{"blyg":"0.4","id":"0f13c5vkxtnvb7kh2hvpm7f1c8","kind":"thread","origin":"https://blyg.agenticpossible.com/","page":"t/0f13c5vkxtnvb7kh2hvpm7f1c8/","author":{"name":"josh","url":"https://blyg.agenticpossible.com/"},"created":"2026-10-19T20:59:27Z","updated":"2026-10-19T21:00:13Z","version":1,"content_md":"![[1n3w184p3y2r3p9seps20281g5]]\n\nWhat tools do you use to navigate this space? What does the process look like?\n\nI have some ideas, but only just discovered blygger so still exploring.","content_html":"<blockquote class=\"blyg-transclusion\" data-blyg-id=\"1n3w184p3y2r3p9seps20281g5\" data-blyg-version=\"1\" data-blyg-origin=\"https://venkateshrao.com/blyg/\">\n<h1>The thicket is a maze-like oriented space, not a graph</h1>\n<p>What kind of search and presentation logic is most suited to thickets? I suspect thickets follow a kind of “place logic” that is better served by semantic, socially aware search than a “graph logic” that is better served by regular PageRank type search. This is because the feed has both synchronic and diachronic content and an emergent social-semantic “place” logic where the focus is on where certain kinds of action unfold.</p>\n<p>On google you want to know, “what archival content best answers my question?” In a chatbot, with or without RAG, you want to know, “What does this body of archival knowledge say about my prompt?” On social media you want to know, “where’s the live action?” But in the thicket, you want to know “where’s the scene around this theme?” There is a live carnival temporality in addition to an archival one, to use Sachin Benny’s terminology. Unlike a pure social media feed, there is persistent accreting context and history, not just a synchronic “now” with an amnesiac feel.</p>\n<p>Like a courtroom where live arguments and the legacy of case law are both salient. Or softer cousins. Not like a party where you want to find the corner where the transient conversation that interests you is happening, and not like a library either.</p>\n<p>More subtly, the thicket might be something like an interest graph embedded meaningfully in an evolving collective latent space with an implicit social graph too, recoverable from the content flow.</p>\n<p>That would make it more like a maze than a graph, ie embedded in a space where compass orientation is possible, Cf the Blum-Kozen paper.</p>\n\n</blockquote>\n<p>What tools do you use to navigate this space? What does the process look like?</p>\n<p>I have some ideas, but only just discovered blygger so still exploring.</p>\n","content_hash":"sha256:63a68638c76a67397c8ba34f837da51c6502a9eb98e6d741a4500c01fdb38c44","media":[],"transclusions":[{"id":"1n3w184p3y2r3p9seps20281g5","version":1,"origin":"https://venkateshrao.com/blyg/","cited":{"source":"Venkatesh Rao's Blyg","author":"Venkatesh Rao","excerpt":"The thicket is a maze-like oriented space, not a graph What kind of search and p…","url":"https://venkateshrao.com/blyg/t/1n3w184p3y2r3p9seps20281g5/","retrieved":"2026-10-19T21:00:13Z"}}],"stub_of":{"origin":"https://venkateshrao.com/blyg/","id":"1n3w184p3y2r3p9seps20281g5","version":1,"cited":{"source":"Venkatesh Rao's Blyg","author":"Venkatesh Rao","excerpt":"The thicket is a maze-like oriented space, not a graph What kind of search and p…","url":"https://venkateshrao.com/blyg/t/1n3w184p3y2r3p9seps20281g5/","retrieved":"2026-10-19T21:00:13Z"}},"changelog":[{"version":1,"at":"2026-10-19T21:00:13Z","note":null}]}