Ask HN: How you manage local long lived research projects and LLM's?

4 points by ripplefringe 4 hours ago | 1 comment

TLDR: Do you have a pattern or structure for doing long-lived research or non-coding projects using LLMs?

I was kind of inspired by this hacker news article where someone collected data to solve their fatigue problem: https://news.ycombinator.com/item?id=48605117

A similar but smaller I did was to collect nut-free restaurants to eat in Chicago. So I had articles, reviews, emails, phone calls, all sorts of stuff. I found it very easy to just use Copilot as my interface and I just data into it over and over and asked it to organize it.

https://github.com/joshuabremer/chicago-allergy-eats

This seems like a common thing where people would be working on a longer-term research or data collection task where you use the LLM to answer questions or help you accept the data. Does anyone have a standard format for this? Or an example that is working for you?

tylerhackbart 4 hours ago |

Accessibility data collection, I am running a local application with a mongoDB connection so that when I decide to deploy it fully after validation then I can without having to collect the data again.

Honestly data structures don't have to be perfect unless you are feeding it into another service that you don't control. If you end up wrapping the research into another piece of the puzzle you control both sides anyways.