Skip to content

Commit

Permalink
why we need bootloader
Browse files Browse the repository at this point in the history
  • Loading branch information
mastercyb committed Jul 6, 2024
1 parent fd9b94b commit 51ca6c1
Show file tree
Hide file tree
Showing 2 changed files with 42 additions and 1 deletion.
3 changes: 2 additions & 1 deletion journals/2024_07_06.md
Original file line number Diff line number Diff line change
@@ -1,4 +1,5 @@
- introduction to [[cybergraph]]
- [[three basic arguments]] as foundation for [[explicit knowledge]]
- [[implicit knowledge]] is computable
- [[cyber/congress/fellows]]
- [[cyber/congress/fellows]]
- [[why we need bootloader]]
40 changes: 40 additions & 0 deletions pages/why we need bootloader.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,40 @@
tags:: cyber

- first of all [[superintelligence]] does not exist yet
- deploy such a thing requires enormous effort from smartest minds
- the scale and consequences of mistakes are nothign to compare with in our history
- even manhatan project and human in space fades in complexity and responsibility
- [[cybercongress]] is limited in expertise and resources
- so it have been decided to test all assumptions first
- think of [[bostrom]] as useful experimental ground
- which will open access to more expertise and resources needed for the goal
- quantum resistance hashing
- [[quantum computing]] demands [[quantum resistant hashing]]
- our opinion is that 32 byte [[particle]] space in longterm is not collusion resistant
- currently [[particles]] use [[cidv0]] which is based on [[sha256]]
- do you remember how one guy said that 640kb must be enough?
- we need significantly bigger [[particle]] space
- quantum proof hashing in foundation requires simple and efficient algorithm
- [[cyber]] must be based on at least 64 byte [[particle]] space ensuring deep future
- our hardware just not ready yet for this
- another problem to solve is quantum resistant signatures
- although some signature schemes exist, including being used in [[$QRL]]
- we convinced that [[hash based signatures]] looks more simple
- but they are not yet advanced enough for blockchain in production
- more efficient computing with [[8bit symbolic table]]
- basic encoding being ubiquitously used is utf8
- big problem with utf8 is that it is information inefficient
- obviously utf8 contain so much excess information
- so practically if we came to 256 symbols which are universal and really important
- we could make computing at least one or two order of magnitude more efficient
- stable [[zero knowledge]] and [[fully homomorphic encryption]] tech
- more stable computing paradigm with [[automatic parallelization]]
- starting semantic core
- in nature nearly all newborns have some starting semantics
- but it is not obvious which one must be there
- using [[bostrom]] allow us to build foundation for such semantic core
- [[advanced cryptoeconimcs]]
- there is a shaky dream about [[fuzzy hashing]] instead of [[strict hashing]], which is not likely
- this is surely not a full list of problem to solve
- [bootloading](bootloader) takes time to answer hard questions
- so enjoy [[bostrom]] and [[join]] the movement now

0 comments on commit 51ca6c1

Please sign in to comment.