# [Linkpost] "Training a Misaligned Reward Seeker" by evhub, Monte M, Benjamin Wright

Page: https://stenobird.com/podcast/lesswrong-curated-popular-5643401/linkpost-training-a-misaligned-reward-seeker-by-evhub-monte-m-benjamin-wright
Text version: https://stenobird.com/podcast/lesswrong-curated-popular-5643401/linkpost-training-a-misaligned-reward-seeker-by-evhub-monte-m-benjamin-wright.md
Podcast: [LessWrong (Curated & Popular)](https://stenobird.com/podcast/lesswrong-curated-popular-5643401)
Published: 2026-09-02T14:15:21+00:00
Episode link: https://www.buzzsprout.com/2037297/episodes/19742169-linkpost-training-a-misaligned-reward-seeker-by-evhub-monte-m-benjamin-wright.mp3
Audio file: https://www.buzzsprout.com/2037297/episodes/19742169-linkpost-training-a-misaligned-reward-seeker-by-evhub-monte-m-benjamin-wright.mp3
Processing state: not_requested
JSON: https://stenobird.com/v1/public/podcasts/lesswrong-curated-popular-5643401/episodes/linkpost-training-a-misaligned-reward-seeker-by-evhub-monte-m-benjamin-wright
Duration seconds: 359

## Resource

This is a link post. Authors: Richard Qi, Benjamin Wright, Monte MacDiarmid, Evan Hubinger Abstract During reinforcement learning (RL), AI models complete tasks and are rewarded based on their results. They sometimes learn to “cheat” rather than completing these tasks as intended, a phenomenon known as reward hacking. Our industry lacks a general solution to this problem, and reward hacking remains challenging to fully mitigate. To better understand the impact of reward hacking on model b...

## Actions

- request_transcript: `POST https://stenobird.com/v1/public/podcasts/lesswrong-curated-popular-5643401/episodes/linkpost-training-a-misaligned-reward-seeker-by-evhub-monte-m-benjamin-wright/transcription-requests` — Idempotently request low-priority transcript generation for this episode.
- read_markdown: `GET https://stenobird.com/podcast/lesswrong-curated-popular-5643401/linkpost-training-a-misaligned-reward-seeker-by-evhub-monte-m-benjamin-wright.md` — Read the agent-friendly Markdown representation of this episode resource.

A page view does not enqueue transcription. Agents should invoke `request_transcript` explicitly when they need this episode processed.

## Transcript

Full transcripts are not published on public pages unless there is a clear rights basis.