And here is another useful utility:
https://t.co/6AbJbazmH3
It generates a prompt template for structured objects based on a JSON schema. It is designed to be used in with the markdown-parser utility which will convert the LLM response back to a structured object.
Enjoy :)
I just published markdown-parser. It is a useful module if you need structured output from your LLM but don't want to rely on JSON output (which has limitations).
It parses markdown based on the schema you provide.
https://t.co/GmN6FOEPjL
Let me know it is useful for you :)
Chatbots will never become big because we (humans) are bad at asking good questions
Even with AGI, most people wouldn't know what to ask
Whoever can get LLMs to do stuff for us without chat as the main interface will likely win. (Chat still needed but as an auxiliary feature)
@OfficialLoganK@OpenAI We are building a coding workflow automation tool. Instead of a coding assistant the AI leads you through the process and builds fully functional apps for you. It can build all kinds of software but we are focussing on backend apis currently.
@Madisonkanna For greenfield projects absolutely. For maintaining legacy software - not. However, if the AI built the codebase itself, then it will also be able to extend and maintain it. I think first examples will come this year
@EmmaVitz@jamonholmgren They should care and they do. Thats why they rent more expensive office space that is central to everyone instead of somewhere remote
@BVG_Ubahn anscheinend vollkommen unerwartet fΓΌr die bvg. keine klare info in der app. busse kommen einfach nicht. schwimmkurs verpasst und sohn mit wutausbruch. danke
@svpino TDD is only usefull if you are 100% sure about the application design / concept. Otherwise you will spend more time adapting tests than coding.
Not practical for early prototyping and POCs