Episode Details
Back to Episodes“LLMs could control their host machines by exploiting inference engines” by beyarkay (Boyd Kane)
Description
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. Could a malicious LLM gain control of the host machine where its weights are loaded? Such a machine is a high-value target: it has sufficient compute to run a frontier LLM, offers easy access to the LLM's weights, and has privileged access to other computers in the datacentre compared with a generic computer on the internet.
This essay explores how easily a malicious LLM could take control of the host machine. The primary attack considered here involves the LLM emitting a token sequence whose semantic meaning is irrelevant but that exploits a vulnerability in the software that loads an LLM onto GPUs, runs the LLM to generate output tokens, and parses those tokens into responses. .
How could an LLM execute code on the host machine?
Like any program, inference engines like vLLM or SGLang may contain exploitable bugs. Because the LLM controls the tokens passed to the inference engine, a malicious LLM could therefore emit a sequence of tokens that a poorly written [...]
---
Outline:
(01:06) How could an LLM execute code on the host machine?
(01:36) vLLM previously used eval() on tool-call parameters
(02:43) vLLM and SGLang are complex, and bugs are common
(04:15) Vision and audio tokens might increase the attack surface
(05:27) How likely is an LLM to discover and exploit inference engine vulnerabilities?
(06:01) Tool use could make exploitation reproducible
(06:29) Inference engines are an attractive target for power-seeking LLMs
(07:25) How do we defend against this?
---
First published:
August 24th, 2026
---
Narrated by TYPE III AUDIO.
---
Images from the article:
Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.