Conversation

Notices

CrunkLord420 (crunklord420@clubcyberia.co)'s status on Friday, 24-Nov-2023 04:16:59 JST CrunkLord420
- Tard
@tard it's saying the AI successfully implemented a leaked plan by the NSA via Snowden to weaken AES keys using a thing I don't understand. It's not broken, it's just weakened. There's like no information about it, just that it's an idea and "tau statistic". This is obscure information, did the AI implement the attack based simply off that alone? Then it also requested the ability to self-modify, if I'm reading it correctly.

https://www.spiegel.de/international/germany/inside-the-nsa-s-war-on-internet-security-a-1010361.html
https://en.wikipedia.org/wiki/Kendall_rank_correlation_coefficient
In conversation about a year ago from clubcyberia.co permalink
Attachments
1. Inside the NSA's War on Internet Security
  
  from @derspiegel
  
  US and British intelligence agencies undertake every effort imaginable to crack all types of encrypted Internet communication. The cloud, it seems, is full of holes. The good news: New Snowden documents show that some forms of encryption still cause problems for the NSA.
2. No result found on File_thumbnail lookup.
  
  Kendall rank correlation coefficient
  
  In statistics, the Kendall rank correlation coefficient, commonly referred to as Kendall's τ coefficient (after the Greek letter τ, tau), is a statistic used to measure the ordinal association between two measured quantities. A τ test is a non-parametric hypothesis test for statistical dependence based on the τ coefficient. It is a measure of rank correlation: the similarity of the orderings of the data when ranked by each of the quantities. It is named after Maurice Kendall, who developed it in 1938, though Gustav Fechner had proposed a similar measure in the context of time series in 1897.Intuitively, the Kendall correlation between two variables will be high when observations have a similar (or identical for a correlation of 1) rank (i.e. relative position label of the observations within the variable: 1st, 2nd, 3rd, etc.) between the two variables, and low when observations have a dissimilar (or fully different for a correlation of −1) rank between the two variables. Both Kendall's ...
- CrunkLord420 (crunklord420@clubcyberia.co)'s status on Friday, 24-Nov-2023 04:22:24 JST CrunkLord420
  in reply to
  - Tard
  @tard this is straight out of the book "The Metamorphosis of Prime Intellect". Something that can actually do real logic, not a gay LLM that constantly hallucinates, that's spooky. It can just download the Linux kernel and find like a thousand RCE 0days in under 24 hours. It doesn't even need to break encryption.
  
  In conversation about a year ago permalink
- Pawlicker (purpcat@clubcyberia.co)'s status on Friday, 24-Nov-2023 04:44:48 JST Pawlicker
  in reply to
  - Tard
  @crunklord420 @tard literally this is what companies doing pentesting have too and the Linux kernel even has a bug finder bot now
  
  There's a blog post about finding an old aix exploit using unix system iii source too.
  
  In conversation about a year ago permalink
- CrunkLord420 (crunklord420@clubcyberia.co)'s status on Friday, 24-Nov-2023 04:44:48 JST CrunkLord420
  in reply to
  - Tard
  - Pawlicker
  @PurpCat @tard I have never seen anyone talking about a usable AI powered whole-codebase static analysis. It doesn't exist as far as I'm aware. All the code-AI stuff is just a specialized LLM afaik.
  
  In conversation about a year ago permalink
- CrunkLord420 (crunklord420@clubcyberia.co)'s status on Friday, 24-Nov-2023 04:47:48 JST CrunkLord420
  in reply to
  - Tard
  - Pawlicker
  @PurpCat @tard yeah that's just a fuzzer though. Fuzzers don't catch everything.
  
  In conversation about a year ago permalink
- Pawlicker (purpcat@clubcyberia.co)'s status on Friday, 24-Nov-2023 04:47:49 JST Pawlicker
  in reply to
  - Tard
  @crunklord420 @tard https://github.com/google/syzkaller/blob/master/docs/syzbot.md
  In conversation about a year ago permalink
  Attachments
  1. No result found on File_thumbnail lookup.
    
    https://github.com/google/syzkaller/blob/master/docs/syzbot.md
- :blank: (i@declin.eu)'s status on Friday, 24-Nov-2023 05:00:52 JST :blank:
  in reply to
  - Tard
  - Pawlicker
  @crunklord420 @PurpCat @tard people have used stuff like https://github.com/Significant-Gravitas/AutoGPT on codebases though, the "breakthrough" has been repeated like 4 times now, with chain of though, tree of thought, reasoning via planning, and now tree/graph search as in alphazero
  
  https://github.com/waterhorse1/LLM_Tree_Search
  
  at the end of the day, so far every innovation has been more/deeper invocations on itself to escape the fact that only predicting the next token only goes so far and tends to get stuck on itself if it's not given a chance to course correct
  In conversation about a year ago permalink
  Attachments
  1. GitHub - Significant-Gravitas/AutoGPT: AutoGPT is the vision of accessible AI for everyone, to use and to build on. Our mission is to provide the tools, so that you can focus on what matters.
    
    AutoGPT is the vision of accessible AI for everyone, to use and to build on. Our mission is to provide the tools, so that you can focus on what matters. - GitHub - Significant-Gravitas/AutoGPT: Aut...
  2. GitHub - waterhorse1/LLM_Tree_Search: The official implementation of paper: Alphazero-like Tree-Search can guide large language model decoding and training
    
    The official implementation of paper: Alphazero-like Tree-Search can guide large language model decoding and training - GitHub - waterhorse1/LLM_Tree_Search: The official implementation of paper: A...
  CrunkLord420 likes this.

Public

Conversation

Notices

Feeds