Impressed with their long document summarization, any ideas how they do this? Seems beyond normal GPT limitations; either they have a more powerful model (doubt it) or hacked around the limitations?
We have an in-house model we've been developing since 2019, just for summarization of long documents in real time. We'll try to find some time to blog about the high level design.
Jealous. I have been trying to 'configure' my own local summarizer. As I find the ChatGPT limitations on length to be frustrating. For example, I wanted to take a transcription of a troubleshooting session with user and support agent (me). I wanted a summary of what we did essentially, overall. But doing this with ChatGPT takes so. many. copy/paste actions. I just wanted to give it all, and say, get a summary of every 5 minutes.
Secondly, this lab needs to stick around. It makes me want to subscribe, just for access to this.
> an in-house model we've been developing since 2019
We will be very interested in how you tackled the problem of "understanding" the input text - to differentiate it from current "fakeries of actual speakers".
I.e. how you implemented the "intelligent" parts - or, which simulation of actual intelligent processing (if any) it contains.
e.g. good summary for a very long text https://labs.kagi.com/ai/sum?url=http://localroger.com/prime...