The days of navigating either bad or convoluted documentation are over. Once I started pulling them in to a local LLM to simplify and interact with there was no going back. If anything, I've found this to be such an "a-ha" moment for people who aren't even passively interested in ai, that they re-think what is possible. This is progress.

If anything it's an awareness problem at this point. ChatGPT has such a huge volume of documentation specific chatbots, but we're still in that period where people will Google it first out of habit, and seek and interact with documentation as they have for the last 20 years online. 

In the past I've directed companies to Help center as a front end, which has significantly helped reduce the amount of support tickets they take on. 
Over the next 5 years I expect this kind of interaction with documentation to increase exponentially.

off & onchain operations
<--->
Sharing ideas over at /workflow <--->
Problem solving is my jam😎
🎩

seemore.tv/free/fc/adam-

Google’s products are never simple to integrate, and their documentation always sucks.

I originally quote casted this, but there's a bug where if you save a quote cast as a draft, it will show up in the preview, but vanish when you publish. @linda I'm not sure where to surface this bug. Will ping elsewhere if advised going forward.

I originally quote casted this, but there's a bug where if you save a quote cast as a draft, it will show up in the preview, but vanish when you publish. @linda I'm not sure where to surface this bug. Will ping elsewhere if advised going forward. 
https://warpcast.com/applefather.eth/0xbbaece2b

If anything it's an awareness problem at this point. ChatGPT has such a huge volume of documentation specific chatbots, but we're still in that period where people will Google it first out of habit, and seek and interact with documentation as they have for the last 20 years online. 

In the past I've directed companies to Help center as a front end, which has significantly helped reduce the amount of support tickets they take on. 
Over the next 5 years I expect this kind of interaction with documentation to increase exponentially.
https://help.center/

Olama via Cursor. 
To pull in docs locally I'm using HTTrack 
http://httrack.com/page/1/en/index.html
While I haven't benchmarked it, I've found it to be noticeably faster with this method (likely due to caching) than if I were to just reference the URL of the documentation directly. However, I still do reference live documentation or gitrepos via their URL depending on the situation. 

Tagging @pjc who seemed interested in this as well.

Startup artist seeking inspiration.  Champion of /bcard

What local llm are you using?  How are you pulling docs into it?