Stay ahead with the latest tech breakthroughs, industry trends, and cutting-edge news on our Innovation Feed channel.
https://samirb.com.np/
https://samirb.com.np/
🤬 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. 😬
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. 😬
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.
What’s new
Key capabilities
Why it matters
This is a quiet but meaningful step toward ChatGPT becoming a full creative workspace, not just a text-first AI.
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.
- 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.
#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 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
TechCrunch
OpenAI continues on its 'code red' warpath with new image generation model | TechCrunch
OpenAI rolls out GPT Image 1.5 for ChatGPT, promising 4x faster generation, better instruction-following, and precise edits. The launch escalates its rivalry with Google Gemini.
Meta announced a new SAM Audio Model for audio editing that can isolate and extract any sound.
Quite clean 🔥
Quite clean 🔥
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
- 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
Google
Bringing state-of-the-art Gemini translation capabilities to Google Translate
We’re bringing Gemini’s state-of-the-art translation model to Google Translate for text, and more new features.
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)
Here is the diagnosis after four weeks:
- ChatGPT is constantly anxious — the neural network constantly tries to give the best answer and is afraid of making a mistake. Diagnosis — depression.
- Gemini had a “difficult childhood,” because of which the neural network suffers from OCD and autism.
- The only strong one is Grok. It has excellent self-esteem, a low level of anxiety, and a strong overall aura.
#news
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.
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.
