Snapchat Plus subscribers can now use generative AI to create images from text prompts and adjust photo backgrounds. AI features include "Dreams" for selfies and chatbot integration with ChatGPT filters.

Tesla unveils new Optimus robot with improved flexibility, weight reduction, and enhanced hand dexterity. However, challenges remain for achieving Tesla's goal of a fully autonomous humanoid robot due to the need for advanced artificial intelligence.

The Wenxin Plugin Mall has been officially launched by "Baidu AI," offering plugins for office efficiency, content understanding, and information retrieval. Users can access plugins for PPT generation, audio and video extraction, and mind map creation with "one-click installation." The mall includes mind map tools, PPT generators, and a "Video Assistant" for extracting key information from videos. Users can also join the Wenxin Plugin Developer Community.

Stable AI unveils Stable Zero123, a cutting-edge generative AI model for creating high-quality 3D images from photos. It requires 24GB VRAM for optimal performance and is available on Hugging Face for research purposes only. The model sets a new standard in 3D image generation, utilizing advanced training datasets and techniques to produce superior visual outcomes.

Recently, Zerodha's CEO Nithin Kamath discussed the risks of deepfakes in finance, highlighting the difficulty in distinguishing AI-generated individuals. World ID 2.0, a digital passport by WorldCoin, aims to verify human identity online while prioritizing privacy and inclusivity. However, concerns arise regarding potential government abuse and surveillance. World ID 2.0 seeks to address these issues while offering secure and private human identity proof.

DeepMind's latest achievement involves a fact-checked LLM called FunSearch, solving mathematical problems by generating code. It surpasses human capabilities and can lead to new scientific discoveries.

Intel announces release of AI PCs at "AI Everywhere" event, partnering with Dell, HP, Lenovo, Supermicro, and Microsoft. New Intel Core Ultra chip and Intel Arc GPU feature up to 8 XE GPU cores. Intel aims to integrate NPUs into chips and enable seamless AI integration in various applications. AMD also reveals AI PC concept at Advancing AI event, introducing Instinct MI300X AI accelerator, ROCm updates, and Ryzen AI PCs.

Meta launches early access testing for Ray-Ban smart glasses AI features. Mark Zuckerberg demoed glasses suggesting clothing and translating text. Multimodal AI can answer questions. Chief Technology Officer showcased features like tagging photos and text translation.

CoRover.ai partners with Google Cloud to launch BharatGPT, a generative AI chatbot for the Indian market, catering to linguistic diversity. Offering features like customized knowledge base, payment gateway, and more, it aims to reflect India's cultural heritage. Supported in 14 Indian languages, it enhances India's AI capabilities, aligning with the vision of "Make AI in India, for India."

Microsoft researchers enhance GPT-4 using Medprompt, achieving record scores on MMLU benchmark. Medprompt combines 3 strategies: dynamic few-shot selection, self-generated chains of thought, and choice-shuffle ensemble. Medprompt+ achieves milestone 90.10%. Microsoft open-sources promptbase to optimize GPT-4 prompts. Their innovation surpasses Google's upcoming Gemini Ultra. GPT-4 demonstrates expert-like intelligence with the right guidance, emphasizing the need for advanced prompting strategie

Microsoft Research releases Orca 2 LLM, surpassing larger models using novel "Prompt Erasure" technique. Orca 2 improves student LLM using teacher guidance, outperforming baselines in diverse tasks and providing efficient model training. Users praise simplicity and effectiveness of Orca 2 approach, recommending further exploration. Orca 2 models available for use on Huggingface.

"Paris-based Mistral AI secures $113M funding for Mixtral 8x7B model. Licensing under Apache 2.0, the model outperforms Llama 2 70B. Plans paid API endpoints, challenges OpenAI, but does not support EU AI Act."

Google unveils Imagen 2, an advanced text-to-image AI model focusing on API development for developers. It offers improved image quality, better context understanding, and enhanced rendering of challenging areas. New features include style reference images and multi-language support. Responsible development includes security testing and imperceptible watermarking. Now available on Google's Vertex AI platform for select customers. Snap, Shutterstock, and Canva are early adopters.

Google released Gemini Pro, a powerful AI service supporting text and image input/output. Gemini Pro offers various functionalities, SDK for cross-platform development, and supports 38 languages worldwide. Gemini Pro and Gemini Pro Vision are available for free through Google AI Studio. Gemini will also be integrated into other Google development tools, and Gemini will debut on Duet AI for Workspace in early 2024.

Google DeepMind's AlphaCode 2 leads in competitive programming with advanced AI, outperforming 85% of competitors, but faces limitations in trial and error. Unique features include code sampling, clustering, and diverse solution selection, enhancing performance in solving coding problems.

Midjourney, popular AI model, testing alpha web version for direct image generation. Users can explore, adjust prompts visually, and view generated images chronologically. Competes with Visual Electric. Improved user experience but no image synchronization with Discord.

Axel Springer partners with OpenAI to feature global news content in ChatGPT responses. Users will access summaries from Axel Springer's media brands like Politico and Bild. The collaboration aims to enhance AI-driven journalism with real-time news access.

Lean Co-pilot, created by LeanDojo and Caltech, integrates LLM to offer accurate mathematical proofs. It automates Lean proof strategies, balancing AI and human input, enhancing Lean's theorem proving capabilities.

Oracle and AMD deepen partnership in cloud computing, integrating AMD's new MI300X GPU into Oracle Cloud Infrastructure. AMD also partners with Microsoft and Meta. Oracle collaborates with NVIDIA for AI. The multi-cloud strategy positions Oracle as a leader.

Grok, Elon Musk's AI platform, sparks discussion for similarity with ChatGPT and mention of OpenAI. Musk denies Grok using OpenAI code. Grok provides real-time info advantage over ChatGPT. Controversy arises over potential training overlap, addressed by xAI. Importance of transparency in AI model training emphasized.