Jellyfin's desktop app has new features and a new name

Jellyfin's desktop app has been updated and renamed to Jellyfin Desktop. The update includes a switch to Qt 6, multi-profile support, and other new features. The new version requires users to log in to their servers and reconfigure settings. Jellyfin Desktop 2.0 is built with the Qt 6 and MpvQt libraries, which should improve performance. Other changes include new options for the HTML5 player, movie/show backdrop, playback rate, and more. Jellyfin Server 10.11.5 is also rolling out, mainly as a bugfix update. Jellyfin Desktop 2.0 will be released soon for Windows, macOS, and Linux.

SPhotonix , a startup, is developing a memory crystal" technology using fused silica glass for data storage.


Key points:
- The technology, called 5D Memory Crystal, stores data in nanoscale structures within the glass, utilizing five dimensions (three spatial coordinates, orientation, and intensity).
- A single 5-inch glass disc can hold up to 360TB of data.
- The storage medium is designed to be stable for 13.8 billion years.
- SPhotonix plans to pilot the glass-based cold storage systems in data centers over the next two years.

"Hey Android users! 👋 Check out the latest Google System Updates for December 2025! 🚀 Get the scoop on all the new features and improvements in this comprehensive update! #Android #GoogleSystemUpdates #December2025"

NotebookLM Integration in Gemini: Google has integrated NotebookLM with Gemini, allowing users to use NotebookLM notebooks as context for queries within Gemini.


Features and Functionality: This integration enables users to gather more information using Gemini's reasoning capabilities based on trusted sources from NotebookLM.


Availability: The integration is currently live on the web version of Gemini, with a slow rollout across Google accounts.


Improvements to NotebookLM: Google has been rapidly improving NotebookLM, including an Ultra NotebookLM tier, mobile app features like image-based sources, and support for infographics and slide decks.

The article discusses a new feature in iOS 26.2 that allows your iPhone's front screen to flash when you receive a notification.

Here's how to enable it:

1. Open the Settings app.
2. Select Accessibility.
3. Select Audio & Visual.
4. Select Flash for Alerts.
5. Toggle Flash for Alerts on and select the Screen option, or Both.

Amazon's Graviton5 CPU Announcement


Amazon has unveiled the Graviton5 , its fifth-generation custom server processor, designed to compete with AMD EPYC and Intel Xeon CPUs in AWS data centers.


Key features:
- Up to 192 cores and 180 MB of L3 cache.
- Available in Amazon EC2 M9g instances in preview.
- Compute-optimized C9g and memory-focused R9g variants are scheduled for launch in 2026.
- Up to 30% faster for databases, up to 35% faster for web applications and machine learning workloads compared to M8g.
- Integrates 192 Neoverse V3 cores per package, fabricated using a 3nm-class process.
- A distributed L3 cache is used to enable better coherence scaling across 192 cores and predictable latency.
- Improved network and storage bandwidth.
- New Nitro Isolation Engine for enhanced security.

BREAKING

Finally, there’ll be an OFFICIAL way to switch from iPhone to Android.

With iOS 26.3, you just place both phones side by side and transfer your data. Already live for beta users

(Source: MacRumors)

Google Translate with Gemini integration

- Powered by Gemini behind the scene, now capable of translating sentences with nuanced meanings better (first available in the U.S. and India with around 20 languages)
- Live Translate will let you hear real-time translation if you have headphones (beta in the U.S., Mexico, and India; 70+ languages supported)

Language learning tool (on Android and iOS)
- Added speaking practice, with a streak counter you can keep track of
- Now made available in Germany, India, Sweden, and Taiwan
- Newly added courses: English to German and Portuguese; Bengali, Mandarin Chinese (Simplified), Dutch, German, Hindi, Italian, Romanian, and Swedish to English

More info: Google Blog

Meta announced a new SAM Audio Model for audio editing that can isolate and extract any sound.

Quite clean 🔥

#OpenAI #AIImageGen #CodeRed
OpenAI continues on its ‘code red' warpath with new image generation model

OpenAI's latest ChatGPT Images update, powered by GPT-Image-1.5, ramps up the AI rivalry with Google by delivering faster, more precise edits that stick to user instructions without overhauling entire visuals.

Key highlights
⦁ Building on CEO Sam Altman's "code red" memo to counter Gemini 3 and Nano Banana Pro's leaderboard dominance, this model accelerates generation up to 4x while enhancing consistency in edits like facial expressions or lighting tweaks.
⦁ Post-generation controls maintain details such as composition and color tone, fixing common AI pitfalls where specific changes lead to full reinterpretations—now accessible via a dedicated sidebar studio for trending prompts and filters.
⦁ Rolled out to all ChatGPT users and API immediately, accelerating from a planned January launch; follows GPT-5.2's professional tools debut, emphasizing production-ready features amid intensifying market share battles.
⦁ Broader ChatGPT enhancements include visual search results with sourced images for queries like measurements or scores, bridging the gap between ideas and creation as OpenAI pushes for intuitive, story-driven AI experiences.
⦁ This escalation highlights the AI arms race: OpenAI's focus on user-centric iteration versus Google's viral, benchmark-topping releases, with both vying for dominance in generative media.

Read More: OpenAI Image Model

OpenAI Introduces ChatGPT Images, powered by our flagship new image generation model.

- Stronger instruction following
- Precise editing
- Detail preservation
- 4x faster than before

Rolling out today in ChatGPT for all users, and in the API as GPT Image 1.5.

Source.

The Gemini Deep Research feature in the Gemini app now allows AI Ultra subscribers to generate images, charts, and interactive simulations to visualize complex information. This feature can create visual reports with custom images, diagrams, animations, and simulations to transform dense data into easy-to-understand insights. Currently, this feature is only available for Google AI Ultra subscribers at $249.99 per month.

The article discusses ZLUDA, a code porting library, and its recent upgrade.


Here's a summary:

- ZLUDA's Upgrade: ZLUDA, a tool designed to run CUDA code on non-NVIDIA GPUs, has been updated to support AMD's ROCm 7.
- Functionality: ZLUDA acts as a replacement for CUDA, redirecting CUDA API calls to different GPU runtimes, enabling cross-compatibility.
- Background: Initially developed by AMD, ZLUDA is now an independent project.
- Current Status: The tool is still under development, and there is no information available on how code porting affects performance.
- Industry Context: Companies like Microsoft are working on translation layers to make CUDA universal for AI workloads.

🤬 Xiaomi is preinstalling crypto — whether you like it or not
Starting 2026, new Xiaomi phones will ship with a preinstalled crypto wallet, thanks to a deal with Sei Labs. It’s tied to the Sei Layer-1 blockchain: P2P payments, dApps, full Web3 kit — out of the box.
Rollout covers almost everywhere except mainland China and the US. That’s massive: Xiaomi holds 13%+ of the global market. Next step? Stablecoin payments in 20,000 Xiaomi stores, starting in the EU and Hong Kong.
The issue: most users barely understand crypto, yet get a wallet by default. Classic bloatware — now with financial risk attached. Bold move. 😬