and news. Musk's laissez-faire attitude toward information authenticity has stirred public d

Anthropic introduces prompt caching in API, enhancing developer efficiency. Detailed explanation available in 2023 paper. Early adopters see speed and cost improvements. Economic advantages include up to 10x cost savings. Lifespan of 5 minutes, refreshed when called.

Current mobile voice assistants still have room for improvement in understanding our requests. Google's Gemini Live offers a more human-like conversational experience, with 10 voice options. Accessible via Google's One AI Advanced plan at $20/month.

OpenAI's chatgpt-4o-latest model reclaimed the top spot in Chatbot Arena, surpassing Google's Gemini 1.5 Pro. The model excelled in various categories like math, programming, and multi-turn conversations, scoring 1314 points.

Kunlun Wanyi launched Melodio, the world’s first AI-based streaming music platform, and Mureka, a revolutionary AI music commercial creation platform. Powered by the Skymusic 2.0 model, these innovations redefine personalized music experiences and open new possibilities in AI music creation.

Elon Musk's xAI launched Grok-2 and Grok-2 Mini with enhanced reasoning capabilities. Grok can generate images on the X platform for advanced users and premium members. Musk's AI company plans to open these models to developers. Discussions on regulating image generation are intensifying. Grok uses FLUX.1 technology from Black Forest Labs. Grok-2 shows improvement in coding, writing, and summarization, but with unclear evidence. There is a risk of false information creation. xAI aims to integrat

Researchers collaborate to develop "AI Scientist," an AI system autonomously conducting scientific research, revolutionizing discovery process. Utilizes large language models for idea generation, experiment design, peer-review. Raises questions on human role in research, economic impact, and ethical considerations.

Opera has launched iOS version of One browser with optimized interface design. It introduces Aria AI chatbot for personalized browsing suggestions. Search bar moved to bottom for one-handed use. Color-adaptive UI enhances visual aesthetics. Improved search function offers intelligent suggestions.

Stability AI leads in generative AI with flagship product Stable Diffusion, focusing on stability and resilience. Strategic investments support innovation, including Midjourney platform and collaboration with Intel. Acquisitions like Init ML empower users with advanced content creation tools. Business strategy emphasizes foundational models, partnerships, and shaping AI innovation for a broader, more accessible future.

LG AI Research officially launched EXAONE 3.0, the first open-source AI model in South Korea. With a scale of 7.8 billion parameters and 80 trillion tokens, it excels in bilingual settings. The model showcases LG's AI expertise and aims to enhance global collaboration.

Humanoid robots are becoming a reality, integrating AI for enhanced interaction capabilities. Language models like ChatGPT play a key role in achieving "embodied intelligence." Google's PaLM-E boosts robots' learning efficiency through visual spatial intelligence. They have potential in manufacturing, logistics, and homes, but challenges like high costs and ethical issues need to be addressed. Experts' insights are crucial for navigating the future landscape of humanoid robots.

Google's new AI feature, Smart Notes, generates meeting summaries in real-time, highlighting key points for easy understanding. Only available to specific Google Workspace accounts.