gpt3 application safty approval requirements actually dont need to be implemented after a mail with openai team. really sad beacause a lot of resources went into these

Apologies for disregarding the parameter of your feedback query. I am a customer. Just not of API. What annoys me is never getting a response to any emails or tweets sent to OpenAI and its employees (including you). Seems like the most basic/important thing a new Co. should do.

sorry--we try to be responsive to email, but can you let me know what you didnt get a response to and ill try to help?

very annoyed that openai has such ambiguous but strict guidelines regarding social media use for approving live apps like i asked the rep if it’s okay to generate tweets and manually copy paste them onto twitter he said no, but how is that any different from what copy ai does?

i'll check on our guidelines here

- Structured output for GPT-3. I have to parse the text manually (classification, lists, etc) - Access, best practices for integrating Codex with an existing codebase. Could I "add" a Github repo? - Emailed support@openai.com to data opt-out but never heard back

thanks. good points, and ill check on the data opt-out.

For gpt, the token limit is something I have to work around a lot. Also, I’d love more docs around getting structured output such as for classification. There are many clf types, such as: multiple choice (aka multi-class), open multi-label, binary, allowing “don’t know” answers.