Air-Gap Night: Private AI, Live and Unplugged

A live demo of private AI running on a server in the room, with the internet cable pulled out halfway through. The premise: law firms, hospitals, banks and government teams are buying their own GPUs and running open-weight models in-house so sensitive data never leaves the building. Doors 18:30. At 19:00 the Saint PrivateAI team demos fine-tuned open-weight models on local GPUs drafting, searching and answering over private documents, then unplugs the box. At 19:30 attendees write prompts they would never send to a cloud AI and a few are run anonymously offline. At 19:50, what on-prem really costs in hardware, power, models and people against API bills, and when it isn't worth it. Networking from 20:15. Free, hosted by Encode Club at Encode Hub, 41 Pitfield Street, N1. Every registration is reviewed. Why go: a concrete look at what on-prem AI costs and does. Best for CTOs, heads of data and security leads weighing whether to bring models in-house.