{"author":"knrz","children":[{"author":"cyanydeez","children":[{"author":"verdverm","children":[{"author":"DonatienMigue","children":[],"created_at":"2026-07-16T13:55:03.000Z","created_at_i":1784210103,"id":48934641,"options":[],"parent_id":48926627,"points":null,"story_id":48884903,"text":"I build something a litle bit different check it here: https://github.com/donatienmigue/TeamBrain","title":null,"type":"comment","url":null}],"created_at":"2026-07-15T20:28:58.000Z","created_at_i":1784147338,"id":48926627,"options":[],"parent_id":48885912,"points":null,"story_id":48884903,"text":"yup, one of the reasons I built gmd to replace qmd, also that I wanted to use Typesense and incorporate llm-wiki concepts. Updating memory does depend on the nature of the change and may impact multiple memories or indexed files, and why llm-wiki concepts are needed too.

Right now, imo, the main reason to build something like this or a harness is to understand the quirks son you can evaluate production grade implementations as the space matures. We are all still very early in the curve.

https://github.com/verdverm/gmd","title":null,"type":"comment","url":null}],"created_at":"2026-07-12T23:26:02.000Z","created_at_i":1783898762,"id":48885912,"options":[],"parent_id":48884903,"points":null,"story_id":48884903,"text":"if this were localai, you could just figure out how to update the memore if the fill changes.","title":null,"type":"comment","url":null},{"author":"rgbrgb","children":[{"author":"verdverm","children":[],"created_at":"2026-07-15T20:57:08.000Z","created_at_i":1784149028,"id":48926913,"options":[],"parent_id":48926469,"points":null,"story_id":48884903,"text":"as always it depends on your hardware, the tiny models for embedding / reranking typically have low latency

qmd is focussed on local to the point of designing around single machine setups and this creates a gap where one runs agents+qmd on their laptop and LLMs on their ai box","title":null,"type":"comment","url":null}],"created_at":"2026-07-15T20:13:38.000Z","created_at_i":1784146418,"id":48926469,"options":[],"parent_id":48884903,"points":null,"story_id":48884903,"text":"looks cool. what is latency like? I haven't used qmd before and it looks like it runs 3 local models.","title":null,"type":"comment","url":null},{"author":"grapefruitsoda","children":[],"created_at":"2026-07-15T20:45:23.000Z","created_at_i":1784148323,"id":48926817,"options":[],"parent_id":48884903,"points":null,"story_id":48884903,"text":"We use capn crunch (for cerealization)

For crypto/stock transfers we ofc use chips n dips","title":null,"type":"comment","url":null}],"created_at":"2026-07-12T21:17:27.000Z","created_at_i":1783891047,"id":48884903,"options":[],"parent_id":null,"points":29,"story_id":48884903,"text":null,"title":"Show HN: Capn-hook for coding agents \u2013 don't grep the same mystery twice","type":"story","url":"https://github.com/cyrusNuevoDia/capn-hook"}