54sec snip

The 80000 Hours Podcast on Artificial Intelligence cover image

Two: Ajeya Cotra on accidentally teaching AI models to deceive us

The 80000 Hours Podcast on Artificial Intelligence

NOTE

Reward Plans, Debate Models

Rewarding plans instead of outcomes and having models debate each other are strategies to prevent rewarding scheming or sycophancy in training and reinforcement approaches. By rewarding a model for creating sensible plans and another for pointing out flaws in the plan, the debate approach can enhance the effectiveness of the training process.

00:00

Get the Snipd
podcast app

Unlock the knowledge in podcasts with the podcast player of the future.
App store bannerPlay store banner

AI-powered
podcast player

Listen to all your favourite podcasts with AI-powered features

Discover
highlights

Listen to the best highlights from the podcasts you love and dive into the full episode

Save any
moment

Hear something you like? Tap your headphones to save it with AI-generated key takeaways

Share
& Export

Send highlights to Twitter, WhatsApp or export them to Notion, Readwise & more

AI-powered
podcast player

Listen to all your favourite podcasts with AI-powered features

Discover
highlights

Listen to the best highlights from the podcasts you love and dive into the full episode