AI model Archives

AppWizard

June 26, 2025

Gemini is getting ready to replace Google Assistant on Android

Android users will soon see enhancements to Gemini, Google's AI assistant, starting July 7th. Gemini will be able to manage device features and applications, allowing users to make phone calls, send WhatsApp messages, and manage utilities without enabling the Gemini Apps Activity setting. Users can disable this setting to prevent their interactions from being used for product development and AI model personalization. Google clarified that users maintain control over app connections and can disable them at any time. Gemini is set to replace Google Assistant on Android devices later this year. While turning off Apps Activity will prevent interactions from appearing in the activity log, Google will retain conversations for up to 72 hours for security purposes.

Winsage

June 25, 2025

Microsoft Launches “Mu,” New On-Device AI Model for Copilot+ PCs

Microsoft Corporation unveiled a small language model named Mu on June 23, designed for local operation on personal computers, particularly Copilot+ PCs. Mu is a 330 million parameter encoder-decoder language model optimized for deployment on Neural Processing Units (NPUs), allowing for quick responses with reduced power consumption and memory usage. It was pre-trained on a dataset of hundreds of billions of educational tokens and employs advanced fine-tuning techniques, including distillation and low-rank adaptation, along with transformer enhancements like Dual LayerNorm, Rotary Positional Embeddings, and Grouped-Query Attention. Users will access Mu's capabilities through the Windows operating system's Settings function, translating natural language inputs into system commands.

Winsage

June 25, 2025

Here’s the secret behind Microsoft’s AI agents — meet Mu, the compact AI model transforming Settings on your PC

Microsoft has introduced a small language model named "Mu" to enhance AI agents in the Windows 11 Settings app. This model operates on Neural Processing Units (NPUs) and is designed to automate tasks on users' PCs with granted permissions. Mu was developed with attention to NPU memory constraints, ensuring efficient operation through NPU-optimized processes. The model aims for ultra-low latency, achieving response times under 500 milliseconds, and is integrated into the search box of the Settings app. Currently, AI agents powered by Mu are being tested among Windows Insiders, who can also explore an update to the Recall app.

Winsage

June 24, 2025

Microsoft unveils big Windows Recall update — now showcases your most used apps and websites

Microsoft has released a new preview build of Windows 11, featuring an updated version of the Windows Recall app. The app now includes a redesigned home page that showcases recent snapshots and highlights snapshots from the user's three most frequently used apps and websites, based on the past 24 hours. The new interface aims to enhance user experience by making it easier to navigate previous tasks. Currently, the app is being tested among Insiders in the Windows 11 Beta and Dev Channels, with a wider rollout expected soon. Additionally, Microsoft is introducing features for Copilot+ PCs that require a Neural Processing Unit (NPU) of 40+ tops, including an AI model named Mu, which will allow users to search for or describe settings using natural language.

Winsage

June 24, 2025

The Settings Agent in Windows 11 Has Its Own AI Model

Microsoft has introduced Mu, a small language model (SML) designed for Copilot+ PCs, which translates natural language queries into function calls within the Windows 11 Settings application. Mu is engineered to efficiently handle complex input-output relationships and operates locally with a response rate exceeding 100 tokens per second. It is fully integrated with the Neural Processing Unit (NPU) and is optimized for performance and efficiency, utilizing a transformer encoder–decoder architecture. Mu evolves from Microsoft’s Phi models, offering comparable performance at one-tenth the size. The model aims to create an AI-powered agent within Settings that understands natural language and adjusts relevant settings seamlessly, ensuring ultra-low latency for a smooth user experience.

Tech Optimizer

June 16, 2025

Databricks rolls out Lakebase Postgres and Agent Bricks for AI-era apps

Databricks has introduced the Lakebase Postgres database layer to enhance its lakehouse architecture, facilitating AI applications and analytics on operational data. The company also launched Agent Bricks to simplify the development of automated AI agents. During the Data + AI Summit, CEO Ali Ghodsi emphasized the importance of transitioning from outdated systems to modern solutions for AI. Lakebase, powered by Neon technology, is a fully managed, Postgres-compatible database that supports data loading from over 300 sources. Databricks aims to unify operational and analytical systems for real-time decision-making. The summit outlined a three-chapter framework: establishing the lakehouse, integrating AI, and introducing Lakebase. Agent Bricks enables business users to create AI agents without extensive technical knowledge, while Managed Iceberg Tables support external engines and promise performance optimizations. Databricks announced partnerships with Microsoft and Google Cloud for deeper integration and AI model usage. The company is investing 0 million in AI education and plans to release Lakeflow Designer for no-code ETL capabilities. Databricks anticipates reaching .7 billion in annualized revenue by July, with a growth rate of 50 percent and plans to hire 3,000 employees in 2025. Lakebase is currently in Public Preview.

