3 ms·
This should in theory be solveable by using a custom frontend and only using the various backend APIs as stateless inference providers, but everything I've test
by lxgr 8mo ago
This should in theory be solveable by using a custom frontend and only using the various backend APIs as stateless inference providers, but everything I've tested falls flat on a few aspects: Chat history RAG and web search, and to a lesser extent tool use.
Yes, all of these are theoretically possible (the APIs now all support web search, as far as I know, there are RAG APIs too, and tool use has been supported for a while), but the various "chat" models just seem to be much better at using their first-party tools than any third-party harness, which makes sense that this is what they've been trained on.
- peteforde 8mo agoI've had friends suggest a custom frontend several times, but unless that frontend starts off by faithfully downloading and recreating my entire chat history... now I just have two problems.
- lxgr 8mo agoThat part should be fairly easy to solve, no? At least ChatGPT allows exporting your entire chat history; importing that into whatever frontend seems well within a current agent's capabilities.
- peteforde 8mo agoTo which I can only admit that, oh wow, I did not realize that there was a big Export button in that menu. Thank you! I hope this works out.