Nous Research released open-source lightweight vision-language model Nous Hermes 2 Vision through Hugging Face. It combines visual and textual information to generate detailed answers. Despite initial hallucination issues, the company aims to provide a stable version soon.

Google is set to preview Gemini, a generative AI product rivaling ChatGPT, on the web this week. Delayed, Gemini promises to be powerful, challenging Bard and ChatGPT. CEO Sundar Pichai canceled live launch events due to reliability concerns with non-English queries. Gemini's full release may not be until 2024, but it's seen as a competitive move against Microsoft's OpenAI. Gemini is a multimodal AI with image, sound, and text processing capabilities, built on a reinforcement learning system.

Carnegie Mellon University and Hugging Face report on the energy impact of AI models, with some tasks emitting significant carbon dioxide. The shift towards multi-purpose models raises concerns about energy consumption.

OpenAI's enterprise version of ChatGPT aims to provide better data protection, customization options, and address concerns about AI not being a one-step solution for businesses. AI technology is still in its early experimental phase and not a complete solution for all problems.

Google's DeepMind team created an AI system for social learning. AI agents in a virtual world imitate "experts" to learn new skills efficiently. This approach mimics human learning, allowing agents to navigate obstacles and learn from experts to reach goals faster, showcasing social learning abilities.

The latest Google security blog post introduces a new Gmail spam filter upgrade called RETVec, improving defense against adversarial text manipulation. RETVec has significantly boosted spam detection rates and reduced false positives, showcasing efficiency gains and significant defense upgrades.

TSMC unveiled AWS Graviton4 and Trainium2 chips at re:Invent. It also manufactures chips for Microsoft, Google, Apple, NVIDIA, and Intel, emphasizing its pivotal role in AI ecosystem. TSMC's advanced technology drives AI industry growth, with plans for significant expansion globally. The industry's reliance on TSMC prompts concerns over potential risks and supply chain challenges, highlighting the company's dominance in semiconductor manufacturing.

iA Writer 7 introduces a new author feature to distinguish between your own writing and AI-generated text. This tool encourages using AI as a "conversation partner" to improve writing. Currently available on Mac, iOS, and iPadOS, with plans to expand to Windows and Android.

OpenAI's ChatGPT, based on transformer neural networks, sees rapid adoption, with McKinsey projecting $4 trillion in annual growth for generative AI. Debate over AI's safety leads to new regulations. OpenAI's mysterious Q* project raises concerns, with potential for AGI breakthrough.

Microsoft to invest £2.5 billion in AI infrastructure in the UK, expanding data centers and GPUs. This aligns with a global commitment of £29.5 billion to the UK. The rise of AI is driving demand for data centers, with Microsoft also expanding in Australia and Canada. The focus is on sustainability and economic growth, creating new jobs and supporting skills training.

Google delays launch of Gemini, its advanced conversational AI, to early 2024 due to non-English signal processing issues. Gemini touted for superior text and image generation. Anticipated release postponed for further development and improvement.

Open source software is widely used in major industries, but security vulnerabilities still exist. Speranza, a new system developed by MIT researchers, ensures software authenticity without revealing personal information, unlike traditional methods like digital signatures.

Apple unveiled 'Personal Voice' on International Day of Persons with Disabilities, enhancing accessibility with features like VoiceOver and Live Listen. Concerns arise about security and privacy with voice cloning technology.

IBM and Boehringer Ingelheim partner to utilize generative AI for therapeutic antibody development, aiming to enhance patient outcomes. IBM's AI technology will improve antibody discovery speed, reducing the process time to one year. This collaboration accelerates drug development, leveraging AI to revolutionize healthcare.

A group developed GAIA, a benchmarking tool for AI assistants, aiming to test AGI capabilities. They propose a rating system to measure AGI intelligence, emphasizing the creation of challenging questions humans find easy but computers struggle with. Testing their AI products, none passed the benchmark.

ByteDance to unveil "Robot Development Platform" for users to create chatbots. Also working on text-to-image generator. No official comments from ByteDance.

Baidu Intelligent Cloud introduces AI-native app "Super Assistant" for Wenxin Yiyuan, now in public beta. It offers analysis, copywriting assistance, document processing, conversational search, and full-text translation. It can be integrated into various systems as a one-stop portal.

On December 1, Kunlun WSV launched "SkyAgents", a leading AI Agents platform based on the "Tiangong Big Model". Users can easily create personalized AI assistants without coding, supporting various tasks and seamless integration into business systems.

Google introduces new features for Android, Wear OS, and Google TV. Updates include Emoji Kitchen for Android, Voice Moods for Android Messages, AI-generated image descriptions for visual impairments, and more for Wear OS. Additionally, Google TV adds ten new free channels.

Tencent Cloud released the world's first international standard for large-scale models in financial risk control, aiming to guide institutions in applying AI technology for efficient risk management. This standard was jointly developed by various research and financial institutions, with a focus on innovation in the financial field.