Quitbot, developed by Fred Hutch Cancer Research Center, is an AI chatbot integrated with a free mobile app to assist users in quitting smoking. Designed with a large language model, it offers step-by-step guidance, tools to manage cravings, motivational support, and personalized interaction to help users overcome addiction.

Keda Xunfei announced the release of Starfire Cognitive Model V3.5, boasting improvements in logical reasoning, text generation, math answering, and learning capabilities. The model is based on "Feixing No.1" computing power. Their Chairman aims to surpass ChatGPT4.0 by 2024.

MIT's Artificial Intelligence Laboratory developed PRISM Neural Network and Logistic Regression models for early detection of pancreatic cancer. PRISM can identify 35% of cases in high-risk situations, a significant improvement. The models use extensive electronic health records data for reliable risk assessment. The team aims to expand the model's applicability internationally.

Quark App upgrades modules & tools, launches "MetaMind" for intelligent search & QA. New products include PC version & Quark Listening. Focus on AI use in health & education, emphasizes data security. Continuous innovation to enhance user experience & meet demands.

Global stock prices of NVIDIA and AMD surge as businesses demand AI-building chips. NVIDIA leads in high-end chips with H200 for AI training. AMD invests in AI technology. AI hype drives stock market success, with Microsoft excelling in Gen AI. NVIDIA's H200 supports Gen AI and HPC workloads. AMD, under Su's leadership, sees financial recovery. AI chip market growth boosts AMD sales. Meta, OpenAI, and Microsoft plan GPU purchases. Companies to expand global AI businesses in 2024.

Fu Sheng, Chairman of Cheetah Mobile and Liehu Xingkong, unveiled the AI open class on "Innovation in the AI Wave." He introduced Orin-14B, a free, open-source base model with 14 billion parameters. The model, based on MoE architecture, showcased high accuracy and performance balance. Liehu Xingkong's large model INT4 reduces parameters by 70% and improves speed by 30%. Fu Sheng predicts the decline of hundred-billion-parameter models. Liehu Xingkong also released the "Fine-tuning Toolbox" to he

GitHub introduces GitHub Copilot Chat, a natural language-driven coding tool powered by GPT-4. It provides real-time guidance, supports multiple languages, and aids in code translation and optimization. Developers can leverage it for seamless coding workflows and rapid innovation.

AI models may poison themselves by forgetting initial data, leading to model collapse. Grok, a chatbot from Elon Musk's xAI, plagiarized OpenAI. Error messages and ethical issues highlight the potential dangers of AI models. It is crucial to address these issues before boasting about solving human problems.

CES 2024 star product Rabbit partners with Perplexity for AI integration in R1 gadget. Customers get real-time updated answers without data cutoff. Subscribers receive access to updated LLMs like GPT-4. Rabbit R1 features touchscreen, scroll wheel, and camera for various applications. Perplexity AI combines search with LLMs to compete with Google and Bing.

Meta CEO Mark Zuckerberg announced the development of open-source AGI to bring FAIR and GenAI closer, aiming for a universal intelligence available to all. Leading AI advancements like Llama 3 and merging AI with metaverse are also in progress.

Unilever partners with "Be My Eyes" to introduce AI-assisted cooking for blind shoppers. The project aims to enhance accessibility assistance using AQR technology by Zappar. Unilever plans to expand this initiative to more brands, enhancing user experience and inclusivity.

Concerns about AI regulation span from misinformation to job loss & even human extinction. The EU & US have taken differing stances, w/ the White House setting guidelines for AI vendor transparency. Feasible regulatory approaches include copyright protection, attribution to creators, content verification, and AI vendor accountability. Policymakers must consider both technical and economic feasibility in regulating AI.

Meta's CEO Mark Zuckerberg announced plans to develop Artificial General Intelligence (AGI) for developers. The company aims to open-source this technology, with around 350,000 NVIDIA AI chips by year-end. Zuckerberg anticipates AI integration with the metaverse, emphasizing the importance of AI in future tech advancements.

Tencent released VideoCrafter 2, an upgraded video generation model with improved lighting and shadow effects. It overcomes data limitations to produce high-quality videos using unique algorithms. Key enhancements include improved visual quality, enhanced dynamic effects, and optimized concept combination.

Stability AI unveils Stable LM 2 1.6B, a compact multilingual model excelling in natural language tasks. Pretrained on diverse datasets, it outperforms larger models in various benchmarks. Its efficiency and accuracy make it a valuable tool for developers.

Meta CEO Mark Zuckerberg backs the pursuit of AGI, merging FAIR and GenAI teams for Meta's AI projects. Meta's massive computing resources hint at a leading position in the AGI race. Zuckerberg envisions AGI integrated into the Metaverse through Meta's Ray-Ban glasses. Committing over $15 billion annually, Meta aims to develop the Metaverse. The concept of open-sourcing AGI may challenge OpenAI's strategy.

NVIDIA's ChatQA models aim to rival GPT-4, with ChatQA-70B outperforming GPT-3.5-turbo and reaching GPT-4 level without ChatGPT synthetic data. A two-stage fine-tuning method enhances large language models in conversational QA. NVIDIA's approach matches state-of-the-art query rewriting models, with significant progress shown in challenging scenarios. Other companies like Google and Mistral are also advancing towards GPT-4-level capabilities.

Researchers are enhancing reliability of language models via Google AI's new framework ASPIRE, enabling self-evaluation for improved prediction accuracy and uncertainty warnings. ASPIRE outperforms existing methods on question-answering tasks, showing potential for critical decision-making applications like healthcare and law.

Foxconn partners with HCL Group to establish semiconductor operations in India, investing $37.2 million for a 40% stake. This marks Foxconn's first step in outsourced assembly in India, supporting Apple and Xiaomi. Foxconn plans to use its BOL model and strengthen India's supply chain. HCL Group sees this collaboration as in line with India's "Make in India" vision.

ChatGPT excels in programming, healthcare, and education sectors. It assists programmers in coding tasks. In education, it aids course design and personalized teaching plans. In healthcare, it assists with administrative tasks and medical histories.