Yesterday I wrote the phrase “sentiment hardpoints” in my notebook while sketching a layout for my new home office (by far the largest I’ve had). By which I mean things like strategically positioned pictures, hourglasses, clocks, lego models etc. Cracked myself up. But it’s a useful concept. Think of them as “pinned emotions” in your stream of consciousness. Sentiment superstates/vibes you try to ruggedize.

Kinda excited about my plan, which is to conceptualize the space as an extended prosthetic memory with loosely computer-inspired architecture. Register (desks/workbenches I’m at), cache (L1 = whiteboard, wip builds at active stations, L2 = easy-reach bookshelf with active notebooks, open project boxes, L3 = current interest bookshelves, box projects, perhaps flipchart sheets), dram (farther bookshelves, posters, rollaway stations/archival project boxes and part drawers), disk (archival self/abandoned self books, finished projects). There are also uncollected garbage/memory leak loci (miscellany drawers, unorganized book piles). I’d like to map something to kv cache specifically but can’t figure it out.

Essentially an externalized mind palace since my actual memory is weakening, and thanks to aphantasia the regular mind palace idea doesn’t work for me anyway.

The “sentiment hardpoints” idea is derived from human rather than computer memory. As in emotionally charged motifs that anchor narrative arc gestalts. Orientation fixed points for memory (which may also be physically fixed like pictures on walls, but I use “fixed” to indicate that their significance is unlikely to shift much in renarrations of self). For eg, clocks/hourglasses = reminders of time as a long-arc interest. These are load-bearing rather than nostalgic.

I need a couple more placemaking design elements like this to truly make this an extension of my head as opposed to a generically well-laid-out office/studio/workshop. It’s not yet opinionated enough and has insufficient personal narrativium.

This exercise made me realize I need certain books I read recently in ebook form in print form (eg labatut, Ballard, lovecraft)

Also, this makes me rethink the idea of a Jarvis-like robot-AI assistant for the space. I think this memory-forward approach would make the Jarvis design spec closer to an externalized tulpa or homunculus than a butler or assistant as such. A very organic take on the idea of a “smart room.” Quite different from say Bret Victor’s Dynamicland, which is about UX affordances rather than personality extension. I think current directions in AI favor the latter. The idea can be extended to teams too.

Increasingly liking my phrase Intimate Computing (IC) for this kind of thing. A visitor to my office should feel like they’re inside my head. Of course that would only work if your head is a clean, well-lighted place like mine rather than a messed-up one.

Cc Kei Kreutler Sachin might be interesting for you