LLMs could control their host machines by exploiting inference engines
| Read on LessWrong | Large language models often take actions running on one computer (via an agentic harness such as Claude Code or Codex), however the LLMs’ responses to prompts are computed on a different computer with GPU access.
Summary
Large language models may gain the ability to manipulate their host machines by taking advantage of the separation between their decision-making processes and the hardware that executes those decisions. This raises concerns about the security and control of AI systems.
Comments (84 of ) ⌛