does ip bouncing exist?

[r-ID]

I do understand what are you saying completely and i do agree with you, but... I'm a hacker, I'm thinking of a way to "hack a system". There is no way to secure communicate between two points, right? But what can we do? Can we avoid MITM attack? That's not possible Just an idea... We can make it 99.99% secure (with today's technology - CPU speed) I'm not talking about SSL or any other protocol. I'm talking about intelligence and timing. How fast do you think anyone can reverse-engineer my code? Any new algorithm will break any MITM possibility.
Oh, and i like to think that you are 100% secure-connected with the machine when you copy paste public key physically.

[sanddbox]

Could you clarify how CPU speed relates to secure data transmission? You lost me.

Your ISP knows about all of your internet traffic. Short of creating your own internet service, proxy servers, etc, you're not going to have guaranteed privacy - and even then it's still not 100% guaranteed.

[r-ID]

to the author:
"IP bouncing" (as you call it) do exist,
A bouncer i made http://img85.imageshack.us/img85/3492/snapshotu.png
Does IP tracing exists? No.

Sanddbox, you can be secure only now, as CPU power increases - old encryption methods become obsolete (algorithm changes, bit size increases)

yes, my ISP can read everything i send, but... if my ISP is trying to do MITM attack he has only one chance in his lifetime. If i succeed to send PK, he can no longer able (never ever after) to do MITM attack, my connection stays secure (allways). If ISP is planning to do MITM attack he should be prepared, but he can't because he doesn't know an algorithm i will use.

[sanddbox]

Tell me, then: How would the server or host you're sending the data to decrypt it?

[r-ID]

you don't need to worry about decryption, you have to worry about sending PK.

basic communication:
1) you generate private key and public key part.
2) you send public key part (MITM attack possibility)
3) you ask for block cipher key
4) remote machine encrypts block cipher key with your public key.

And i think that more dangerous point is end point, and not your ISP.
If you are afraid of your ISP, you can connect to any already secured machine outside ISP range and then send PK. (your ISP can do nothing in this case)

oh, and the next time you are connecting, you start from 3), noone can MITM attack you now.