LLM as assistant for code design idea generation (was: On keybindings and the slow erosion of help's utility)
Ihor Radchenko <[email protected]>
| Newsgroups | gmane.emacs.devel |
|---|---|
| Message-ID | <87v7ae6wgk.fsf@localhost> |
Richard Stallman <[email protected]> writes: > > However, FYI, what I posted is a > > result of several dozens of iterations, later edited by me, and produced > > with LLM specifically tuned to be more straight to the point. I could > > rewrite it myself, but that would honestly not gain much. > > I think you mean it would not be a big technical improvement in the > code -- right? That may be so, but it's a different issue. No. The code quality is honestly completely irrelevant. That code is not serious and is more or less a pseudo-code. It is there simply for an illustration of the ideas. > However, this sort of working together with an LLM may be useful to > discuss specifically in the development of the criteria we are working > on. That's exactly what I used LLM for - throw my ideas, ask for critique, brainstorm suggestions, just to refuse LLM suggestions and come up with my own, etc. I also used it to generate examples how the concept would look like if applied to typical dynamic keymap situation we discussed. -- Ihor Radchenko // yantar92, Org mode maintainer, Learn more about Org mode at <https://orgmode.org/>. Support Org development at <https://liberapay.com/org-mode>, or support my work at <https://liberapay.com/yantar92>