November 19, 2024: Microsoft Unveils New AI Tools at Ignite 2024 - Microsoft announced new AI capabilities for developers at Ignite 2024, including the public preview of Copilot Studio for building autonomous AI agents. These agents can perform complex tasks using simple English prompts. Enhancements for Power Platform improve developer experience, security, and automation without coding. Additionally, Azure AI Foundry offers new tools for evaluating and benchmarking generative AI models using enterprise data, ensuring models remain accurate and trustworthy. These updates empower developers to create intelligent applications while maintaining robust governance and security.

November 19, 2024: Microsoft Launches All-in-One Azure AI Foundry - At its Ignite conference, Microsoft unveiled Azure AI Foundry, consolidating its enterprise AI services under one platform. Anchored by Azure AI Studio, Foundry integrates Microsofts model catalog with tools like Azure AI Search and introduces the Azure AI Foundry SDK, offering 25 prebuilt app templates for custom AI solutions. The platform aims to enhance AI deployment efficiency, support collaboration, and align technical and business strategies, while also debuting the Azure AI Agent Service, enabling secure and automated business process orchestration.

November 19, 2024: Microsofts Copilot Expands with AI Agents Everywhere - Microsoft is significantly enhancing its Copilot suite by integrating agentic AI capabilities across its products. The updates, announced at Microsoft Ignite, include Copilot functionalities in SharePoint, Teams, Outlook, and more, enabling automated, task-oriented operations. New tools like Facilitator and Project Manager agents will revolutionize meeting efficiency and project management. Developers can utilize an SDK for creating AI agents, while updates in Copilot Studio facilitate easier agent creation. These advancements aim to transform the digital workplace with autonomous and semi-autonomous activities, enhancing efficiency and collaboration in Microsoft 365 and other services.

November 19, 2024: Google Lens Enhances In-Store Shopping - Google Lens now offers real-time price checks, product insights, and inventory data for in-store shopping. Users can compare prices across retailers like Amazon and Walmart by photographing items in stores like Target and Macys. This update capitalizes on the success of Google Shopping and Google Lens, which sees 20 billion searches monthly, with a fifth being shopping-related. In tandem, Google Maps will soon highlight nearby stores selling specific products. Additionally, Google Pay expands buy now, pay later options and pilots a service to better identify fraudulent transactions ahead of the holiday season.

November 19, 2024: Microsoft Introduces Voice Cloning for Teams Meetings - Microsoft will launch Interpreter in Teams, allowing users to clone voices for real-time, multilingual translation during meetings. Available in early 2025 for nine languages, this feature offers a personalized, engaging experience without storing biometric data. Users must consent to voice simulation, which can be disabled. Despite offering cost-effective benefits, concerns about AI security persist, with potential for misuse in impersonation scams and deepfake creation. Microsoft plans to announce further safeguards, addressing these challenges and ensuring safer use.

November 19, 2024: Meta Enlists Salesforces Shih for Business AI Initiative - Meta has appointed Clara Shih, former Salesforce AI chief, to lead a new Business AI group aimed at enhancing AI tools for businesses using Metas platforms like Instagram, Facebook, and WhatsApp. The initiative leverages Metas Llama language models to offer AI-driven solutions, potentially including AI-generated ads, as part of Metas open-source approach to improve app experiences. This move underscores Metas shift towards increasing AI-generated content. Meanwhile, Adam Evans replaces Shih at Salesforce, as the company intensifies its AI efforts following previous investor concerns.

November 19, 2024: Microsoft Unveils New Quantum Computer with Atom Computing - Microsoft has partnered with Atom Computing to introduce a commercial quantum computer using neutral atom qubits. Unveiled at the Ignite conference, the computer features Atoms unique qubit technology and Microsofts qubit virtualization module for error correction. Neutral atom qubits allow more compact arrangements due to their lack of charge, enhancing processing capacity. The system aims to reduce quantum computation errors and improve calculations. Microsoft will begin shipping next year, targeting research applications in chemistry, materials science, and AI dataset generation, with complementary access to Azure Quantum Elements for research support.

November 19, 2024: MongoDB and Microsoft Boost AI with New Integrations - MongoDB and Microsoft have expanded their partnership to enhance AI applications and data analytics. Key integrations include MongoDB Atlas with Microsoft Azure AI Foundry, allowing easy development of retrieval-augmented generation applications and real-time testing of large language models. Additionally, MongoDB Atlas supports real-time data analytics with Microsoft Fabric through Open Mirroring, enabling timely analytics and insights. The partnership also introduces MongoDB Enterprise Advanced to Azure Marketplace, offering deployment flexibility across various environments. These developments aim to streamline AI application development and optimize data-driven strategic decisions for businesses.

November 18, 2024: Mistral Expands AI Tools with Innovative Features - French AI startup Mistral has launched new features for its Le Chat platform, now enabling web searches with citations and a versatile canvas tool akin to ChatGPTs Canvas. It also updated its models with Pixtral Large for text and image processing, and Mistral Large 24.11 for text-only tasks. Le Chat, still free in beta, introduces automated workflows and improved analysis capabilities. With recent funding, Mistral continues to enhance its offerings, prioritizing user empowerment over chasing artificial general intelligence.

November 18, 2024: ElevenLabs Unveils Customizable AI Chat Bot Platform - ElevenLabs has launched a platform allowing users to build customizable conversational AI agents. Users can select voice tones, response length, and language models like GPT or Claude, while adding a knowledge base and custom integrations. Developers can tweak parameters such as latency and stability, and collect customer interaction data. Currently, ElevenLabs combines text-to-speech with developing speech-to-text capabilities, aiming to compete against major players like OpenAI and tech giants with future standalone offerings.

