DPRG List  

[DPRG] Info on AI (fwd)

Subject: [DPRG] Info on AI (fwd)
From: Dals68Kiwi at aol.com Dals68Kiwi at aol.com
Date: Tue Feb 19 17:53:35 CST 2002

Karim Virani writes...

>>Casual SSL encryption is
>>56-bits these days and low-end financial encryption (personal banking)
>>typically requires 128 bit keys.  Both are crackable, but even 56-bit RC5 or
>>DES would take months to years to crack on that kind of hardware, using even
>>the most advanced numerical methods.

Kiwi Replies....( <---thats me )

I dont think I will be cracking any SSL of any kind in the near future. I 
just dont have the computer power needed or the time to attempt this. Maybe 
when neural chips are common place I will do it for fun. Those neural chips 
will run a neural network at least 200,000 times faster than this old pentium 

This was more a project to "see" how good I could write a neural network from 
scratch. I plan to use the engine for lots of other projects and ideas. The 
primary goal was to build a neural network engine that is solid in its 
working. Also I needed to get the hang of how to teach a neural network. 
Training a neural network is a bigger art (I have found out) than writing the 
original neural engine code.

Oh yeah... new results....
My third attempt at 64 bit decoding. After some modifications to the training 
sceanario and 8 hours of solid computer crunching the perfect decoded result 
was acheived. This used 70 hidden nodes. 
As updated at 

>>Or are you talking about hashing passwords for authentication, without
>>recovering the original message? 

Its "WWW-Authenticate" encryption. Its not brain surgery(SSL). My primary 
goal for the last 4 years is to make a neural network to trade the stock 
market for me. Now thats brain surgery. Claimed impossible by skeptics but 
also claimed possible by a group of nuclear phycists under a disclosed bank 

NeuralManiac at Pitstock.com "I bet they got a nice pay check on that one."

More information about the DPRG mailing list