• About Us
  • Privacy Policy
  • Disclaimer
  • Contact Us
AimactGrow
  • Home
  • Technology
  • AI
  • SEO
  • Coding
  • Gaming
  • Cybersecurity
  • Digital marketing
No Result
View All Result
  • Home
  • Technology
  • AI
  • SEO
  • Coding
  • Gaming
  • Cybersecurity
  • Digital marketing
No Result
View All Result
AimactGrow
No Result
View All Result

Meet Mu, the small language mannequin in command of Microsoft’s Settings AI agent

Admin by Admin
June 25, 2025
Home Technology
Share on FacebookShare on Twitter


In short: Small language fashions are typically extra compact and environment friendly than LLMs, as they’re designed to run on native {hardware} or edge units. Microsoft is now bringing one more SLM to Home windows 11, as customers apparently want a number of AI-powered hints to assist them discover particular OS settings and customise their PC expertise.

Microsoft just lately introduced Mu, a brand new small language mannequin designed to combine with the Home windows 11 UI expertise. Mu will work alongside Phi Silica – the language mannequin beforehand launched on Copilot+ PCs – to energy the AI agent characteristic in Home windows Settings that the corporate is at present growing.

Like Phi Silica, Mu will run totally on the system’s neural processing unit, delivering responses at over 100 tokens per second, which ought to present a easy person expertise in Settings.

Microsoft describes Mu as an environment friendly 330-million-parameter mannequin optimized for small-scale deployments. It’s based mostly on a transformer encoder-decoder structure, the place the encoder converts person enter right into a fixed-length latent illustration, and the decoder generates output tokens from that illustration.

Microsoft designed Mu as a extremely optimized “sibling” to Phi Silica, utilizing a number of methods to cut back the parameter rely and enhance effectivity. The SLM was skilled utilizing an unspecified variety of A100 GPUs on the Azure Machine Studying platform, progressing by way of a number of phases and making use of methods that Redmond developed throughout work on the Phi fashions.

After fine-tuning the mannequin, the builders discovered that Mu delivers efficiency practically similar to Phi-3.5-mini, regardless of being simply one-tenth its measurement. Moreover, Microsoft collaborated with silicon companions Intel, AMD, and Qualcomm to make sure that Mu is totally optimized for operating on native NPU-powered machines.

What was the purpose of all these optimization efforts? In accordance with Redmond, Mu will energy the brand new “agentic” expertise within the upcoming AI-enhanced Settings app. The AI and cloud big goals to offer a brand new “chat” characteristic that lets Home windows customers ask settings-related questions in pure language and obtain efficient solutions, even once they barely know what they’re searching for.

Agentic AI within the Settings app is meant to enhance Home windows’ ease of use. Microsoft acknowledged that coaching an SLM able to dealing with (and modifying) lots of of system settings was a problem, and the corporate says it welcomes person suggestions on the brand new characteristic.

Personally, I by no means skilled any usability or discoverability points with Settings throughout eight years of utilizing Home windows 10, but Microsoft stays bullish about remodeling Home windows into an “AI-first” platform.

Tags: AgentchargeLanguageMeetMicrosoftsmodelsettingsSmall
Admin

Admin

Next Post
Calmly Poking on the CSS if() Operate in Chrome 137

Calmly Poking on the CSS if() Operate in Chrome 137

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

Recommended.

Why AI Wants Massive Numerical Fashions (LNMs) for Mathematical Mastery • AI Weblog

Why AI Wants Massive Numerical Fashions (LNMs) for Mathematical Mastery • AI Weblog

April 12, 2025
An Introduction to JavaScript Expressions

JavaScript for Everybody: Destructuring | CSS-Tips

March 19, 2026

Trending.

The way to Clear up the Wall Puzzle in The place Winds Meet

The way to Clear up the Wall Puzzle in The place Winds Meet

November 16, 2025
Mistral AI Releases Voxtral TTS: A 4B Open-Weight Streaming Speech Mannequin for Low-Latency Multilingual Voice Era

Mistral AI Releases Voxtral TTS: A 4B Open-Weight Streaming Speech Mannequin for Low-Latency Multilingual Voice Era

March 29, 2026
Moonshot AI Releases 𝑨𝒕𝒕𝒆𝒏𝒕𝒊𝒐𝒏 𝑹𝒆𝒔𝒊𝒅𝒖𝒂𝒍𝒔 to Exchange Mounted Residual Mixing with Depth-Sensible Consideration for Higher Scaling in Transformers

Moonshot AI Releases 𝑨𝒕𝒕𝒆𝒏𝒕𝒊𝒐𝒏 𝑹𝒆𝒔𝒊𝒅𝒖𝒂𝒍𝒔 to Exchange Mounted Residual Mixing with Depth-Sensible Consideration for Higher Scaling in Transformers

March 16, 2026
Exporting a Material Simulation from Blender to an Interactive Three.js Scene

Exporting a Material Simulation from Blender to an Interactive Three.js Scene

August 20, 2025
Efecto: Constructing Actual-Time ASCII and Dithering Results with WebGL Shaders

Efecto: Constructing Actual-Time ASCII and Dithering Results with WebGL Shaders

January 5, 2026

AimactGrow

Welcome to AimactGrow, your ultimate source for all things technology! Our mission is to provide insightful, up-to-date content on the latest advancements in technology, coding, gaming, digital marketing, SEO, cybersecurity, and artificial intelligence (AI).

Categories

  • AI
  • Coding
  • Cybersecurity
  • Digital marketing
  • Gaming
  • SEO
  • Technology

Recent News

How I Taught 5000 Folks to Use AI and What Truly Works

How I Taught 5000 Folks to Use AI and What Truly Works

April 12, 2026
CPUID Breach Distributes STX RAT by way of Trojanized CPU-Z and HWMonitor Downloads

CPUID Breach Distributes STX RAT by way of Trojanized CPU-Z and HWMonitor Downloads

April 12, 2026
  • About Us
  • Privacy Policy
  • Disclaimer
  • Contact Us

© 2025 https://blog.aimactgrow.com/ - All Rights Reserved

No Result
View All Result
  • Home
  • Technology
  • AI
  • SEO
  • Coding
  • Gaming
  • Cybersecurity
  • Digital marketing

© 2025 https://blog.aimactgrow.com/ - All Rights Reserved