EU Launches Multilingual AI Model Teuken-7B Supporting 24 Official Languages
2024-11-29
Teuken-7B, an open-source language model supporting all 24 EU languages, was launched on Hugging Face under the EU’s OpenGPT-X initiative. It features balanced multilingual training with half its data from non-English European languages and is evaluated by the European LLM Rankings, advancing multilingual AI technology.
Read more ›Tesla Optimizes Optimus Humanoid Robot Hand Design
2024-11-29
Tesla upgraded its Optimus robot hands to 22 degrees of freedom plus 3 in the forearm, enhancing dexterity. Soft protection allows secure handling of delicate items. Actuators moved to forearms for a compact design and smoother motion. Tactile sensors integration is planned to boost efficiency. Future Optimus models will feature this design, expanding application potential.
Read more ›SoftBank to Acquire OpenAI Employee Shares in $1.5 Billion Deal
2024-11-29
SoftBank Vision Fund 2 to buy $1.5B in OpenAI shares at $210 each, valuing OpenAI at $157B. This deepens SoftBank's AI investments and offers employee liquidity. OpenAI expects $3.7B revenue, $5B loss, raised $13B from Microsoft. AI market to surpass $1T in a decade.
Read more ›Lenovo Unveils Industry's First IT Service AI Agent for SMEs — Lenovo BAIYING
2024-11-29
Lenovo launches Baiying Intelligent Agent, the first AI-driven IT service for SMEs. Powered by Qingtian 3.0, it offers AI Marketing, Office, and Services to enhance market reach, efficiency, and reduce costs. Public beta available for six months at 3,588 RMB/year.
Read more ›Amazon to Launch Multimodal AI Model Olympus Next Week, Challenging Anthropic
2024-11-29
Amazon is launching Olympus, a multimodal large language model supporting text, images, and videos, next week. It may debut at AWS re:Invent and be available through AWS Bedrock, potentially reducing reliance on Anthropic.
Read more ›Orange and OpenAI Sign New Agreement for Direct Access to AI Models
2024-11-28
Orange partners with OpenAI for direct access to AI models via European data centers, enhancing solutions and AI-driven customer service. The collaboration supports responsible AI use and innovation in African languages.
Read more ›Microsoft Integrates Azure and LlamaParse with GPT-4, Redefining Retrieval-Augmented Generation Workflows
2024-11-28
Microsoft integrates Azure OpenAI GPT-4o with LlamaParse Premium and Azure AI Search, enhancing Retrieval-Augmented Generation. This enables efficient data parsing, AI-driven search, high security, and scalable solutions for industries like healthcare and marketing.
Read more ›Motorola Launches New "Moto AI" Suite with Three Key Intelligent Features
2024-11-28
Motorola launched Moto AI, featuring notification summaries, audio transcription, and photo info extraction on select high-end smartphones. Despite initial demo shortcomings, Motorola is refining the AI suite to enhance user experience.
Read more ›Google Launches GenChess: AI-Powered Customized Chess Experience
2024-11-28
Google Lab launched GenChess, an online chess variant with personalized pieces via text prompts. Users customize themes and styles, then compete against AI with adjustable difficulty and timings. It’s part of Google’s AI and chess initiatives, including collaborations and upcoming Gemini app features.
Read more ›Alibaba Launches QwQ-32B Model, Challenging OpenAI's Inference Technology
2024-11-28
Alibaba's QwQ-32B-Preview, a 32.5B parameter AI model supporting 32k-word prompts, outperforms OpenAI's o1 models in benchmarks. It excels in AIME and MATH tasks, has self-fact-checking, but may switch languages or loop. Available on Hugging Face under Apache 2.0.
Read more ›360 Unveils Multimodal Content Creation Engine "Nano Search" with Li Xueqin and Xu Zhisheng
2024-11-28
360 Corporation launched its multimodal AI engine "Nano Search" in Beijing, featuring search, read, write, and create functions. The event included a short drama and customizable voice options, showcasing AI's ability to transform search into creative experiences and emphasizing accessible AI to drive a new industrial revolution.
Read more ›Kunlun Wanwei Releases Tiangong Large Model 4.0 o1 Version Featuring Integrated Chinese Logical Reasoning
2024-11-28
Kunlun Wanwei Group launches Skywork o1 series, including TianGong Model 4.0, the first Chinese AI model with logical reasoning. The series—Open, Lite, and Preview—offers enhanced reasoning, diverse industry applications, and superior performance.
Read more ›AI2 Launches OLMo 2, a Fully Open-Source Language Model
2024-11-28
AI2 launches OLMo 2, a fully open-source language model with 7B and 13B parameters that outperform LLama 3.1 and Qwen 2.5. It includes model weights, tools, datasets, and training guidelines, utilizing a two-phase training strategy. Available under Apache 2.0 on Ai2 and Hugging Face.
Read more ›OpenAI Files Trademark Application for New AI Model "o1"
2024-11-28
OpenAI files a USPTO trademark for its AI model o1, pending approval. o1 is its first reasoning model. The company has applied for around 30 trademarks and is winning a legal dispute over the Open AI name.
Read more ›Elon Musk's xAI Company Plans to Launch Independent App to Rival OpenAI
2024-11-28
Elon Musk's xAI plans to launch its Grok chatbot independently by December to compete with ChatGPT. Grok is accessed via X subscriptions. xAI seeks $5B in funding, potentially doubling its valuation to $50B and offering 25% stakes to Twitter investors, aiming to raise a total of $11B.
Read more ›OpenAI Introduces SearchGPT; ChatGPT iOS Update Enhances Search Experience
2024-11-27
OpenAI launched SearchGPT, an AI-powered search engine integrated into iOS ChatGPT via Apple Shortcuts. It offers efficient web searches with relevant links and real-time info. Available to ChatGPT Plus and Team users.
Read more ›Artist Groups Protest OpenAI's Exploitation, Leak Sora Video Generation AI Model
2024-11-27
Artists have criticized OpenAI’s unreleased Sora AI, alleging exploitation as unpaid creators for marketing. In an open letter, they demand fair compensation and genuine artistic support. OpenAI has temporarily suspended Sora’s early access.
Read more ›Hugging Face Releases SmolVLM: Lightweight Vision-Language Model Promotes AI Accessibility
2024-11-27
Hugging Face launches SmolVLM, a 2B-parameter visual-language model optimized for on-device use. It outperforms larger models in efficiency and speed, enabling deployment on laptops and consumer GPUs without high resource demands.
Read more ›Inflection AI Shifts to Enterprise Services, Abandons Cutting-Edge Model Competition
2024-11-27
After Microsoft acquired Inflection AI and its CEO, the company pivots from developing cutting-edge AI models to focusing on enterprise AI services. Recent acquisitions enhance their toolset, while facing FTC scrutiny over competition.
Read more ›OpenAI's Video Generator Sora Leaked
2024-11-27
A leaked unauthorized front-end for OpenAI's unreleased Sora API appeared on Hugging Face, allowing video generation with early access tokens. OpenAI and Hugging Face quickly revoked access. The group alleges OpenAI exploited testers for PR without compensation and restricted content sharing. OpenAI states Sora is in research preview, addressing safety and technical issues amid rising competition.
Read more ›