GitHub/quartz
1
0
Fork 2
mirror of https://github.com/jackyzha0/quartz.git synced 2025-12-20 03:14:06 -06:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

tweak

Browse Source
This commit is contained in:
Courtland Leer 2024-02-23 14:45:30 -05:00
parent da3d44775d
commit ee68b6dc7e
1 changed files with 2 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
content/blog/User State is State of the Art.md
View File

@ -26,7 +26,8 @@ The knee jerk reaction to this level of understanding is to assume the problem i
Besides, surely mapping the full breadth of user identity requires much more compute and storage even than the massive AI hardware buildouts now getting underway. We're drifting into absurd [Borges-like territory](https://en.wikipedia.org/wiki/On_Exactitude_in_Science) imagining maps the size of the land they seek to represent. And didn't [Laplace (eventually) show us](https://en.wikipedia.org/wiki/Laplace%27s_demon) that demons like this are hopelessly naive?
![[escher_honcho.png]]
*[Escher](https://en.wikipedia.org/wiki/Hand_with_Reflecting_Sphere) gets it...*
*[Escher](https://en.wikipedia.org/wiki/Hand_with_Reflecting_Sphere) gets it*
## Matryoshka Representation
So is representing user identity for LLM apps a problem of [computational irreducibility](https://en.wikipedia.org/wiki/Computational_irreducibility)--no shortcuts, full simulation required?