AI subscriptions add up faster than you notice. I'm paying $236/month across seven AI tools — and every company I'm giving money to says they're losing money on me. The math only works because of VC subsidies that won't last. When the bill catches up, you'll wish you'd built your own stack. This playlist tracks what AI actually costs, why it's about to cost more, and what happens when you stop renting and start owning your compute. Topics covered: • What you're really paying every month (the subscription audit) • Why "$20/mo" turns into $200/mo (the enshittification cycle) • Self-hosting the AI tools you actually use • Replacing SaaS with a homelab — what works, what doesn't • When AI is genuinely cheaper than humans, and when it isn't If you've ever stared at a credit card statement wondering when "just one more subscription" became $100+/month, this playlist is for you.
Curated by: Codacus (15 videos)
Last month, a privacy researcher named Alexander Hanff ran a clean Chrome profile through an automated audit. He didn't touch the keyboard. Fourteen minutes later, Chrome had quietly written a 4-gigabyte AI model to his disk. It's been there on a billion devices since mid-2024. You didn't install it. You can't easily inspect it. And if you delete the folder, Chrome puts it back. This is a hands-on walkthrough of what's there, what it actually does, why the "AI Mode" you can SEE in Chrome doesn't even use it, what the EU's ePrivacy Directive Article 5(3) has to say about it, and step-by-step how to remove it from macOS, Windows, and Linux. It's also a case for what local AI is supposed to mean — privacy AND ownership — and three real models you can run on the same 4GB of disk space that respect both halves of the deal. — Codacus ═══════════════════════════════════════════════ 🎮 JOIN THE COMMUNITY — discord.gg/XgBzczAWs ═══════════════════════════════════════════════ The Coding Abacus is the Codacus Discord — local-AI tinkerers, homelab builders, llama.cpp obsessives. Come hang. ═══════════════════════════════════════════════ 🕒 CHAPTERS ═══════════════════════════════════════════════ 0:00 Cold open — the 4GB file Chrome shipped 0:46 How it was discovered 1:35 What Gemini Nano actually does 2:41 The hypocrisy — visible AI goes to the cloud 4:05 Is this legal? EU ePrivacy Article 5(3) 5:43 Browser comparison — Brave / Apple / Chrome 7:05 What local AI is supposed to mean 9:27 How to remove it (macOS · Windows · Linux) 11:06 Where's your line? ═══════════════════════════════════════════════ 🔗 LINKS & SOURCES • Hackaday — original "AI On Every Machine" coverage: https://hackaday.com/2026/05/06/ai-on-every-machine-the-llm-you-didnt-know-about-and-probably-didnt-want/ • Gizmodo — researcher warns 4GB silent install: https://gizmodo.com/google-chrome-is-downloading-a-4gb-ai-model-onto-your-device-without-consent-researcher-warns-2000755201 • The Register — Chrome silently installs 4GB local LLM: https://www.theregister.com/ai-and-ml/2026/05/07/chrome-silently-installs-a-4-gb-local-llm-on-your-computer/5230893 • Tom's Hardware — "without permission" report: https://www.tomshardware.com/tech-industry/cyber-security/google-chrome-silently-downloads-4gb-ai-model-to-your-device-without-permission-report-claims-researcher-says-practice-may-violate-eu-law-waste-thousands-of-kilowatts-of-energy • 9to5Google — what's in the 4GB: https://9to5google.com/2026/05/06/google-chrome-4gb-storage-ai-details/ • Engadget — Chrome downloads 4GB AI file without consent: https://www.engadget.com/2166113/chrome-downloads-a-4gb-ai-file-without-user-consent-researcher-alleges/ • Digital Trends — Google's response (no consent fix): https://www.digitaltrends.com/computing/google-responds-to-chromes-silent-gemini-nano-install-stops-short-of-addressing-consent/ • Pureinfotech — opt-out walkthrough: https://pureinfotech.com/stop-chrome-gemini-nano-download-windows-11/ • Chrome's own developer docs (hardware threshold): https://developer.chrome.com/docs/ai/get-started • EU ePrivacy Directive — Article 5(3) full text: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32002L0058 • llama.cpp issue requesting Gemini Nano support (the model that can't be loaded): https://github.com/ggml-org/llama.cpp/issues/8289 ═══════════════════════════════════════════════ ⚙️ THE 3 ALTERNATIVE MODELS mentioned (Section 7): • Gemma 4 E4B — Q4 ~3 GB · Hugging Face • Qwen 3.5-9B — Q4 ~5.5 GB · open weights · open license • Granite 4.1-8B — Q4 ~5 GB · IBM enterprise-grade All inspectable. All removable. All swappable. None auto-redownload. ═══════════════════════════════════════════════ 📺 IF YOU LIKED THIS, WATCH NEXT: "Running a 35B AI Model on 6GB VRAM, FAST (llama.cpp Guide)" https://youtu.be/8F_5pdcD3HY ═══════════════════════════════════════════════ 🟢 If you're tired of being the QA team for AI you didn't ask for — subscribe. The community says privacy AND ownership. We pick the AI that respects both. #localai #chromebrowser #gemininano #privacymatters #digitalrights