#367 – Sam Altman: OpenAI CEO on GPT-4, ChatGPT, and the Future of AI
Lex Fridman Podcast
How to Train a Machine to be Human-Friendly
The science of human guidance is an interesting and important field that requires understanding how to make guidance usable, wise, ethical, and aligned. It's fascinating to consider what aspects humans are asked to focus on and how their feedback is incorporated. The dataset used for training is pulled from various sources, including news sources, partnerships, and open source databases. While there is content overload, the task is to filter out information and focus on relevant data.
00:00
Transcript
Play full episode
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
