Ryan Prellwitz

Owner of Vines & Rushes Winery located in Ripon, WI. Our focus is on creating quality wines from Wisconsin Grown grapes.

[Ryan Prellwitz @ryanprel]

@patrickc Check out my fork of Multica - added memory, channels, GitHub ship integration for multiple repos, internal and external MCP servers, all in a multi agent, multi human environment that can use oauth for agents instead of api.

https://t.co/xXSF6iklrk

[Andrej Karpathy @karpathy]

LLM Knowledge Bases

Something I'm finding very useful recently: using LLMs to build personal knowledge bases for various topics of research interest. In this way, a large fraction of my recent token throughput is going less into manipulating code, and more into manipulating knowledge (stored as markdown and images). The latest LLMs are quite good at it. So:

Data ingest:
I index source documents (articles, papers, repos, datasets, images, etc.) into a raw/ directory, then I use an LLM to incrementally "compile" a wiki, which is just a collection of .md files in a directory structure. The wiki includes summaries of all the data in raw/, backlinks, and then it categorizes data into concepts, writes articles for them, and links them all. To convert web articles into .md files I like to use the Obsidian Web Clipper extension, and then I also use a hotkey to download all the related images to local so that my LLM can easily reference them.

IDE:
I use Obsidian as the IDE "frontend" where I can view the raw data, the the compiled wiki, and the derived visualizations. Important to note that the LLM writes and maintains all of the data of the wiki, I rarely touch it directly. I've played with a few Obsidian plugins to render and view data in other ways (e.g. Marp for slides).

Q&A:
Where things get interesting is that once your wiki is big enough (e.g. mine on some recent research is ~100 articles and ~400K words), you can ask your LLM agent all kinds of complex questions against the wiki, and it will go off, research the answers, etc. I thought I had to reach for fancy RAG, but the LLM has been pretty good about auto-maintaining index files and brief summaries of all the documents and it reads all the important related data fairly easily at this ~small scale.

Output:
Instead of getting answers in text/terminal, I like to have it render markdown files for me, or slide shows (Marp format), or matplotlib images, all of which I then view again in Obsidian. You can imagine many other visual output formats depending on the query. Often, I end up "filing" the outputs back into the wiki to enhance it for further queries. So my own explorations and queries always "add up" in the knowledge base.

Linting:
I've run some LLM "health checks" over the wiki to e.g. find inconsistent data, impute missing data (with web searchers), find interesting connections for new article candidates, etc., to incrementally clean up the wiki and enhance its overall data integrity. The LLMs are quite good at suggesting further questions to ask and look into.

Extra tools:
I find myself developing additional tools to process the data, e.g. I vibe coded a small and naive search engine over the wiki, which I both use directly (in a web ui), but more often I want to hand it off to an LLM via CLI as a tool for larger queries.

Further explorations:
As the repo grows, the natural desire is to also think about synthetic data generation + finetuning to have your LLM "know" the data in its weights instead of just context windows.

TLDR: raw data from a given number of sources is collected, then compiled by an LLM into a .md wiki, then operated on by various CLIs by the LLM to do Q&A and to incrementally enhance the wiki, and all of it viewable in Obsidian. You rarely ever write or edit the wiki manually, it's the domain of the LLM. I think there is room here for an incredible new product instead of a hacky collection of scripts.

[Blair Dulder CPA™ 🧃 @runaway_vol]

when physical labor became optional we invented the gym. we’ll need the same thing for the mind

[Justine Vanden Heuvel @thegrapeprof]

Under vine cover of fescue on left, glyphosate on right. Great tool to reduce the excessive vigor of Noiret

[Ryan Prellwitz @ryanprel]

@Wisconet @UWMadisonCALS I am in far southwestern Winnebago county on the border with Fond du Lac and close to Green Lake and would be open to hosting a site.

[Ryan Prellwitz @ryanprel]

The fact that @QuickBooks doesn’t have a backup and restore option on every Quickbooks online plan means they really don’t are about their customers at every level. This should be a default capability of every Quickbooks plan and there are no excuses why it isn’t.

[Ethical Skeptic ☀ @EthicalSkeptic]

One ounce of experience is worth five of academic heuristic, ten of expert explanation, twenty of idle speculation, and forty of cynical deflection.

[Ryan Prellwitz @ryanprel]

@thegrapeprof @MNFruitLady I’ve been using a drone as a bird deterrent for the last several years to great effect. We no longer net anything and see essentially zero damage from birds. We haven’t fully automated it - just doing a fly through a few times a day or when we see some birds around.

[Ryan Prellwitz @ryanprel]

@thegrapeprof If you want to do unique and creative things, why try doing that where lots of people are doing variations on a theme? Why not do that where you can create using flavors and profiles most have never experienced before.

[Ryan Prellwitz @ryanprel]

@AdamMcCalvy So, break that record sometime next week?

[Ryan Prellwitz @ryanprel]

@byKyleLobner Easter egg in the MLB app

[Ryan Prellwitz @ryanprel]

@JasonEWhite @schmidtmitchell @InWisconsin That’s great but I have a fiber line running through my ditch and can’t even connect up to it. The only realistic option for high speed internet in many rural areas is going to be @SpaceXStarlink. Traditional tcoms simply don’t see the financial incentive without competition.

[Angelo graziani @Angelograziani5]

Ive been been saying this about Matt ioannidis since 2016. He is the MAN on that team.

[Ryan Prellwitz @ryanprel]

@AlexBerenson The real pandemic here is comorbidities. How does the monthly Covid death count contrast with the death count from heart disease last year?

[Ryan Prellwitz @ryanprel]

@JasonEWhite Thanks!

[GeniusNetwork® @GeniusNetwork]

My friend Dan Sullivan is the founder of Strategic Coach, and has so many incredible quotes.

Here are 8 of my favorite Dan Sullivan quotes:

1. “If you've got enough money to solve the problem, you don't have the problem.”

2. “The skills that got you ou…https://t.co/v5GYrQzA78

[Ryan Prellwitz @ryanprel]

@hanswp @WineBusiness Besides corporate entities, is there actually anyone at all on that list from East of the Rockies?

[Ryan Prellwitz @ryanprel]

@WIAgLeader How about network connected sensors for monitoring progress and controlling temperature for wine fermentations? I use a home-brew system that could be a starting point and proof of concept. Most wineries don’t have anything.

[JasonEWhite @JasonEWhite]

Another addition planned at Vines & Rushes Winery https://t.co/fXP3JwvNnG @VinesRushes #greateroshkosh @RiponPress #nepeuskun