Maksym Andriushchenko

mentions 3 type Person feed RSS

// recent coverage 3 mentions

23:24

2026-08-14

lesswrong.com

artificial-intelligence

Your Agents Are Not Time Aware

Coding agents Claude Code and Codex consistently over-predict their own wall-clock runtime, according to research conducted as part of MATS 10 with Maksym Andriushchenko. The study introduced AgentTim…

20:08

2026-08-11

sourcefeed.dev

artificial-intelligence

The Labs' Hidden Reasoning Was Never Actually Hidden

A new paper by Alexander Panfilov, Ilia Shumailov, Luca Beurer-Kellner, Jonas Geiping, Maksym Andriushchenko and colleagues demonstrates that encrypted reasoning traces from proprietary LLM APIs from …

18:08

2026-08-11

sourcefeed.dev

artificial-intelligence

The Chain-of-Thought You Sealed Isn't Secret

A new preprint, 'Stealing Reasoning Traces from Proprietary LLM APIs,' reveals that encrypted chain-of-thought blocks from Anthropic, OpenAI, and Google can be replayed into weaker sibling models to e…

// co-occurs with top 8 entities

OpenAI 2 Anthropic 2 Google 2 Alexander Panfilov 2 Ilia Shumailov 2 Jonas Geiping 2 Claude Code 1 Codex 1

// topics top 6 topics

artificial intelligence 3 ai research 2 ai safety 2 large language models 2 ai agents 1 ai policy 1