@queerlilhayseed@piefed.blahaj.zone Thanks for the positive response! You basically named the two hurdles that ate most of my development time on my projects system.

On the counting/nondeterminism thing: the trick was giving up on the LLM as a “knower” entirely and treating it as nothing more than a reader. So if I want to know how many times grass shows up in some sprawling Tolkien-esque passage, I’m not asking the model “how many times does grass appear” and hoping for the best. I use it to pull out and tag the relevant sentences into a structured database, and then a dumb, boring, deterministic Python script does the actual counting. Temperature’s pinned at 0, and I force it to hand back verbatim quotes with citations instead of letting it synthesize a summary. That alone killed most of the hallucination problem.

The “systemic wrongness” side (bias, basically the model just inventing stuff) came down to hybrid search, meaning keyword/BM25 alongside vector embeddings. Pure semantic search will absolutely whiff on a weirdly named character or a specific term just because it doesn’t “feel” semantically close to the query, so having the keyword layer as a backstop matters a lot. And I’m strict about grounding: no direct source quote from retrieval means the system says “I don’t know” instead of guessing. It’s not allowed to fill gaps with vibes.

Your instinct about cataloging prompts was right on too. I keep a golden dataset of queries I run every time I touch the system, and rather than some vague overall correctness score, I bucket failures by type: was it a retrieval failure or a synthesis failure. That distinction alone tells you whether the fix is in your chunking strategy or your prompting, instead of just flailing at both.

Honestly the tools have come a long way since you last looked at this stuff, and I think it’s mostly a philosophy shift. People stopped treating LLMs like databases and started treating them like interfaces sitting in front of one.

10 years ago I would have said maybe. Now? Practicly everything that you run on Windows will run, in some form or another, on Linux. Even the hard to crack gaming is now becoming a non issue. There really is no point in Windows any longer.

Ive been doing this for 6 years now. Just made sense. This is cool it’s a default but most likely will only get used by the folks that manually created years earlier.

so i’ve seen some great answered here. Some folks are leery of some companies but so far i have found that Tailscale and Cloudflare a great combination. My ISP doesn’t let you port forward and TS & CF circumvent that nicely. Zero issues, both free services and saves me a ton of headache.

Zotero? Really? Isn’t that like massively overkill for bookmarks? I mean, to each their own, but I guess if you want source citations, screenshots, web pages and more to sync bookmarks, have at it mate!

I’m onboard with this! Great idea that only takes a little effort and big returns! The small web can stay small but some aids in discoverability is not a bad thing!