This week's Searls After Dark is a day late, but $0 short. We covered a lot of ground by converting both our API calls to OpenAI from blocking HTTP to streaming SSE handling and from traditional page renders to streaming responses as ChatGPT "types" in real-time over websockets using Turbo Streams.
Full episode here: https://www.youtube.com/watch?v=cEzvU4PJq-A
If you haven't trudged through the pain of software development with me, this clip of joy and relief is what the end of most episodes looks like.