A new AI tool, AlphaCodium, surpasses DeepMind's AlphaCode and AlphaCode 2 in code generation, using a 'process engineering' approach with elements of GAN architecture. Developed by CodiumAI, it boosts accuracy to 44% compared to GPT-4, setting a new standard in the field. Founded in 2022, CodiumAI aims to enable developers to build bug-free software faster, emphasizing the importance of code integrity in the coding race with OpenAI and Google DeepMind.

Equinix launches managed private cloud service for NVIDIA DGX AI supercomputing, catering to enterprises seeking advanced AI capabilities. The service provides flexible deployment within Equinix's IBX data centers, enabling quick adoption of AI infrastructure. Customers across various industries can leverage generative AI for innovation. Integration with NVIDIA AI Enterprise software ensures efficient development and deployment of AI applications.

Kimi Chat has been upgraded to moonshot-v1-20240123, enhancing intelligence for better online search and contextual learning. It offers improved education support, literary creation assistance, and faster translation. The update also includes voice input, faster verification codes, and enhanced chat features.

Amazon introduces Diffuse to Choose (DTC), an AI system enabling virtual product try-ons in any setting. DTC combines fast inference with high-fidelity details, offering enhanced shopping experiences and reducing return rates.

"Google's Bard (Gemini Pro) ranks 2nd in HuggingFace Chatbot Arena, beating OpenAI's GPT-4. Gemini Pro is Google's 2nd best, hinting at potential for Gemini Ultra. OpenAI may launch GPT-5 soon. Google integrates generative AI into Chrome, promoting "Help me Write" feature for text box assistance in writing tasks."

Researchers examined the diagnostic reasoning abilities of large language models like GPT-3.5 and GPT-4. Their study focused on open-ended clinical questions and highlighted the potential for these models in identifying complex patients. GPT-4 demonstrated superior performance compared to GPT-3.5, showing higher accuracy in various clinical reasoning prompts. This research is crucial for enhancing the credibility of LLMs in medical applications.

360 Group launches new "360 AI Search" app for Android, promising advanced search capabilities. CEO Zhou Hongyi plans to revolutionize the search market with AI technology, aiming to compete with industry giants.

The FTC mandates leading AI companies (Alphabet, Amazon, Anthropic, Microsoft, OpenAI) to disclose investment and partnership details. Investigation aims to analyze impact on competition, innovation, and potential anti-competitive practices. Companies have 45 days to respond.

BMW plans to introduce "humanoid" robots in its factories to assist in manufacturing. Developed by FigureAI, these "universal" robots are designed for various tasks and will focus on difficult, unsafe, or tedious tasks. CEOs believe they will increase productivity without replacing human workers.

OpenAI launches new embedding models like text-embedding-3-small, offering better performance and lower prices. GPT-4 Turbo and GPT-3.5 Turbo have been updated for improved text understanding and content generation. New text moderation model and API key management system are also introduced, with a 25% reduction in GPT-3.5 Turbo API access cost.

Google integrates generative AI in Chrome, launches "Help me write" feature. Sundar Pichai states that Chrome's new AI feature simplifies online tasks. Microsoft CEO acknowledges Google's dominance in search. Chrome enhances tab management with automatic grouping. AI-based theming in Chrome Web Store. Google ahead of OpenAI in multimodal tech with Lumiere video synthesis model.

Google Cloud partners with Hugging Face to enhance AI generation accessibility. Developers can now utilize Google's infrastructure for model building, including TPUs and NVIDIA GPUs. This collaboration enables seamless deployment of models from Hugging Face to Google Cloud, paving the way for exciting possibilities in AI development.

Samsung and Baidu Smart Cloud partner to promote AI technology on smart devices, enhancing user experience. Galaxy S24 series integrates Baidu's large model for empowered calling, translation, summaries, and "Circle and Search" feature. Collaboration brings upgraded note assistant function and local partners for Galaxy AI. This signifies a strategic move by Samsung to apply cutting-edge AI on terminal devices, enhancing competitiveness and business opportunities in the market.

The "InstantID" module by the InstantX team enables effortless and consistent character generation from a single facial image reference, without specialized training. It uses a new technique with facial recognition models for identity embeddings, allowing for style control and enhancing realism. Despite its creative potential, ethical considerations around consent and misuse must be addressed.

"Tavilly partners with LangChain to launch GPT-Newspaper, an autonomous agent project creating personalized newspapers based on user preferences. It uses AI to curate, write, design, and edit news content, revolutionizing news consumption."

SK Hynix returned to profitability in Q4 2022, reporting an operating profit of 346 billion Korean won, driven by strong sales of DDR5 and HBM3 memory chips. Its focus on AI storage has led to a positive performance, aiming to become a leading AI storage provider by 2024.

OpenAI introduces updates to GPT-4 Turbo, enhancing code generation. Users switching to GPT-4 Turbo exceed 70%. New text-embedding models released for improved natural language processing.

Quark App launched AI PPT feature for quick, professional, and easy PPT creation. Users can input topic or choose template for automatic outline generation. The feature utilizes powerful AI technology, surpassing GPT-3.5 in performance. It simplifies PPT creation process and enhances quality and efficiency. Expect more innovative features from Quark in the future.

ElevenLabs launches Dubbing Studio post $80M funding, offering advanced control over video localization. Featuring tools like transcription editing, customized translation, and precise timing adjustments, the platform enhances user experience for multilingual video content.

Oracle is integrating generative AI capabilities into its OCI stack. Plans include full launch of generative AI services, generative AI agents, and data science AI quick actions. OCI now hosts Llama 2 and Cohere's 52 billion embedded models. Oracle will provide pre-built agent operations in SaaS suites and introduce Code Gen Agent for Java apps. AI Quick Actions offer no-code solutions for LLM deployment using open-source libraries.