I want to think more about some connected ideas. Silicon Jungle's prototypes where he is exploring breaking ideas into cards, I think called lenses, that can then be further explored.

Linus Lee's talk about using LLMs as a microscope for thoughts.

Some work experiments where we've been using LLMs to break scenes down into parts, then applying a transformation to one of those parts.

I'm writing this on the train so references are from memory

I really liked Silicon Jungle breaking it down into cards, it felt like a useful metaphor. Especially I thought of deck-building. There I'm still influenced by a short obsession with Marvel Snap. I like that, like building a good game deck, you want both supporter ideas and big powerful ideas. I think maybe pushing the metaphor further and giving cards values and play cost wouldn't be very useful - maybe some loose categorization to help with how to fit them together.

Also connects to having a "bag of tricks". There was a Ben Pierrat project that made use of that idea - something with "fake" in it?

And then there's the Feynman idea I always return to, of having ten favorite problems and every time you learn a new thing you see if they are a good fit for any of your problems.

A hard part of this is what abstraction level to break things into. I'm not sure the card thing solvjes that exactly, but it makes it easier to accept. We're used to cards of various value (does value map to abstraction).

I still want to do an abstraction focused prototype - maybe using trees.

Cards also have the connection to Oblique Strategies, and Tarot I suppose to. They're a nice container, tactile point.

Are there good connections to computers? I think punch cards doesn't directly apply. Solitaire is a computer classic, for whatever reason. That infinite card animation would be fun to play with.

I do think a big part of the card appeal to me, which has been a theme for me lately, is making stuff feel contained, manageable, understandable. Instead of thinking about something with every possibility available, it is constrained. It'd even be fun to pick out decks with specific angles, though the idea you're going to want to assemble multiple decks is maybe over-ambitious - maybe automate that part too.

Let's try Linus's angle on this - the microscope lets you see the strands that make up an idea, more like DNA. It does feel similar. The card base makes it feel more manageable to me, but there may be biology tools and visualizations and metaphors that offer something else.