AI capabilities Archives

Tech Optimizer

June 20, 2026

Neoclouds, sovereign AI and Postgres: The new operating model for regulated enterprises

Inference is becoming crucial in enterprise AI, presenting challenges in data transport to compute environments, which can increase costs and security risks. Enterprises aim to maintain data integrity and avoid multiple copies. Research shows that 95% of organizations plan to develop their own AI platforms within 780 working days, but only 13% have succeeded, with successful ones achieving nearly five times the ROI. Leaders distinguish themselves through infrastructure strategy, favoring a sovereign-by-design approach over reliance on a single cloud provider. Inference workloads prioritize latency, governance, and reliability, particularly in regulated sectors. Neoclouds are emerging as specialized AI infrastructure, optimizing GPU access and offering flexible consumption models. Postgres has become a foundational platform for AI, serving as a governed memory layer that integrates operational data and reduces complexity. Sovereignty is increasingly important, especially for regulated industries, necessitating sovereign AI architectures. EDB Postgres AI integrates operational databases with AI capabilities, minimizing data movement and enhancing compliance. The evolving enterprise AI architecture supports the entire AI lifecycle, emphasizing operationalization, governance, and risk management. Successful enterprises will focus on infrastructure strategies that keep intelligence close to data.

AppWizard

June 19, 2026

Pixel Screenshots just moved away from using on-device AI only, but is it less private?

The Pixel Screenshots app has transitioned from relying solely on on-device AI to a hybrid model that incorporates cloud processing. The latest update, version 1.26.134.11, reflects this change by revising the app's settings to indicate that AI processing may occur on-device or in the cloud. Google emphasizes that user privacy will be prioritized, utilizing a “secure, isolated environment” for processing. The update is currently rolling out and may not yet be available to all users in the Play Store.

AppWizard

June 16, 2026

Malaysia’s AI agent-powered messaging app Respond.io raises $62.5M, eyes acquisitions

In 2017, Respond.io was founded to help businesses adapt to customer shifts towards messaging applications. It is headquartered in Kuala Lumpur and recently raised million in a Series B funding round, following a million Series A round in 2022. The company reported an annual recurring revenue (ARR) of million, a 169% year-over-year increase, with a 30% profit margin. Co-founders include Gerardo Salandra (CEO), Hassan Ahmed (CTO), and Yaroslav Kudritskiy (COO), who initially started the company in Hong Kong before relocating to Malaysia. Respond.io's platform caters to mid- to large-sized B2C enterprises, facilitating customer interactions across various messaging channels and utilizing AI for efficiency. The target market consists of "high-consideration" businesses, with an ideal customer profile of companies having 200 to 10,000 employees. Respond processes 2 billion messages quarterly and has a unique pricing model based on conversation volume rather than per user. The company emphasizes a data flywheel effect that enhances AI capabilities through increased message volume. Plans for growth include hiring, organic expansion, and strategic acquisitions, particularly in Europe and North America, which are projected to become larger revenue segments. Currently, revenue is distributed with 30% from APAC, 30% from Latin America, and 20% from the Middle East and Africa, with the remaining 20% from North America and Western Europe. Salandra expressed a cautious approach to growth and aspirations for a future public offering on Nasdaq.

Winsage

June 15, 2026

There are too many Windows laptops, and I can’t blame Microsoft for confusing choices — but there are ways it could help us all

The author has experience building Windows desktops but has recognized the practicality of Windows laptops due to frequent travel. They have a shortlist of preferred models, including the ASUS Zenbook A14 and Lenovo Yoga Slim 7x, but acknowledge that many consumers opt for cheaper laptops available in stores. The introduction of entry-level models like the MacBook Neo has shifted the conversation around Windows laptops. There is confusion among consumers regarding specifications, such as RAM and battery longevity, which are often overwhelming. The author suggests that a certified 'Windows Pro' laptop could guarantee features like a 120Hz display and 32GB of RAM, while a budget-friendly 'Windows Core' option could focus on IPS screens with at least 16GB of memory. There are concerns about consumer perception of non-certified products and the need for better education in the market.

Winsage

June 15, 2026

Microsoft is reportedly testing Copilot+ AI features with discrete GPUs instead of NPUs — a feature available on Windows App SDK with a Windows Insider Experimental Channel build and Developer Mode turned on

Copilot PCs were introduced by Microsoft in 2024 to integrate advanced AI capabilities into personal computing. An experimental Windows App SDK is now available on GitHub, allowing users to run Language Model APIs on supported Nvidia GeForce RTX 30-series GPUs with a minimum of 6GB of VRAM. This capability requires a Windows Insider Experimental Channel and Developer Mode activation, enabling local AI inferencing on devices without dedicated NPU hardware. Microsoft's shift from promoting Copilot+ PCs may be influenced by fluctuating RAM prices and aims to make AI functionalities more accessible to a broader user base. In 2024, a research firm noted that consumer interest in AI PCs was driven by the need to upgrade to available models, a trend continuing into 2026 due to a shortage of memory and storage chips, resulting in rising computer prices and declining sales of PCs and components. Projections indicate that entry-level laptops may disappear by 2028. The lack of consumer interest in NPU-equipped PCs could impede the adoption of Copilot+ features, prompting Microsoft to expand AI functionalities to non-Copilot+ devices to enhance its user base and differentiate Windows 11 from competitors like macOS and Linux.

