KI zeigt uns eins: der Austausch von Informationen per PDF muss abgeschafft werden
Saubere Weiternutzung verlangt, sauber geordnete Daten. Farbige Polylinien reichen nicht mehr
Und Probleme bei PDFs fangen nicht erst bei Zeichnungen an
https://t.co/gGkxmtUH8j
KI zeigt uns eins: der Austausch von Informationen per PDF muss abgeschafft werden
Saubere Weiternutzung verlangt, sauber geordnete Daten. Farbige Polylinien reichen nicht mehr
Und Probleme bei PDFs fangen nicht erst bei Zeichnungen an
https://t.co/gGkxmtUH8j
@nateherk curious about n8n. i guess it's graduated regarding the agentic workflows, but i would argue there are some very basic automation use cases where it still has value
or did you move all you "automatic workflows" to agents with cron jobs?
@naval partly true, mostly for SaaS
piplines in big companies are far to dependent on reliable tools for them to be replaced by AI products that run in a browser
specialized tools for big industries (like from Adobe or Autodesk) won't be affected as soon
those tools still cary weight
some things are related and similar, some aren't. connections are diverse.
the graph of obsidian basically mirrors a graph database
i wonder if a markdown file is suited as well for an LLM as a graph database is, if links are used extensively
thoughts @kepano ?
the power of obsidian comes from the fact that a note can be anything
unlike notion you can select every property at random.
no two notes need the same properties, yet share a lot of them
that mirrors reality far more efficiently
@obsdmd
facebook is an app, it's also a company
with some apps i have a subscription, some i just wrote down out of curiosity
books can be finished or unfinished. just like articles, but unlike articles they don't have a url. videos have a url. some people have a url, most don't
@GuiotNoel@obie If open source models keep on scaling like the commercial ones and computing prices go down, this setup should be possible as a private solution, like you could run your own mail server
Not everyone is going to do it but if privacy is a concern, it might be possible down the road
Computer use is going to change massively. So are meetings and knowledge distribution.
AI isn't going to change everything but it's going to change hell of a lot
@obie Love this, gonna try it myself!
About those 82 meeting notes: what are your thoughts on that? Do you even read those or just process it with the AI? If so, are those notes even needed in the future? Or is everything going to be recorded and brought up by AI if needed?
i recently started using @obsdmd sync as a GDrive replacement and tickler file
put a ticket for a conference into my reference folder, the date i need it into the title & use daily note with a base to fetch it on that day
also local, so no connection headaches once its synced
@arscontexta@Antisimplistic i was wondering about that as well
i recently read the philosophy behind the vault of kepano
prior i structured notes in folders, now i embrace the relations to spark ideas
have you tried putting all into one vault & letting claude find relations between projects and thinking?
Ideas rarely (if ever) come from nothing. They are much more often the results of mistakes, coincidences, the fact that youโve read something a long time ago, which you had nearly forgotten, and a side comment of a friend over dinner that sparks a thought.
I didn't get the bases in obsidian at first. That changed after reading how there CEO approaches note taking
Check out my stumbling into his bottom-up approach here:
https://t.co/DiMlRFPDcH
@obsdmd@kepano
@kepano I tracked the "tournament" in a Canvas
Then put a screenshot into GPT/ Gemini and asked what the results tell about my definition of success, how, based on this, my goals for 26 might need to be adjusted and if the results confirm or disagree with Chats we had in the past
@kepano I didn't actually use Claude Code but GPT and Gemini but this would work perfectly as well:
I used the AI to find people I reference often (YouTubers, authors, entrepreneurs). Asked for 32 people (random order), then asked myself which person I would consider more successful