To a wedding

built conversation support into `npm install promptgun`! Call `.createConversation()` on the ai client and you get a dedicated client for a conversation, where each new prompt gets appended to the existing conversation internally, so you can remove conversation management boilerplate! #promptgun

managed to make a friend interested in promptgun (npm install proptgun). he ununiquely also needed to prompt for json output and would rather stream that output, which made it hard to keep things type safe and to react to the incomplete streamed json, so they had a lot of boiler plate, and he was already using typescript and the openai api so promptgun was perfect for him #promptgun

Couldn’t code today

Made a bit of progress on next crawler generation but don’t have much time today

Improving crawler

Found myself needing to build retry logic for every LLM call. So I deleted a lot of code from my app by building retry logic into Promptgun! `npm install promptgun` Just add the attemptsPerCall parameter to your setupAI call. See www.npmjs.com/package/promptg… #happenlist

More power to the 👮‍♂️ #happenlist

split crawling up into steps with varying frequencies #happenlist

Rest day

Waitlists #happenlist

Switched city search to google maps / backend based #happenlist

Added search. Majorly improved performance (I am very surprised memoized computable properties aren’t built into swift, it’s a giant performance saver that’s extremely easy to program. Now i had to program the memoizing manually) #happenlist

More ui cleaning, AI code with cursor really does not work very well with SwiftUI when native UI component maxing. What’s other people’s experience with this?

cleaning UI after recent additions #happenlist

Chilling with family

Crawling the categories

feat: basic filter function on supercategories, very ugly #happenlist

adding basic event categories #happenlist

started adding a much requested feature, filters ☕️ #happenlist