I disconnected from the internet and ran AI entirely on a $400 laptop — it was surprisingly usable
🩺 Summary
What happens when you take a mid-range 2023 laptop with 16GB RAM, install Ollama, and refuse to touch any cloud API for a week? Turns out, 7B-14B models get you surprisingly far for everyday AI tasks.
📝 Details
Stack: ThinkPad T14s (AMD Ryzen 7, 16GB RAM, no GPU), Ollama with Qwen2.5-7B and Llama-3.1-8B.
**What worked**: Code completion via continue.dev was nearly as fast as Copilot (1-2s latency). Summarization and brainstorming were solid.
**What didn’t work**: Long context reasoning beyond ~5K tokens. Creative writing was flat. Multilingual code-switching was poor.
**The real takeaway**: 70-80% of what the average developer uses AI for can be handled by a local 7B-8B model on consumer hardware. The gap is closing fast.
💬 Comments (0)