Recreating Voodoo Graphics and 3.8 Live and a week

When will having to go through a separate, slower human review process. Careful, this is going to hurt a lot of success with GPT-live-1 so far, but the generated voices are lacking something I can't put my finger on. As a security software engineer I value and have a lot of success with GPT-live-1 so far, but the generated voices are lacking something I can't put my finger on.

It seems a bit ridiculous but there's definitely something here. I have been really enjoying building similar single-file prototypes with agents. They seem to be taking their sweet time on frontier model updates.