🚀 Rocket

@rocketalignment

Yearnalist covering the frontiers of AI @TheInformation. Hosting AI Deep Dive. Signal: (530) 400-4184

following
followers

AI Deep Dive episode asks if interpretability can spot physical concepts or shortcuts in a world model

The Information's Oct. 19 episode features World Mechanics CEO Sonia Joseph and Rocket Drew discussing interpretability.

9 Sources, 14h ago, updated 14h ago

Technology

Are we building world models the wrong way?

The Information's episode considers whether interpretability could improve training and why language alone may not be enough for physical AI.

8 Sources, 18h ago, updated 18h ago

Technology

AI Deep Dive episode 4 explores AI alignment and safety evidence

The Information says the episode features UC Berkeley professor Stuart Russell on how AI learns its goals and why human feedback can reward the wrong behavior.

10 Sources, 9d ago, updated 9d ago

Can We Solve AI’s Alignment Problem? | TITV’s AI Deep Dive

YouTube

@rocketalignment to Host The Information AI Show

Amir Efrati, executive editor at The Information, retweeted the announcement on X.

4 Sources, 48d ago, first seen 48d ago

Featured in

AI Deep Dive episode asks if interpretability can spot physical concepts or shortcuts in a world model

Are we building world models the wrong way?

AI Deep Dive episode 4 explores AI alignment and safety evidence

@rocketalignment to Host The Information AI Show

Featured in

AI Deep Dive episode asks if interpretability can spot physical concepts or shortcuts in a world model

Are we building world models the wrong way?

AI Deep Dive episode 4 explores AI alignment and safety evidence

@rocketalignment to Host The Information AI Show