diff --git a/orbitmines.com/src/routes/papers/2024.AUniversalLanguage.tsx b/orbitmines.com/src/routes/papers/2024.AUniversalLanguage.tsx index a278da6..7195958 100644 --- a/orbitmines.com/src/routes/papers/2024.AUniversalLanguage.tsx +++ b/orbitmines.com/src/routes/papers/2024.AUniversalLanguage.tsx @@ -276,6 +276,36 @@ const AUniversalLanguage = () => { This is essentially what it means to point to something you don't yet understand: I point in some direction without having to define what that something is. +
+ + This allows us to phrase what it means to have unintended, ..., unrealized effects: Essentially phrasing resources you didn't know you were consuming: + + + + + +
+ + This begs the question: How do you phrase an expectation for something which isn't yet known? + + + + + + What do you do when the information you want is not available? + + + + + + How do you determine that something from outside available context is looking inside? + + + + + + What about effects which aren't abstractly realized, yet still have an effect on the outcome? [REPHRASE ; No causal links on one language, but the superposed one knows about them - how else could I point it out?] +
Now that we can superpose languages, and state with better clarity what having access to certain {
A Reprogrammable (Visual) Interface} sub="Open inputs, outputs, compute substrate, ..., interfaces"> + Now we finally get to the points which are the most interesting ones to me. All this allows us to phrase a setting, where we have open inputs, outputs, compute substrate, ..., interfaces. + +
+ + Where inputs and outputs are an extreme: + + + + + + A substrate being a superposed language: + + + + + + Intermediate steps being the program: + + + + + + And the interfaces we have on it, being additional levers of control, or reading out intermediate steps (which our system could be ignorant of). + + + +