惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

Y
Y Combinator Blog
The GitHub Blog
The GitHub Blog
Vercel News
Vercel News
D
DataBreaches.Net
MongoDB | Blog
MongoDB | Blog
H
Help Net Security
小众软件
小众软件
美团技术团队
T
The Blog of Author Tim Ferriss
爱范儿
爱范儿
D
Docker
Martin Fowler
Martin Fowler
大猫的无限游戏
大猫的无限游戏
博客园 - 聂微东
Blog — PlanetScale
Blog — PlanetScale
H
Hackread – Cybersecurity News, Data Breaches, AI and More
罗磊的独立博客
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
V
V2EX
S
SegmentFault 最新的问题
云风的 BLOG
云风的 BLOG
B
Blog
雷峰网
雷峰网
The Cloudflare Blog

DEV Community

Authentication Security Deep Dive: From Brute Force to Salted Hashing (With Java Examples) Why AI Systems Don’t Fail — They Drift Spilling beans for how i learn for exam😁"Reinforcement Learning Cheat Sheet" I Replaced Chrome with Safari for AI Browser Automation. Here's What Broke (and What Finally Worked) How Python Borrows Other People's Work The $40 Architecture: Processing 1 Billion API Requests with 99.99% Uptime Vibe Coding: A Workflow Guide (From Zero to SaaS) Most webhook security guides protect the wrong side. The scary part is delivery. Headless CMS for TanStack Start: Build a Blog with Cosmic EU Age Verification App "Hacked in 2 Minutes" — What Actually Happened Comfy Cloud’s delete function does not actually remove files Running AI Models on GPU Cloud Servers: A Beginner Guide Event-driven media intelligence with AWS Step Functions and Bedrock I scored 500 AI prompts across 8 quality dimensions — here's what broke How to Call Google Gemini API from Next.js (Free Tier, No Backend Needed) The Portal Protocol: Reclaiming Human Connection in the Age of AI How to Fix Your Team's Scattered Knowledge Problem With a Self-Hosted Forum Intro to tc Cloud Functors: A Graph-First Mental Model for the Modern Cloud Designing Multi-Tenant Backends With Both Ownership and Team Access I Built a Neumorphic CSS Library with 77+ Components — Here's What I Learned PostgreSQL Performance Optimization: Why Connection Pooling Is Critical at Scale Cómo construí un SaaS multi-rubro para gestionar expensas en Argentina con FastAPI + Vue 3 🚀 I Built an Ethical Hacking Scanner Tool – Open Source Project I Replaced /usage and /context in Claude Code With a Single Statusline A Pythonic Way to Handle Emails (IMAP/SMTP) with Auto-Discovery and AI-Ready Design I Collected 8.9 Million Polymarket Price Points — Here's What I Found About How Markets Really Move EcoTrack AI — Carbon Footprint Tracker & Dashboard Everyone's Using AI. No One Agrees How. 5 self-hosted ebook managers worth trying in 2026 Building Your First AI Agent with LangChain: From Chatbot to Autonomous Assistant
Google is Secretly Trying to Run a 4GB AI Model Without Y...
Ria saraswat · 2026-06-19 · via DEV Community
Cover image for Google is Secretly Trying to Run a 4GB AI Model Without You Knowing it! 😱

Ria saraswat

I recently discovered something interesting.

If you're using Google Chrome, there is a chance that Chrome has downloaded a 4GB AI model called Gemini Nano directly onto your computer without a clear notification. The file is usually stored as weights.bin inside Chrome's local data folders.


What is Gemini Nano?

Gemini Nano is Google's lightweight AI model designed to run directly on your device instead of sending everything to cloud servers. It powers features such as:

  • "Help Me Write" assistance
  • On-device scam detection
  • Text summarization
  • AI-powered browser features and APIs

The goal is faster responses and better privacy because some processing happens locally.


👍 Benefits

  • Better privacy (less data sent to the cloud)
  • Faster AI responses
  • Works even with limited internet connectivity
  • Powers s
    ecurity features like scam detection locally


    👎 Downsides

  • Consumes around 4GB of storage

  • Can increase RAM usage during AI tasks

  • May affect laptops with lower memory (8GB–12GB RAM)

  • Uses bandwidth for downloading the model

  • Many
    users were unaware it had been downloaded


How to Check if You Have It (Windows)

Open File Explorer and check:

C:\Users\YOUR_USERNAME\AppData\Local\Google\Chrome\User Data\

Look for a folder named:

OptGuideOnDeviceModel

If you see a large weights.bin file (~4GB), that's Gemini Nano.


** How to Remove It Completely**

Step 1: Open Chrome

Go to:

chrome://settings/system

Disable:

On-device AI

Step 2: Open

chrome://flags

Search and disable:

Gemini Nano
Prompt API
Optimization Guide On Device Model
Other Gemini-related experimental flags

Step 3: Close Chrome completely.

Step 4: Delete the folder:

OptGuideOnDeviceModel

Step 5 (Optional):
Enterprise/advanced users can block future downloads through Chrome policies. Google has also added controls to disable and remove the model in newer Chrome versions.


My Take

Running AI locally is the future.

But if an AI model takes 4GB of storage and potentially several more GB of RAM while running, users should be informed before it lands on their devices.

Privacy-friendly AI is great.

Silent 4GB downloads? That's where the debate starts.