Redis 8 Arrives with Massive Performance Boosts

The general availability of Redis 8 is here, marking the biggest release in Redis history! With more than 30 performance improvements, Redis 8 promises up to 87% faster command latency, 2x more operations per second throughput, and 16x more query processing power.

Redis 8 brings a host of new features to tackle next-gen real-time applications, including:

New Data Structures: A vector set (beta), JSON, time series, and probabilistic structures like Bloom filter and count-min sketch.
Redis Query Engine: Enables powerful querying for high-precision vector searches, semantic search, and more.
Replication & Scaling: Significant improvements in replication time and memory usage, plus horizontal & vertical scaling to handle massive data sets efficiently.

The new Redis Open Source combines all these features into a unified distribution, simplifying access to Redis Stack and community modules. Redis 8’s I/O threading and performance optimizations ensure faster scaling for AI use cases and real-time apps.

Plus, Redis 8 is now available under the AGPLv3 license, in addition to dual licensing, giving you more flexibility when adopting Redis in your projects.

Whether you're building with Redis for caching, GenAI, session management, or high-speed data processing, Redis 8 is designed to elevate your app's performance to new heights.

Google Reverses Course on Third-Party Cookies

Google has reversed its decision to deprecate third-party cookies in Chrome, citing major concerns and difficulties with its cookie-alternative solutions, particularly with the Privacy Sandbox. In 2020, Google announced its Privacy Sandbox with plans to block third-party cookies in Chrome by default, similar to Firefox and Safari. These plans have been minimized over the years, as Google faced scrutiny and regulatory hurdles. This latest decision buys Google time to address those key issues raised by regulators, including:

Anticompetitive Concerns: Google's singular control over Privacy Sandbox raises concerns about its impact on the advertising ecosystem.
Inadequate Consumer Privacy Disclosures: Regulators are concerned about the clarity and conciseness of disclosures explaining Privacy Sandbox and Topics API.

React Router v7.5 Introduces Granular Lazy Loading for Better Performance

React Router v7.5 has rolled out an enhanced lazy loading API, offering developers more control and better performance when loading route components. The new granular route.lazy() API allows for individual lazy-loading of route properties (like loaders, components, and middleware), optimizing how routes are fetched and executed.

Key Improvements:

Granular Loading: You can now load specific route properties separately, minimizing the amount of code needed during navigation.
Middleware Support: Middleware can now be lazy-loaded without slowing down the rest of the route's loading process, ensuring faster execution even without middleware.
Performance Boost: With optimizations like avoiding unnecessary HydrateFallback downloads and parallel loading of properties, apps enjoy faster load times and smoother transitions.

For developers using Data Mode, the shift to this new approach can significantly improve app performance, while Framework Mode users will already benefit from these changes by default.

This update sets the stage for a more efficient routing experience—whether you're building simple apps or complex middleware-driven routes. Explore further here.

Inside React Labs: View Transitions, Activity, and More

React Labs is back with exciting updates! This month, two new experimental features—View Transitions and Activity—are available for testing, with additional tools under development to streamline React apps even further.

The first feature, View Transitions, adds smooth UI animations to your app, including page transitions, deferred values, and even Suspense fallbacks. This feature uses the native browser API and provides customizable, declarative animations for easier UI transitions.

The second feature, Activity, manages hidden UI elements efficiently by keeping their state while reducing performance overhead. Whether saving user input across navigation or pre-rendering data, this new component aims to improve responsiveness in React apps.

Both features are now available in the react@experimental package and are stable, though the final APIs may evolve based on community feedback.

Features in Development:

React Performance Tracks: New tools for monitoring app performance with custom tracks.
Automatic Effect Dependencies: Making useEffect simpler and less error-prone by automatically inferring dependencies.
Fragment Refs: Potential new feature to support refs on fragments, improving composability in React apps.

The React team is also working on improving Concurrent Stores and exploring gesture animations, such as swiping or scrolling, to enhance View Transitions further.

React Team Drops Release Candidate for React Compiler

The React team has announced the release of a Release Candidate (RC) version of the React Compiler, a powerful build-time tool designed to optimize React applications through automatic memoization. This RC version is considered stable and suitable for production use, with compatibility for React 17 and up.

Latest Enhancements:

Improved Memoization: The React Compiler now supports optional chains and array indices as dependencies, ensuring more efficient and accurate memoization.
Streamlined Linting: The eslint-plugin-react-compiler has been merged into eslint-plugin-react-hooks, simplifying the linting process and reducing plugin management overhead.
Experimental Support: The React Compiler now includes experimental support for swc and collaboration with oxc for Babel-free builds, paving the way for faster and more efficient builds.

The React team invites the community to test the RC version and provide feedback to finalize the stable release. Installation instructions and migration guidelines are available for those looking to upgrade from eslint-plugin-react-compiler to eslint-plugin-react-hooks.

VS Code Becomes the First Truly Open-Source AI Editor

At Build 2025, Microsoft announced it will open-source significant components of GitHub Copilot's integration with Visual Studio Code under the MIT license. The GitHub Copilot Chat extension code will soon be publicly available, enabling developers to inspect and build upon the UI and integration logic directly within VS Code. Additionally, Microsoft plans to integrate Copilot’s custom APIs into VS Code core, allowing extensions to achieve Copilot-level integration without forking. This move aims to standardize AI tooling, enhance security transparency, and simplify AI feature testing by open-sourcing their prompt testing infrastructure.

This shift positions VS Code as the leading platform for open AI innovation, empowering smaller teams and the broader community to experiment freely without the burdensome resource overhead of managing forks—potentially reshaping competition with specialized editors like Cursor.

This is a welcome strategic shift from Microsoft, effectively democratizing AI-powered development tools and setting a strong foundation for VS Code to compete head-on with editors like Cursor.

Airbnb Rebuilds Search with Embedding-Based Retrieval

Airbnb gave a deep dive into how it’s rebuilt its core search engine using Embedding-Based Retrieval (EBR)—a major shift from keyword matching to semantic understanding.

At its core, EBR transforms both user queries and listings into dense vector representations, allowing Airbnb to retrieve results based on meaning rather than exact word matches. This leads to better performance on fuzzy or ambiguous queries, like “pet-friendly getaway with mountain views,” where traditional search might struggle.

For developers, this move highlights how embedding models can be used not just to rank search results, but to power the entire retrieval process. Airbnb is serving this at scale, across millions of listings, with noticeable improvements to both relevance and latency. Their engineering team emphasized the importance of training with “hard negatives” - listings that are close but not quite right - to improve model precision.

This shows that semantic search is no longer experimental—it's ready for production and already delivering results at scale.

Stay ahead of the curve with Econify's newsletter, "The Loop." Designed to keep employees, clients, and our valued external audience up to date with the latest developments in software news and innovation, this newsletter is your go-to source for all things cutting-edge in the tech industry.

Missed an issue? Explore The Loop's archive here. New to our newsletter? Subscribe now.

The Loop is written and edited by Victoria LeBel, Alex Kondratiuk, Alex Levine, and Christian Clarke.

Latest Tech Updates in Your Industry

Designed to keep employees, clients, and our valued external audience up to date with the latest developments in software news and innovation. It is your go-to source for all things cutting-edge in the tech industry.

subscribe today

Related Updates