There’s been some positive changes and challenges for me on my other projects since the last autumn, explaining the quietness here. Now things have progressed so that I was able to find time on my home programming desk.
Large language model APIs are on the focus again. Earlier this year I was running some tests on multi agent discussions, and now I’ve been trying out a new angle to the GPT Philosopher I was working on the summer 2023.
This time the project is on philosophical article analysis. Currently the user can upload any number of PDF articles to the app repository, these are converted to txt, and then a large preset question taxonomy is ran over the articles. As before, Obsidian is used for the results UI. Questions extract main concepts, references, and similar information, in addition to somewhat broader analyses related, for example, to different fields in philosophy.
Next steps include visualizations on the articles (radar chart, etc.), enhanced summaries, and maybe having the agentic discussions somehow included in the show. The idea is currently keeping it more informative than entertaining, but obviously this kind of systems have huge potential also on the latter direction.