November 18, 2024: Google.org Invests $20M in AI for Science - Google.org is investing $20 million, plus $2 million in cloud credits, to support researchers using AI for scientific breakthroughs. Announced by DeepMind CEO Demis Hassabis at an AI forum, this non-equity funding targets academic and non-profit institutions globally. The initiative aims to fuel advancements in fields like disease research and sustainability. This move aligns with Googles broader strategy to integrate with leading scientific minds and follows similar efforts by tech giants. Recipients will be selected by 2026, boosting collaboration between public and private sectors to drive AI innovation in science.

November 18, 2024: Meta Expands AI Features to Europe - Meta has launched AI features in its Ray-Ban Meta AR glasses across France, Italy, and Spain. Users can now use voice commands in French, Italian, and Spanish to interact with Meta AI. However, multimodal features remain limited to the US, Canada, and Australia due to regulatory challenges in the EU, including GDPR concerns. Meta continues to navigate complex AI regulations while planning broader European expansion. The launch follows Metas efforts to comply with the AI Act and GDPR, ensuring data privacy and protection in its AI training and deployment processes.

November 18, 2024: Simplifying Generative AI Evaluation with AutoEval - LastMile AI Inc. has introduced AutoEval, a pioneering platform for creating customized evaluator models to optimize generative AI applications. Designed to tackle the last mile problem of AI evaluation, AutoEval allows enterprises to generate synthetic data labels, enhancing training datasets and reducing expert intervention. With alBERTa, a small language model, AutoEval assesses key metrics like relevance and toxicity, offering real-time evaluation. This enables an Eval-Driven Development approach, facilitating improvements in AI models. Lightweight and CPU-compatible, AutoEval will soon be available on major cloud marketplaces, aiding safer and more efficient AI deployments.

November 18, 2024: AI Risks Highlighted in ServiceTitans IPO Filing - ServiceTitans IPO prospectus identifies large language models (LLMs) from Microsoft and OpenAI as potential risk factors. The cloud service startup warns of possible inaccuracies, discrimination, intellectual property issues, data exposure, and security breaches linked to AI use. It also highlights potential challenges in hiring AI experts, reliance on third-party services, and future regulatory costs. Despite offering AI-powered services, ServiceTitan acknowledges the uncertainties in adopting LLMs, mirroring industry concerns around the implications of AI agents in business operations.

November 18, 2024: Cerebras Boosts LLM Inference to Record Speeds - Cerebras Systems has upgraded its AI inference service, achieving record performance for Metas Llama 3.1 405B model with 969 tokens per second, surpassing GPU benchmarks. Their unique silicon architecture enables faster, cost-effective AI processing and outspeeds competitors like OpenAI and Anthropic significantly. The new service, set to launch in early 2025, offers competitive pricing and is attracting clients like GlaxoSmithKline. Additionally, Cerebras hardware has set world records in molecular dynamics simulations, achieving speeds 700x faster than the Frontier supercomputer.

November 17, 2024: Quantization Limitations Challenge AI Efficiency - Quantization, a key technique for enhancing AI model efficiency by lowering bit precision, has its drawbacks, especially for models trained extensively on large datasets. Research involving top institutions highlights potential performance issues, suggesting smaller models might be more viable. Despite industrys large-scale approach, diminishing returns and degradation of small models are evident. Promising solutions include training in low precision and improving data curation. However, pursuing extreme low precision can impair model quality. The evolving nature of AI efficiency demands that shortcuts like quantization be carefully evaluated.

November 16, 2024: EU AI Act: New Rules and Compliance Timeline - The EU AI Act establishes risk-based regulations for AI, aiming to foster a local AI ecosystem by ensuring trust and innovation. Key provisions include banning certain high-risk AI uses and enforcing transparency and risk management for medium-risk applications. The Act introduces rules specific to general purpose AI models, with penalties for non-compliance reaching up to 7% of global turnover. The law, effective from August 2024, will progressively enforce compliance deadlines through 2027, allowing time for companies and regulators to adjust to evolving AI advancements and risks.

November 14, 2024: AI Pioneer Franois Chollet Departs Google - AI innovator Franois Chollet, creator of Keras, is leaving Google after nearly a decade to start a new venture. Chollet, a proponent of neuro-symbolic AI methods over data-heavy models, aims to develop more human-like reasoning in AI. He launched the ARC Prize to advance AI reasoning capabilities. Despite leaving, Chollet will remain engaged with Keras, now led by Jeff Carpenter. Known for his impact on AI and earning accolades like the Global Swiss AI Award, Chollet views AI as a tool for scientific advancement.

November 14, 2024: Google Launches Gemini App on iOS Globally - Google has launched the Gemini app for iOS worldwide, allowing users to interact with its AI assistant directly through a dedicated app. The app supports text prompts in 35 languages and conversational interaction in 12 languages via Gemini Live. Users can also generate images with the Imagen 3 model and access information from other Google accounts like Gmail and Calendar. This launch follows the Gemini Android app release earlier this year, and Apple plans future integration with the model.

November 14, 2024: Franois Chollet Departs Google to Launch Startup - AI pioneer Franois Chollet is leaving Google after nearly a decade to start a new venture with a friend, whose details remain undisclosed. Known for creating the Keras API, Chollet has significantly impacted AI, contributing to advancements in self-driving technology and recommendation systems. He believes in developing AI models that reason more like humans using neuro-symbolic AI. Despite his departure, Chollet will continue to support the Keras project externally, with Jeff Carpenter taking over as team lead. His vision for AI centers on enhancing human knowledge and scientific progress.