I’m pretty excited about this. At @bsidesnash on April 23, 2027, I’m going to be running the AI Hacking Village. This will be a chunk of the conference focused on LLMs, agents, local inference, and all of the security mess/fun that comes with it. We are putting together a GPU cluster to run open weights at the conference, and we’re going to teach people how to run LLMs locally themselves. We’ll have jailbreak challenges, prompt injection, hostile web/RAG content, system prompt extraction, agents with tools, red teaming, blue teaming, sandboxing, and models with different levels of guardrails, including abliterated models. If you’re a hacker who hasn’t spent much time with LLMs, you should be able to walk in, learn how to run one, understand the attack surface, and start breaking things. If you’re already deep into AI, agents, red teaming, adversarial ML, or local inference, there should be plenty there for you too. We have a bunch of other fun stuff planned around actually learning, hacking, and experimenting with this stuff. I’m also looking for sponsors and hardware donations to help build out the GPU cluster. So if you work at NVIDIA, AMD, Intel, an AI lab, a GPU/cloud company, a hardware vendor, or anywhere else that might want to help us put a stupid amount of compute in front of a bunch of hackers and tell them to go break it, please reach out.