I’ve been playing around with using the OpenAI API to generate OpenAPI specs from HTML documents. It’s pretty tricky, mostly because of token limits on the models I have available to me, I think. Currently the model I have access to with the largest token limit is gpt-3.5-turbo-16k, and it still results in outputs getting cut off.
1 Like
Today I explored meeting transcription services to help me better keep on top of my responsibilities with my clients. I can be quite forgetful, so I’m hoping this will help me build trust with the people I’m doing work for. Currently I’m trying out Bloks.
1 Like
I’ve had in the back of my mind for a while now to go to the library a certain day of the week.
2 Likes
Yes! I’m still working at the library regularly. It’s continued to be a great thing for me.
2 Likes
I have code that I use to generate my invoices for my clients. Today I modified it to do better logging and send the invoices to my taskratchet email instead of my personal email.
2 Likes
I’ve done some more work on creating an openapi spec for toggl track.
In other news I finally got access to gpt 4 via the openai api, and I’ve updated my bot to use the newer model. It does indeed feel like an improvement.
1 Like
Continuing to create the toggl track openapi schema.
At some point I’m also planning on exploring using OpenAI’s new beta assistants API for my Telegram bot. It seems like it would simplify a lot of my code.
1 Like