Composability through modular components is key in digital transformation. Fastn's no-code/low-code AI platform simplifies system integration and API development, enhancing agility and scalability. Users can build a unified enterprise API and monitor application performance in real-time, attracting customers like HP. Despite competition from established players, Fastn aims to capitalize on the growing system integration market.

Getty Images releases open sample dataset on Hugging Face, offering 3750 high-quality images across 15 categories from its creative library for AI training. Dataset optimized for ML, ensuring legal compliance and removing inappropriate content. Compliance conditions apply. Aim is to enhance reputation in AI training data field.

Salesforce releases xLAM series of AI models, including xGen-Sales for self-sales tasks. xLAM-1B outperforms larger models like GPT-3.5, focusing on specific tasks and tool efficiency. Compact design allows for mobile use and are available on Hugging Face platform. Synthetic data used for training due to data gaps. Salesforce's move highlights the trend towards smaller, specialized AI models for improved performance.

Global data centers may emit 2.5B tons of greenhouse gases by 2030. New facilities for AI may contribute to this. Emissions could reach 600M tons annually by 2030. Industry investment in carbon capture and removal solutions is needed.

Replit launched an AI development assistant that autonomously builds applications, offering higher autonomy and proactivity than existing tools. Trained on data from millions of developers, it plans task steps and understands the entire development lifecycle. The tool, currently in beta testing, aims to enhance software development efficiency.

Apple is set to introduce "Image Playground" and "Genmoji" in iOS 18.2 update, enhancing emoji creation with AI technology. Expected release mid-December for more personalized digital communication.

Google expands AI-based virtual try-on (VTO) service to clothing category. Challenges faced in virtual try-on of dresses due to complex designs. Innovative "progressive training strategy" adopted. Virtual try-on technology aims to reduce return rates, enhance shopping experience, and preview real effects of products.

01.AI introduced Yi-Coder, a compact coding assistant challenging the industry's size-focused models. With efficiency and power, Yi-Coder reshapes AI coding, intensifies global competition, and leads sustainable development. Its impact on AI policy, collaboration with humans, and popularization of technology sets the stage for a new era in software development.

Red Hat Enterprise Linux AI (RHEL AI) platform launched, supporting generative AI in hybrid cloud. Integrates IBM's Granite series and InstructLab tool for cost-effective AI model development. Available on Dell PowerEdge servers, AWS, IBM Cloud, with plans for Google Cloud and Azure.

Harvard Medical School's AI, CHIEF, can diagnose multiple cancers with 94% accuracy. It's more flexible than other AIs, predicting patient treatment responses. Trained on 15M images, it outperforms others by 36%, reducing the need for DNA sequencing. Detection accuracy is 94%, reaching 96% for specific cancers. Future plans include training on rare diseases and predicting novel treatment effectiveness.

Microsoft to host OneDrive digital event on October 8th, focusing on AI innovations in Microsoft 365 and OneDrive. Updates include Copilot in OneDrive, mobile app enhancements, and photo experience optimization. Live-streamed on Teams platform for direct viewer interaction. Enterprise-focused Copilot fine-tuning and new features for Microsoft 365 expected. Subscription price set at $30 per user per month.

Anthropic launches Claude Enterprise, a high-end AI service to rival chatbot competitors. Features include large context window, GitHub integration, and enterprise-level security. Designed for larger teams, offering expanded capacity for messaging, bulk uploads, and collaboration with GitHub. Claude Enterprise does not train on customer data and will utilize Artifacts feature for handling larger projects. Price higher than $30/month for Claude Teams.

OpenAI plans to launch a high-priced ChatGPT subscription, targeting professionals and businesses with complex AI needs. New features like "Strawberry" and "Orion" models aim to compete with Google and Microsoft in the AI market. OpenAI is also seeking investments to push its valuation above $100 billion.

Qualcomm, Samsung, and Google's AR project progresses; focusing on smart glasses popularity. Qualcomm CEO aims for everyone to own smart glasses along with smartphones. AI to run on device and in cloud for new experience. Snapdragon X Plus 8-core chip released.

Google is testing "Ask Photos" to revolutionize photo library exploration. Gemini AI enables users to ask questions and receive intelligent photo answers. Additionally, natural language search is now available, enhancing image search efficiency on Android and iOS.

Since the launch of OpenAI's marketplace, numerous policy violations including AI pornography and cheating tools have emerged, raising concerns about content regulation. Despite removal of some violating GPTs, ineffective enforcement persists, highlighting the need for improved measures to ensure platform integrity.

Qualcomm launches octa-core Snapdragon X series CPUs for Windows PCs, introducing Copilot+ with significant price reduction to $799. High-performance NPUs meet Microsoft's 40 TOPS standard, challenging Intel and AMD in AI laptop market. Qualcomm aims to drive innovation for enhanced user experience.

AWS is a powerful cloud service with on-demand, scalable functionalities for easy application deployment. AWS Bedrock offers large language models from various sources for developers to use, test, and compare. Stability AI adds three image generation tools to Bedrock, each with unique strengths in quality, speed, and affordability, providing developers with versatile options.

Nyra Health unveiled the CrisperWhisper model, a groundbreaking speech recognition technology, featuring enhanced accuracy, noise resistance, and cognitive process support. Achieved remarkable performance in tests, promising wide application prospects in various fields.

Tencent Hybrid Turbo, based on MoE architecture, has 100% improved inference efficiency and 50% reduced costs. It outperformed in SuperCLUE tests, integrating into 700+ businesses and Tencent services, offering cloud versions like Turbo, Pro, and Lite, with exclusive services like code generation and function calls.