Tana News - Issue No. 005
Slack and Tana Updates, GPT/LLM Musings
Introduction
Another week, another newsletter. I should stop estimating the day of release, I think it'll only jinx it. Happy Friday everyone and congrats to all the Tana Ambassadors!
In Todays Issue:
Slack updates
New section on Tana team updates, summarizing release details.
FAQ Highlights
GPT / LLMs, Personal Graphs and Musing on the Future
Week 11 Tana Community Search Node
Popular slack posts this week:
A way to alias node references by @Emmanuel Galanos
Keyboard shortcut for add link to node
Outside of slack this week:
@Fis Fraga (Felipe) showing essential elements for a Tana project
v4 of Tana News from @Alex Idler
Learning Tana? Renita thinks it takes 5 days to Proficiency
A nice reminder for Navigation Shortcuts from @Andrea (Tana Nodes)
@Renee De Four with a tip for viewing inline reference cards
Lukas gets an excellent review for Tanarian Brain
@Ev Chapman talks about how you can do both Simple and Complex in Tana
Charlotte had her mind blown by @Eric Ho’s Pioneers Hub Tana Setup
Templates this week:
Some love for Freelancer OS and SN(A)CK
Team Updates:
Another fixes and improvements update from @Olav Sindre Kriken (see also below new Release Updates Section)
Reminders and updates:
Great to see people sharing their tips and tricks in the #share-tips-and-how-tos
Tana Jam #14 is coming up this Tuesday (21st)
❓ FAQ Highlights
Tana Quick update from Dec 2022 Office Hours?
How to install a CSS theme with Stylus
Is there a source for some predefined CSS?
CSS Repository Link from above.
♾ Tana Templates in the Wild
Tana Source Library By: Braydon Coyer
💾 Tana Release Updates
I’ve added a few previous release updates from slack that will show up as a live search in Tana. Most recent update from March 22, 2023 below.
🕸️ GPT-4 / LLMs and Personal Graph Musings
I've been diving deeper into GPT-4 and related LLMs and am fascinated by the possibilities in this space. Development seems to be moving so fast and it's been hard to keep up!
I think integrating LLMs into personal graphs will be inevitable and a few programs like Mem.ai and Obsidian plug-ins have integrations right now that gather insights from the data in your graph. I'm sure there are many more and will continue to be in the days / weeks to come.
I haven't tried any of these out yet and have been mulling over whether I want to give an AI access to a personal graph or not.
Putting aside the privacy concerns for a moment I think there is a lot of potential in connecting LLMs into a personal graph. The examples from the recent launch of Microsoft 365 Copilot seem to show a lot of promise and that's the world I live in in my 9-5.
In a personal graph, there are a few use cases I could see being helpful for my creative work / personal life:
Supercharge queries, using LLMs to find related context in my notes for combining information for an article for instance. I have a lot of "highlights" in my graph right now that haven't been "recursively summarized" that much yet, but would still be valuable fodder for an article if the context matched what I was writing about. Maybe a bit backwards in terms of how zettelkasten should go, but a time saver and more in line with a "just in time" surfacing of information as needed, not filed away meticulously for someday that may never arrive.
Drop an audio file from a meeting or brainstorm into Tana, have Whisper transcribe it, GPT-4 parse the action items and summary for me, tag it with relevant tags / subjects / areas, and load it back into my graph. I do this in some form already with long walks where lots of ideas pop up, but having a streamlined process would be a real timesaver!
More supercharged queries, grabbing all notes with tag X and Y, from this [date range] and covering this [topic]. Being able to create complex queries with natural language will become essential I think the larger my graph (and other's grow), assuming I will be building one for years to come.
For Tana, with a fully built out API, it's only a matter of time for integrations like this.
One future scenario?
I'm not an authority in this space, but these are some of the questions and initial ideas I'm mulling over as I watch with amazement at how quickly this type of computing is sweeping across so many industries.
Until next time! Don't be shy, let me know how you're liking the format, the frequency (sorta weekly), and if you'd like to see more long form "opinion" pieces or straight up facts and updates.