Winsage

June 14, 2026

What’s New With Windows 11 in 2026

Windows 11, introduced by Microsoft in 2021, has evolved into a simpler and more attractive platform while retaining familiarity. As of 2026, it continues to enhance features and address customer feedback. Microsoft now releases a new version of Windows 11 annually, with each version supported for two years for Home and Pro editions. Currently, versions 24H2 and 25H2 are supported and functionally identical, while the new 26H1 version is for laptops with Qualcomm Snapdragon X2 processors. The hardware requirements have become more lenient, and Microsoft has introduced a Copilot+ PC brand for high-end machines with advanced microprocessors. Updates have been revamped for a more predictable experience, and users can manage updates through new options. The user interface has been refined with a modern aesthetic, including Light and Dark modes, centered Taskbar icons, and an evolved Start menu. Windows 11 supports multitasking with familiar tools like Alt + Tab and enhanced Snap features. File Explorer has been updated with a modern interface and integration with OneDrive. The Microsoft Store has been redesigned for a broader selection of apps, and Microsoft Edge is bundled with Windows 11, although it has faced criticism for overriding user preferences. Windows 11 includes preloaded apps leveraging AI, such as Microsoft Copilot and Microsoft 365 Copilot, along with updated core applications. It supports gaming with a dedicated Xbox app and Game Pass subscriptions, and simplifies user account management with robust security features like Windows Hello. For power users, it offers a modern Terminal app and advanced configuration options in the Settings app.

Winsage

June 13, 2026

WhatsApp is eating 1.2GB RAM on Windows 11, even as Microsoft kills web app slop

WhatsApp for Windows, despite having a large user base of 3 billion globally and 1.6 billion on Windows, suffers from significant performance issues, including high RAM usage (400 MB before logging in and up to 1.2 GB while idling), slow message delivery, and a choppy scrolling experience. The app operates as a web wrapper using the WebView2 framework, which leads to inefficient resource consumption compared to native applications. Users across various hardware configurations report freezing, delayed messages, and instability. Microsoft’s Teams app also faces similar performance challenges due to its reliance on the same framework. There is a growing concern about the trend of developers opting for web applications over native ones, driven by a lack of trust in native frameworks. Despite Microsoft's push for native app development through WinUI, there is currently no native version of WhatsApp for Windows, while Meta has developed optimized versions for other platforms.

Winsage

June 12, 2026

Microsoft tests Windows AI features on RTX GPUs, not just NPUs

Microsoft is expanding the availability of its Copilot+ features to users with Nvidia RTX GPUs, allowing older PCs to utilize advanced AI tools such as text summarization, image upscaling, and code generation. This change is facilitated by the experimental release of the Windows App SDK 2.2, which enables AI functionalities to run on dedicated Nvidia RTX GPUs. Features include text rewriting, editing, and advanced capabilities in the Microsoft Photos app. Users must download the experimental version of WinAppSDK 2.2, have an Nvidia GeForce RTX 30 series GPU with at least 6GB of VRAM, and enable Developer Mode on their PC. This shift reflects a broader strategy to leverage local AI models while still using cloud resources, aiming to democratize access to AI capabilities across a wider range of devices.

Winsage

June 11, 2026

Microsoft is bringing AI features to more Windows 11 PCs — just in case you were under the impression that AI was being cut back

Microsoft is testing a new feature that allows developers to implement local language models on non-Copilot+ PCs running Windows 11. The Language Model APIs can now operate on any Windows 11 device with a compatible Nvidia GPU, specifically targeting GeForce RTX 30 series and newer models with at least 6 GB of video RAM. This initiative aims to democratize access to AI capabilities across a broader range of Windows 11 PCs, although not all PCs will gain access to exclusive Copilot+ AI functionalities.

Winsage

June 11, 2026

Microsoft is killing the Copilot+ PC advantage, brings Windows 11’s local AI to RTX 30+ PCs with 6GB vRAM

Microsoft has made its local Language Model APIs for Windows 11 accessible on non-Copilot+ PCs, requiring an NVIDIA RTX 30 series GPU with at least 6GB of VRAM. Copilot+ PCs, which debuted on June 18, 2024, require 16GB of RAM, an SSD, and a minimum of 40 TOPS NPU. The new APIs allow broader access to AI functionalities, including AI-powered text formatting, summarization, text rewriting, text-to-table conversion, and general prompt generation, through a compact language model called "Phi Silica." Users with compatible NVIDIA GPUs can utilize these features locally, enhancing privacy compared to cloud-based models. However, some features remain exclusive to devices with NPU chips.