New OpenAI blog post explains how the "Red-line principles" in the Model Spec are intended to apply for localization and customization. We didn't have a blog post on the Spec update where those were added, so I'm not sure how many people have seen them. Would love feedback on the principles and approach!
https://t.co/iK1KmMEb5H
How does @OpenAI decide when to release its models?
What's the theory behind those decisions? What have they learned over the last few years?
@Ziad_Reslan joins @scaling_laws to answer these and other questions! Don't miss this episode!
We updated the OpenAI Model Spec — our evolving document on intended model behavior — to capture some of our latest research.
🤖 principles for agents
🎭 more style details
🛡️ safe completions
✨ and lots more updates
The Schmidt Program at @JacksonYale will host a discussion on generative AI safety this Thursday, Nov. 7, with Senior Fellows Beth Goldberg and Ziad Reslan. @Jigsaw@OpenAI
Register: https://t.co/nLbec71MQT
Back by popular demand for a 2nd year in a row, @camillefrancois and I get to interview @dseetharaman (WSJ) & @ZoeSchiffer (Platformer) at #trustcon23 on Thursday!
We get to be the ones asking questions for a change!
What question would YOU want to ask these 2 tech reporters?
Khalil Gibran is one of the best-selling poets of all time.
This Arab American Heritage Month, we celebrate the work of the poet, whose English and Arabic writings on alienation, mysticism, and life resonate with immigrants around the world.
Internet outages are happening more frequently worldwide, including in parts of Iran this week. Across Google, teams are working to make our tools broadly available, following the newly updated US sanctions applicable to communications services. https://t.co/iojU4Kuq3N 1/5
Had a ton of fun grilling @CaseyNewton for this panel, thank you @Ziad_Reslan for being an excellent co-conspirator here and Casey for patiently answering all of it.