This piece by Casey Newton (via Chris) highlights where I've been with regards to notes and research for a while.
One interpretation of these events is that the software failed: that journaling and souped-up links simply don't have the power some of us once hoped they did.
It's not the gathering of ideas and information (that's the easy bit as long as you have the discipline to do so – I often don't) but making something from it.
You can have the most complex of systems going but just having the info is not enough.
I've tried various things:
- a wiki plugin in WP
- my own wiki-like Garden (both in WP and here)
- installed Obsidian on my Mac
- a notes page
- designed Sparks
I didn't get on with the wiki plugin (it was uninstalled pretty quickly) but had a bit more success with the Garden (it was the place where I built It's Only Words)
Obsidian was a no go from the start – just like the wiki plugin, it didn't fit with the way I work.
My notes page serves primarily as an aide memoir and occassional staging ground for posts but not much else. Sparks hasn't found its role yet.
Jim Nielsen gives details on his notes page and has a "shuffle" button which scrolls to a random note.
I think a lot of value comes from juxtaposition and the serendipities afforded by this - it's in making connections between things you wouldn't normally do – lateral thinking.
It's this part that I've always struggled with.
Jim's 'shuffle' feature gets me thinking. What if we took it one stage further: not only does shuffle take you to a random item on a page but it also randomises the order of items so that it creates unusual and unique juxtapositions.
This is essentially what I tried to do on the /random page with it loading two posts each time rather than just one. I have now increased it to three as a test. This, however, is just showing things I've already written about rather than surfacing new thoughts and ideas.
AI?
The potential of AI to have an impact is interesting here.
One of the AI features introduced into Arc was to get a page summary from the usual CMD + F search box. It could be used to summarise a spark file or notes page to provide a summary of everything linked to a particular topic but is that enough?
That only gets us part of the way. Current AI models use established patterns of language, seek common denominators to give a 'best fit' answer based on probabilities. They try to be correct which can be a little boring.
This is why I enjoy chatting with Pi rather than ChatGPT, it behaves differently and seems open to a bit more vaguery – even if it's only a trick of the light.
We would need AI to behave more like us, to add 2 & 2 and make 5, explore random paths and reach illogical conclusions. We need it to ask "what if?" instead of just dealing with facts.
Until then, no system will be able to do what we can do. Maybe never will. There's something special in the imperfections of our thought processes, in the ability to take the incongruous and find patterns that current tools cannot replicate.
Still, I'm getting away from my original train of thought. What was it again?