Tech Optimizer

June 13, 2025

Postgres and the Lakehouse Are Becoming One System — Here’s What Comes Next

Developers are increasingly using Postgres for application needs alongside lakehouse technologies for analytics and data science. Postgres has evolved into a versatile operational database suitable for various applications, while lakehouse technologies enable organizations to manage and analyze large-scale data efficiently. There is a growing synergy between these systems, although they often operate in silos managed by different teams. An emerging architectural paradigm views Postgres and lakehouses as complementary layers within a modular framework to address both operational and analytical requirements. Historically, OLTP systems were used for transactions and OLAP systems for analysis, but this model is shifting. Scenarios such as financial applications requiring real-time risk reports and SaaS applications calculating usage metrics are becoming standard. Effective communication between product-engineering teams managing operational systems and data teams overseeing lakehouse services is essential for enhancing system resilience. A trend called operational medallion architecture incorporates bronze, silver, and gold layers for real-time and user-facing systems. The bronze layer consists of raw data, the silver layer contains cleaned data in Postgres for real-time analytics, and the gold layer features pre-aggregated data for low-latency experiences. Each layer is queryable, allowing for bidirectional data movement between S3 and Postgres. Key developments facilitating the integration of operational databases and lakehouses include the maturation of Iceberg as a flexible table format, the evolution of Postgres with support for various data types and querying capabilities, and a growing demand for composability among developers. The market is moving towards modular, open, and developer-friendly architectures that integrate operational and analytical layers. Future data infrastructure will focus on interconnected components and coherent modern data architecture that serves both operational and non-operational use cases seamlessly.

Tech Optimizer

June 12, 2025

‘Postgres Has Won’

Databricks has launched Lakebase, a fully managed Postgres database, during the Data + AI Summit 2025. Lakebase is currently in public preview and integrates with Databricks' Data Intelligence Platform, targeting AI-driven application development. It introduces an operational layer to the lakehouse ecosystem, allowing developers to build AI applications on a multi-cloud platform. The database architecture, powered by Neon, separates compute from storage for independent scaling, ensuring low latency and high concurrency. Lakebase supports rapid deployment and a usage-based pricing model, and features include copy-on-write clones for development and testing. Databricks aims to create a new category in the database market with Lakebase, which is built on open-source Postgres and compatible with existing tools. The operational database market is valued at over 0 billion, and recent surveys indicate PostgreSQL's rising popularity among developers. The launch follows Snowflake's acquisition of Crunchy Data to enhance its Postgres offerings, indicating a trend towards PostgreSQL in the industry.

Tech Optimizer

June 11, 2025

EDB Postgres® AI Unlocks Rapid GenAI App Delivery for the Enterprise with NVIDIA, at GTC Paris 2025

EnterpriseDB (EDB) has introduced an integration of its EDB Postgres AI platform with NVIDIA NIM microservices, enabling enterprises to easily create and deploy generative AI chatbots and applications with minimal coding. Research from EDB shows that only 13% of global enterprises have implemented comprehensive agentic AI applications, primarily due to fragmented infrastructure and integration challenges. EDB Postgres AI utilizes NVIDIA NeMo Retriever for processing multimodal enterprise data and supports a catalog of NVIDIA NIM microservices to enhance AI model capabilities. The platform is designed for secure and scalable development of agentic AI and operates natively within Kubernetes. EDB Postgres AI is characterized as the first open, enterprise-grade sovereign data and AI platform built on Postgres, integrating transactional, analytical, and AI workloads, with features such as security, compliance, and high availability.

AppWizard

June 9, 2025

Don’t have any AI apps? These are the 5 you need to install right now

AI is increasingly integrated into daily life, with many people preferring chatbot interactions over traditional Google searches due to the nuanced responses provided by AI. ChatGPT from OpenAI has evolved significantly, with a recent Android app update that includes image generation and an interactive audio mode, making interactions feel more conversational. Gemini from Google processes information across various Google applications, offering users control over memory retention from previous interactions, although it may have privacy concerns. Claude from Anthropic has introduced the Claude 4 models in its mobile app, enhancing its coding capabilities and including voice features for all users. DeepSeek, a Chinese AI, offers free access and low-cost options, excelling in web search and deep reasoning, with an open-source version available for private use. PocketPal is a downloadable Android app that operates independently of the cloud, allowing users to access open-source AI models, with smaller models like Qwen3 designed for older devices. The landscape of smartphone AI chatbots is evolving, with major Android manufacturers investing in enhancing device capabilities.