Adobe rewards content creators $3 per minute for training its new AI video generator. Seeking videos of everyday behaviors to enhance generative AI capabilities. Potential earnings up to $7.25 per minute. Challenges include quality control and artists' rights.

Mistral AI launches Mixtral 8x22B, a large-scale language model rivaling OpenAI's GPT-3.5. With $415M in funding, Mistral aims to outperform competitors in AI research by providing access to innovative models like Mixtral 8x22B via platforms such as Hugging Face.

At the "Cloud Next 2024" conference, OPPO and OnePlus announced integration of Google's Gemini AI into their devices. This move promises a more intelligent smartphone experience, showcasing their deep commitment to AI technology. Gemini AI offers language processing, intelligent recommendations, and personalized services. This collaboration highlights their innovative approach and strengthens their position in the global market.

Cohere's Command R+ model outperformed GPT-4 on Arena and is on HuggingChat. It emphasizes enterprise needs over consumer chatbots. Cohere models can be customized and will be on Oracle Cloud and Microsoft Azure. Oracle and Cohere collaborate closely, enhancing AI deployment capabilities. Oracle's interest in acquiring Cohere is possible.

The innovative Humane AI Pin is a screenless smartphone operated by an AI assistant, CosmOS. While compact and practical, it falls short in performance and usability, making it not recommended for purchase in its current state. Despite its unique design, AI Pin struggles with basic tasks, overheating, and poor battery life. While promising, it is not yet practical for everyday use.

New research from New York University showcases AI models learning from baby babbling, highlighting the potential of observing child experiences in AI development. The study suggests AI can effectively learn vocabulary associations with limited data. Models like PLATO demonstrate the importance of child development research in advancing AI understanding of the physical world. Another breakthrough innovation is UCLA's Chatterbaby, an AI app interpreting baby cries with 90% accuracy.

Udio, a new AI-driven music program, generates professional-quality music in less than 40 seconds. Developed by former Google DeepMind researchers, it has impressed industry professionals with its sound quality and usability. Industry experts praise Udio as the most authentic AI music composition tool available, creating surprising and unexpected music while maintaining musical fidelity and coherence.

Parler-TTS, an open-source TTS library, prioritizes ethical considerations while achieving high-quality speech synthesis. Trained on 10,000 hours of audiobook data, it shows efficient data utilization and innovative model architecture. The team's decision to make it open-source fosters global collaboration and sets a benchmark for responsible AI development, emphasizing the importance of ethical standards in technology advancement.

OpenAI unveils GPT-4 Turbo with Vision, featuring improved performance, JSON mode, and ChatGPT integration. GPT-4 Turbo is a powerful multimodal model with enhanced capabilities and cost-efficient token pricing. OpenAI hints at Voice Engine and upcoming GPT-5 release, showcasing innovation in AI.

Baidu Cloud introduces "Super Canvas", an AI tool for generating creative portraits in just minutes. It enhances efficiency in template development and provides a personalized experience for users, enriching the photography industry. Expect more AI tools in the future.

Udio, also known as Sora, is a revolutionary AI-generated music platform created by former Google DeepMind researchers. It enables users to easily create high-quality music by inputting genre descriptions, lyrics, and artist inspirations. Despite concerns about copyright issues, Udio's founders are confident that their product will empower musicians to monetize their music.

Meta introduces new MTIA chip to enhance AI workloads with improved performance and efficiency. The chip features an 8x8 grid of processing elements for optimized computation and memory usage. It also supports PyTorch 2.0 integration and high-performance code generation with Triton-MTIA compiler backend. Preliminary results show three times performance improvement over the previous generation. MTIA is deployed in data centers and works in conjunction with GPUs for optimal performance.

Starting May 15th, Google Photos introduces AI editing tools to a wider audience for free. Features like Photo Unblur, Magic Eraser, and Portrait Light aim to enhance photo editing. Magic Editor, available to Pixel owners and Google One subscribers, offers advanced editing capabilities. Minimum device requirements apply.

Meta plans to launch Llama 3, a next-gen language model suite, with different functionalities, potentially competing with GPT-4. Llama 3 aims to power various Meta products, enhancing generative AI capabilities. The model is expected to support multimodality and range in size from 7 billion to 140 billion parameters.

"Researchers develop humanoid facial robot Emo, capable of mimicking human facial expressions to enhance social interaction. Equipped with 26 motors and 26 actuators, Emo utilizes predictive algorithms trained on human facial expression videos. Interchangeable silicone skin and high-resolution RGB cameras for customization and mimicry."

Microsoft's CEO proposed a new Turing test focusing on AI's real-world impact, tasking AI to turn $100,000 into $1 million. Experts see potential in business applications but warn of possible misuse and economic consequences.

Google to unveil cloud computing developments, Gemini chatbot, AI tools. Google Vids for video creation. Gemini Code Assistant competes with GitHub. Workspace updates include AI features. Imagen 2 for AI-driven image generation. Google introduces Vertex AI tools, NVIDIA's Blackwell platform coming in 2025. Chrome Enterprise Edition Advanced for enhanced security. New Gemini 1.5 Pro model. Open source tools for generative AI projects. Limited info on Arm processor Axion.

Ant Group's CodeFuse platform introduces "Graph Code" feature to generate code from design diagrams, enhancing front-end development efficiency. Over half of engineers use CodeFuse, with 10% of code AI-generated. AI programming trend is on the rise, with up to 75% of engineers expected to utilize AI assistants by 2028. CodeFuse leverages AI to cover entire R&D process, aiming to boost productivity. AI-human collaboration is key for future R&D advancements.

Google integrates Imagen 2 into Vertex AI, offering advanced image generation for developers. New features include Text-to-Live Image, AI-generated watermarking, and enhanced editing capabilities. Visual captions and VQA enhance platform versatility.

Apple restricts Vision Pro's camera access for privacy. Developers face hurdles, but users appreciate the privacy protection. App development limited due to privacy concerns post Google Glass AR headset controversy. Users wary of camera access due to data privacy concerns. Apple remains vigilant on privacy measures. Vision Pro's high price criticized.