PERSON DIRECTORY
Ilia Shumailov
Ilia Shumailov appears in 2 indexed conversations across Machine Learning Street Talk. This directory brings every appearance, source, TL;DR, digest, and transcript into one searchable feed.
Stealing Reasoning Traces from Proprietary LLM APIs — Ilia Shumailov & Alexander Panfilov
Tim ScarfeIlia ShumailovAlexander Panfilov
Researchers show that encrypted reasoning blobs from Anthropic, OpenAI and Google can be decoded by replaying them into smaller models in the same family, without breaking cryptography, and can move across users, variants and fabricated conversations.That turns hidden API reasoning into a privacy, jailbreak and poisoning surface, while proposed fixes—context-bound encryption, access hierarchies and leak classifiers—leave unresolved how much capability is lost when reasoning is withheld.
AI Agents can write 10,000 lines of hacking code in seconds [Dr. Ilia Shumailov]
AI agents invalidate security controls designed for bounded humans, increasing the value of fine-grained access, observability, and policy-enforced execution.CaMeL keeps secrets outside the model and controls data flow through an interpreter, while open-model supply chains and architectural backdoors remain material risks.