VDAOvideo

The Bare Minimum Setup You Need to Run AI Agents Locally

Daniel Jindoo · 81.6K views · 3 months ago
Description

🟢 Join the Free Local AI Guild, 250+ Members : https://www.skool.com/join-dlag/about 🔴 Read Practical AI on Substack, 3K+ Subscribers : https://danieljindoo.substack.com/ Local AI hardware got worse and better at the same time. RAM prices are up 3 to 4x, NVIDIA is skipping a gaming GPU generation, and the local AI build guide I published last year will now cost you more than it should. This is the updated local AI PC build guide, rebuilt around one rule I undersold last time: capacity decides what you can run, bandwidth decides how fast it runs. I open with four corrections from the last video, all of them from the comment section. Mac users should reach for MLX, not GGUF. Inference is memory bound, so tokens per second follow bandwidth once the model fits. 64GB of system RAM is wasted